Lessons from the Cancer Genome
Levi A. Garraway1,2,4and Eric S. Lander3,4,5,*
1Department of Medical Oncology and Center for Cancer Genome Discovery, Dana-Farber Cancer Institute, Boston, MA 02215, USA
2Department of Medicine, Brigham and Women’s Hospital
3Department of Systems Biology
Harvard Medical School, Boston, MA 02115, USA
4The Broad Institute of Harvard and MIT, Cambridge, MA 02142, USA
5Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
Systematic studies of the cancer genome have exploded in recent years. These studies have re-
targets in cancer. The genes affect cell signaling, chromatin, and epigenomic regulation; RNA
splicing; protein homeostasis; metabolism; and lineage maturation. Still, cancer genomics is in
its infancy. Much work remains to complete the mutational catalog in primary tumors and across
the natural history of cancer, to connect recurrent genomic alterations to altered pathways and
acquired cellular vulnerabilities, and to use this information to guide the development and applica-
tion of therapies.
More than a century ago, Theodor Boveri proposed that cancer
is caused by chromosomal derangements that cause cells to
divide uncontrollably (Boveri, 2008)—that, in modern terms,
cancer is a ‘‘disease of the genome.’’ It took 70 years for molec-
ular biologists to prove this concept by showing the existence of
1982). By the mid-1980s, researchers had established two main
suppressor genes) and had defined the genomic alterations
that give rise to them (e.g., nucleotide substitutions, chromo-
somal copy number alterations, and DNA rearrangements; re-
viewed in Macconaill and Garraway ). These studies also
began to suggest considerable complexity in the mutational
origins of cancer, with cancer-causing genes varying across
and within tumor types and with multiple genes contributing to
In aninfluential commentary in 1986, Renato Dulbecco argued
that the complete sequence of the human genome would be an
essential tool for systematically discovering the genes that drive
we must now concentrate on the cellular genome,’’ he wrote.
‘‘We have two options: either to try to discover the genes impor-
tant in malignancy by a piecemeal approach, or to sequence the
whole genome.it will be far more useful to begin by sequencing
the cellular genome.’’ Responding to this and other calls, the
Human Genome Project (HGP) was launched in 1990. A ‘‘draft’’
sequence was completed by 2000 (Lander et al., 2001; Venter
et al., 2001) and a near-complete sequence by 2003 (IHGSC,
With the availability of the genome sequence, cancer
researchers rapidly began to develop a new field of ‘‘cancer
genomics.’’ Cancer genomics involves systematic studies of
(some or all of) the genome to find sites of recurrent derange-
ment in specific cancer types. Pioneering genomic studies at
the Sanger Institute and Johns Hopkins uncovered genes
mutated frequently in melanoma and colon cancer, respectively
(Davies et al., 2002; Samuels et al., 2004). Studies by several
groups in Boston and New York then discovered frequent
activating mutations in lung cancer, which largely explained
patient response to a drug (Lynch et al., 2004; Paez et al.,
2004; Pao et al., 2004). Soon thereafter, a working group of the
U.S. National Cancer Institute proposed a ‘‘Human Cancer
Genome Project’’ (see http://www.genome.gov/Pages/About/
onBiomedicalTechnology.pdf), which came to be called The
Cancer Genome Atlas (TCGA). The NCI launched a pilot project
for TCGA in 2006 and a full project in 2009. In parallel, an Inter-
national Cancer Genome Consortium was launched and has
grown to involve researchers in more than 15 countries (Hudson
et al., 2010).
The notion of taking a genomic approach to characterizing
cancer was not universally endorsed, as reflected in the title of
one commentary: ‘‘Human Cancer Genome Project: Another
Misstep in the War on Cancer’’ (Gabor Miklos, 2005). Some
thoughtful critics felt that hypothesis-driven research was the
best way to study cancer and worried that systematic studies
were so expensive that they would drive out focused investiga-
tion (Weinberg, 2010). Proponents argued that science requires
investment in both hypothesis generation and hypothesis testing
and that unbiased genomic studies were anexcellent way to find
surprises. They also expected the cost of genomic studies to
plummet, with new technologies just over the horizon. Some
scientists were skeptical because they believed that there
were few cancer-related genes left to discover, whereas others
thought that cancers were too hopelessly complicated to yield
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 17
to systematic analysis. This open debate helped to shape the
design of cancer genome projects. In the end, however, the
questions could only be answered with data.
With cancer genome projects now underway for several years
(Table 1), the time is right to assess the early returns and to
consider next steps for the field. (As a complement to this
Review, we recommend an earlier review by Stratton et al.
, which describes many foundational aspects of cancer
genomics.) Here, we describe the remarkable tapestry of biolog-
ical, evolutionary, and therapeutic insights that have emerged
from systematic cancer genome characterization. At the end,
we suggest the next steps for cancer genomics.
Initial cancer genome projects had to be carried out with what
by traditional capillary-based sequencing in which each exon to
be studied was amplified and sequenced individually, and chro-
mosome copy number alterations were surveyed with DNA mi-
croarrays. DNA rearrangements could hardly be cataloged at
all. The high cost and extensive infrastructure needed for
large-scale DNA sequencing placed tight constraints on the
amount of data that could be collected. Exome-scale projects
could only be carried out on small numbers of samples (Sjo ¨blom
et al., 2006; Wood et al., 2007); thus, much effort was spent
ori notions of cancer mechanisms or therapeutic targets (Ding
et al., 2008; Greenman et al., 2007; CGARN, 2008).
The emergence of massively parallel sequencing (MPS) revo-
lutionized the entire enterprise (Bentley et al., 2008; Margulies
et al., 2005). Initially, MPS made it possible to sequence nearly
1 billion bases (1 gigabase [Gb]) in a single run; this number
grew to >600 Gb/run by 2012. In parallel, methods were devel-
oped that employ hybridization to oligonucleotide ‘‘baits’’ in
aqueous solution to capture specific portions of the genome—
exons (the ‘‘exome’’) (Gnirke et al., 2009; Hodges et al., 2007).
MPS also made it possible to use a single technology platform
for all categories of genome analysis (discovering point muta-
tions, assessing copy number alterations and translocations,
measuring transcript levels, identifying alternative splicing, de-
tecting DNA methylation, and mapping chromatin structure).
The first whole cancer genome sequenced by MPS was re-
ported in 2008 (Ley et al., 2008). Whereas initial studies were
confined to single samples (Pleasance et al., 2010a, 2010b),
studies of hundreds of samples have quickly become the
norm. Plummeting costs have propelled an unprecedented
explosion of sequence data. For example, >16,000 cancer
samples had been subjected to genome or exome sequencing
by late 2012 just at our institution alone (Broad Institute).
With the MPS data came a need for completely new analytic
tools. The first challenge was to accurately determine the
sequence in individual tumor and normal samples from the
‘‘raw’’ sequence data. Each type of alteration in the DNA and
RNA required a specialized detection method, including for
single nucleotide variants, small insertions/deletions, chromo-
somal rearrangements, gene fusions, alternatively spliced tran-
scripts, chromosomal copy number alterations, and detection
of foreign DNA (such as from viruses) (Beroukhim et al., 2007;
Chen et al., 2009; Cibulskis et al., 2013; Dees et al., 2012; Kim
and Salzberg, 2011; Kostic et al., 2011; Trapnell et al., 2009).
Algorithms employ probabilistic methods to identify mutations
Table 1. Current Large-Scale Cancer Genome Projectsa
Anatomic SiteTumor Type
pediatric: pilocytic astrocytoma
Head and neckhead/neck squamous cell cancer
lung squamous cell carcinoma
Breast breast lobular carcinoma
breast ductal carcinoma
breast HER-2 positive
breast ER positive vs. negative
esophageal squamous carcinoma
colon cancer (non-Western)
Gynecologicovarian serous cystadenocarcinoma
cervical cancer (squamous + adeno)
Urologic renal: clear cell carcinoma
renal: papillary carcinoma
renal: chromophobe carcinoma
prostate adenocarcinoma, early onset
Soft tissue (Sarcoma) solitary fibrous tumors
extraskeletal myxoid chondrosarcomas
Hematologicacute myeloid leukemia
lymphoma: chronic lymphocytic leuk.
lymphoma: germinal B cell
lymphoma: diffuse large B cell
chronic myeloid disorders
aIn conjunction with The Cancer Genome Atlas, International Cancer
Genome Consortium, and Slim Initiative for Genomic Medicine.
18 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
or rearrangements based on their presence in multiple tumor
sequence reads and absence in the paired normal DNA
sequence (Meyerson et al., 2010).
Detecting mutations with high accuracy turned out to be
surprisingly tricky. Because somatic point mutations in cancer
are so infrequent (?1/Mb), the background error rate must be
are true positives. Many false positives initially arose from
sequencing errors and inaccurate alignment of reads to the
human genome. In addition, false negatives may arise from
admixture of noncancer cells (tumor purity), copy number varia-
tions inherent in cancer genomes (ploidy), and the presence of
variant subclones within the cancer cell population (heteroge-
neity) (Carter et al., 2012). Increasing the sequencing depth
(average number of reads per base) was found to improve
both the specificity and sensitivity of mutation calling. Currently,
tumor sequencing is performed with 100- to 150-fold coverage
for whole-exome analysis and 30- to 60-fold coverage for
whole-genome analysis. (The whole-genome sequence should
ideally be deeper but is currently limited by cost.)
Obtaining accurate mutation calls for a collection of individual
samples is only the first step. The harder challenge is to distin-
guish between ‘‘driver’’ events that are causally related to the
development of cancer and random ‘‘passenger’’ events that
have simply accumulated over the course of development and
cell growth. This requires determining which genes show signif-
icantly more mutations than random expectation. Sophisticated
mathematical methods are needed to ensure that the ‘‘random
expectation’’ properly accounts for (1) variation in background
mutation rates across the genome, (2) variation across tumors,
and (3) variation in purity and heterogeneity (Chapman et al.,
2011a; Dees et al., 2012; Hodis et al., 2012; M.S. Lawrence,
personal communication). Without such corrections, genes
may be spuriously declared to be drivers, with the problem
growing worse as sample size grows (because even modest
deviationsfromexpectation willappeartobesignificant). Recent
studies have highlighted some likely spurious results and
have developed solutions to eliminate them (G. Getz, personal
communication). Perfecting these algorithms remains an area
of active research. Other algorithms have been developed to
study the structure of amplifications and deletions to detect
the possibility of multiple target genes within a given locus (Ber-
oukhim et al., 2007; Beroukhim et al., 2010).
The explosion of genomic data quickly shed light on the muta-
tional processes of cancer, revealing an unexpected richness
of mechanisms. These insights may propel a deeper under-
standing of factors governing genome integrity and tumor evolu-
Initial plans for cancer genome analysis assumed a single
uniform background mutation rate (?1/Mb). In fact, cancer
mutation rates turned out to be much more variable, ranging
from as low as one base substitution per exome (<0.1/Mb) in
some pediatric cancers to thousands of mutations per exome
(?100/Mb) in certain mutagen-induced malignancies (such as
lung cancer and melanoma). Moreover, mutation rates were
found to vary substantially across the genome, governed by
processes such as transcription-coupled repair and replication
Cancer genome sequencing has also revealed a wide array of
mutational patterns both across and within individual tumor
types. Their distinctive characteristics may reflect extrinsic
factors (e.g., UV light or tobacco smoke) or intrinsic patterns
such as DNA repair deficiencies. For example, a recent study
(Nik-Zainal et al., 2012a) pointed to at least five distinct nucleo-
tide substitution patterns, most of which occur by as-yet
unknown mechanisms. One such process, which produces
C > A, C > G, or C > T substitutions at TpCpX trinucleotides, ap-
peared to underpin most nucleotide substitutions in ?10% of
ER-positive breast tumors. These studies also discovered
a new regional hypermutation mechanism characterized by
multiple base mutations that occur in cis near rearrangement
breakpoints. Termed ‘‘kataegis’’ (Greek, ‘‘kataegis’’ = ‘‘shower’’
or ‘‘thunderstom’’), this process likely involves the activation-
induced deaminase (AID) and apolipoprotein B mRNA-editing
enzyme catalytic polypeptide-like (APOBEC) protein families.
New mutation patterns (in this case, A > C transversions at
‘‘AA’’ dinucleotides) have also been discovered in esophageal
cancers by large-scale sequencing (A.M. Dulak, personal
Chromosomal Gains and Losses
cancer genome studies have yielded a systematic assessment
of large-scale (whole chromosome or chromosome arm) and
focal copy number aberrations. The typical cancer cell exhibits
large-scale gains or losses involving a quarter of its genome
and carries focal events affecting 10% (Beroukhim et al.,
2010). Based on current sample collections, many focal amplifi-
cations and deletions have been localized to ‘‘peak’’ regions
containing a median of 6–7 genes (although the number is
150–200 in some cases). For the majority of focal events, the
driver gene(s) still cannot be assigned definitively.
One of the most striking mutational patterns unveiled by
whole-genome sequencing studies consists of a catastrophic
phenomenon that produces dozens or even hundreds of rear-
rangements. The resulting disarray is distinctive for two reasons:
it is typically localized within one or a few chromosomes, and it
usually involves only two distinct copy number states (Stephens
et al., 2011). Termed ‘‘chromothripsis’’ (Greek, ‘‘thripsis’’ =
‘‘shattering’’), this process occurs in ?2%–3% of human
cancers, with an elevated prevalence in bone cancers, pediatric
medulloblastoma, and neuroblastoma (Molenaar et al., 2012;
Rausch et al., 2012). Genomic shattering appears to develop
as a result of erroneous chromosome segregation during mitosis
and the subsequent entrapment of individual chromosomes
within ‘‘micronuclei’’ (Crasta et al., 2012). Micronuclei have a
tendency toward premature chromosome condensation, which
may result in pulverization of chromosomal segments. Chromo-
somes that survive this process, having undergone aberrant re-
assembly through nonhomologous end-joining, emerge with
dense rearrangements that may sometimes dysregulate cancer
genes (Stephens et al., 2011).
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 19
A whole-genome sequencing study of primary human prostate
cancer (Berger et al., 2011) uncovered a distinct category of
complex chromosomal rearrangements.
genomes often exhibit ‘‘chains’’ of copy-neutral rearrangements
that consist of ?4–12 distinct breakpoint junctions distributed
across multiple chromosomes, with the breakpoints forming
a ‘‘closed chain’’ (A to B, B to C, C to D, and finally back to A)
that distinguishes the process from chromothripsis or other
complex rearrangements. Closed chain rearrangement break-
points tend to occur near ‘‘open’’ chromatin (that is, transcrip-
tionally active chromatin) in prostate cancer genomes harboring
ETS transcription factor rearrangements but near‘‘closed’’ chro-
matin in certain ‘‘ETS-negative’’ prostate cancers. These chains
have recently been termed ‘‘chromoplexy’’ (Greek, ‘‘plexy’’ =
‘‘weave’’ or ‘‘braid’’) (Baca et al., 2013).
Other complex DNA rearrangements seem to arise through
errors in DNA replication (Liu et al., 2011). These may include
fork-stalling and template-switching events that trigger micro-
homology-dependent DNA priming, duplications, and DNA
template insertions (for a recent review, see Holland and Cleve-
land ). Interestingly, these replication-dependent rear-
rangements show a strong correlation with TP53 mutations in
subtypes of medulloblastoma (Rausch et al., 2012). Thus,
somatic alterations in DNA-damage-sensing pathways may
render tumor progenitor cells vulnerable to ensuing catastrophic
Insights into mutational patterns may bring a deeper under-
standing of tumor evolution. In contrast to a simple gradualist
notion in which somatic mutations accumulate steadily, tumor
evolution can be punctuated by various types of catastrophic
events (Baca et al., 2013). A fuller knowledge of mutational
processes—particularly those that preferentially enact cancer
genes—may help to identify driver mechanisms in tumors.
New Cancer Genes
A key question is whether cancer genomics has led to the
discovery of new genes and, ideally, to new classes of genes
not previously known to play a causal role in cancer. Although
much work still lies ahead, the answer is clear. The trickle of bio-
logical discoveries from early studies has become a wave, impli-
cating a wide range of cellular processes in cancer. Whereas
some of the new cancer genes encode classical signaling
proteins, most populate new and sometimes surprising cate-
gories, such as metabolism, epigenetics, chromatin biology,
splicing, protein homeostasis, and cell differentiation (Table 2).
The insights from these studies are already guiding hypoth-
esis-driven cancer research ranging from basic cell and molec-
ular biology to novel therapeutics.
Signal Transduction Pathways
Studies from the 1980s and 1990s revealed that signaling path-
ways linked to proliferation and survival played a crucial role in
many cancers. Mutations were discovered in key genes that
encode members (or regulators) of receptor tyrosine kinase
(RTK)-signaling pathways (HER-2, c-KIT, ABL, RAS, NF1, NF2,
MET, PTEN), the WNT/b-catenin pathway (APC), and the TGF
b pathway (SMAD2 and SMAD4), among others. Moreover, the
pharmaceutical industry showed that drugs could be developed
toinhibit protein kinases.The posterchild was imatinib (Gleevec)
against the ABL and KIT kinases, which proved remarkably
effective in treating malignancies driven by activating mutations
in these oncoproteins (chronic myelogenous leukemia [CML]
and gastrointestinal stromal tumors [GIST]) (Demetri et al.,
2002; Druker et al., 2001).
Recognition of the importance and druggability of RTKs moti-
vated the first unbiased sequencing surveys in the early 2000s,
which employed Sanger sequencing to examine dozens of
genes in dozens of patients. The studies quickly hit pay dirt,
with the finding of mutations in BRAF in 50% of melanomas,
PIK3CA in ?25%–30% of breast and colorectal cancers,
EFGR in 10%–15% of non-small cell lung cancers, FGFR2 in
15%–20% of endometrial cancers, and JAK2 in myeloprolifera-
tive diseases (Davies et al., 2002; Dutt et al., 2008; Kralovics
et al., 2005; Levine et al., 2005; Lynch et al., 2004; Paez et al.,
2004; Pao et al., 2004; Pollock et al., 2007; Samuels et al.,
2004). (Some of the findings came to have a major impact on
drug development and clinical treatment, including the develop-
ment of selective RAF and MEK inhibitors that have produced
dramatic remissions in melanoma and the ability to target the
use of EGFR inhibitors to the subset of lung cancer patients
who derive benefit.)
In turn, these successes led researchers to scale up
sequencing surveys to discover additional candidate genes in
signaling pathways and eventually to all genes. Recurrent muta-
tions were found in genes involved in several—sometimes
surprising—pathways not previously suspected to drive cancer.
These included the MAP3K1 and MAP2K4 genes in breast
cancer (encoding serine/threonine kinases involved in the P38-
JNK signaling pathway) (Banerji et al., 2012; Ellis et al., 2012;
Stephens et al., 2012; CGAN, 2012), RAC1 in melanoma
(a GTPase involved in the RAC/PAK-signaling module involved
in focal adhesion) (Hodis et al., 2012; Krauthammer et al.,
2012), ELMO1 and DOCK2 in esophageal cancer (two genes
that activate RAC/PAK signaling) (A.M. Dulak, personal commu-
nication), MYD88 in diffuse large B cell lymphoma (activates NF-
kB signaling) (Ngo et al., 2011), and PREX2 in melanoma (a
guanine nucleotide exchange factor that controls RAC/PAK
and PI3K signaling) (Berger et al., 2012). Remarkably, a pathway
involved in axon guidance in neurons (the ROBO/SLIT pathway)
turned out to be a target of mutations in ?20% of pancreatic
adenocarcinomas (Biankin et al., 2012). And, a pathway that
governs the oxidative stress response in all cells (the KEAP1/
NRF2-signaling pathway) is activated by mutation in >30% of
squamous lung cancers (Hammerman et al., 2012; Shibata
et al., 2008). Many of these results would have eluded hypoth-
In addition, genome-wide studies of copy number alterations
based on DNA microarrays revealed recurrently amplified genes
in signaling and cell survival pathways. MCL1 and BCL2L1,
which encode anti-apoptotic proteins that are critical regulators
of tumor cell survival, were found to be amplified in a wide range
of cancers, including breast, lung, colorectal, melanoma, and
glioblastoma (Beroukhim et al., 2010). FGFR1 was found to be
amplified in >20% of lung squamous cancer (Weiss et al.,
2010) and in ?10% of breast cancers (Chin et al., 2006).
20 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
CRKL, which encodes a signaling adaptor protein, was found
amplified in a subset of lung cancers (Kim et al., 2010).
Despite their successes, these studies were sobering in
revealing that the early promise of using a single kinase inhibitor
(imatinib) to treat prevalent oncoprotein mutations (as in CML
and GIST) was not going to be widely generalizable: most
cancers lacked a highly recurrent mutation in genes encoding
kinases (or other readily druggable targets) (Greenman et al.,
2007). This underscored the importance of more deeply probing
the cancer genome.
If there were any doubt that genomic approaches would reveal
surprises, they should have been put to rest by a pioneering
study in 2008. In this paper (which predated the maturation of
MPS technology), Vogelstein and colleagues employed an
impressive ‘‘brute force’’ approach to PCR amplify and
sequence 175,471 exons from 20,661 genes (Parsons et al.,
2008). They were rewarded with the discovery of highly recurrent
mutations in the IDH1 gene, which encodes the cytoplasmic
metabolic enzyme isocitrate dehydrogenase, a seemingly
unlikely candidate for a cancer gene (Figure 1); the mutations
affected a single amino acid in the active site. Subsequent
studies found that specific mutations in IDH1 and IDH2 (IDH1’s
mitochondrial homolog) occurred in >70% of secondary glio-
blastomas, oligodendrogliomas, and high-grade astrocytomas
(Parsons et al., 2008; Yan et al., 2009) and in ?15%–30% of
acute myelogenous leukemias (AML) (Mardis et al., 2009).
Because isocitrate dehydrogenases convert isocitrate to
Table 2. Discoveries from Cancer Genome Characterization
Cellular Process Altered by Genomic AlterationsExamples of Cancer Genes Discovered (or Extended to New Cancers*) by Genomics
TERT promoter mutationsa
MAPK signaling (oncogenes)
MAPK signaling (TSG)
PI3K signaling (oncogenes)
PI3K signaling (TSG)
Notch signaling (oncogene or TSG)
TOR signaling (TSG)
Wnt/b-catenin signaling (TSG)
TGF-b signaling (TSG)
NF-kB signaling (oncogene)
Epigenetics DNA methylation
Epigenetics DNA hydroxymethylation
Chromatin histone methyltransferases
Chromatin histone demethylases
Chromatin histone acetyltransferases
Chromatin SWI/SNF complex
Transcription factor lineage dependency or oncogene
Transcription factor other
Cell cycle (oncogene)
Cell cycle (TSG)
aActivating mutation or amplification.
bInactivating mutation or deletion.
cBoth activating and inactivating genomic events observed.
dEffect of mutations on protein function unknown.
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 21
a-ketoglutarate (a-KG) in the tricarboxylic acid (TCA) cycle, the
observation suggested a previously unrecognized link between
cell metabolism and cancer. It soon became clear that the muta-
tions caused a gain-of-function (or ‘‘neomorphic’’) activity,
whereby isocitrate was converted to a distinct metabolite: the
R-enantiomer of 2-hydroxyglutarate (2HG; Figure 1) (Dang
et al., 2009; Ward et al., 2010). How this ‘‘oncometabolite’’ might
drive cancer, however, remained a mystery.
The answer emerged from a different type of genomic anal-
ysis: genome-wide surveys of DNA methylation. The methylation
studies revealed that a subset of glioblastomas (the ‘‘proneural’’
subtype) showed a DNA methylation pattern that strongly
resembled the CpG island methylator phenotype (CIMP) origi-
nally described in colorectal cancer (Noushmehr et al., 2010).
Remarkably, the CIMP-like phenotype was tightly correlated
confirmed that introduction of mutant IDH actually caused the
CIMP-like phenotype (Turcan et al., 2012).
Unexpectedly, the mechanism was clarified by yet another
(AML). This large-scale study showed that IDH1/IDH2 mutations
were mutually exclusive with inactivating TET2 mutations (Fig-
ueroa et al., 2010), suggesting that the two types of mutations
had similar effects and were thus functionally redundant. The
TET2 protein catalyzes 5-methylcytosine hydroxylation in
a a-KG-dependent manner, and loss of TET2 produces
a CIMP-like phenotype. Studies then showed that 2HG appears
to inhibit several a-KG-dependent enzymes (Xu et al., 2011),
including Jumonji-C domain histone demethylases that affect
gene expression (Lu et al., 2012) and prolyl-4-hydroxylases
(EGLN1/2/3) that regulate hypoxia inducible factor (HIF), which
is involved in certain cancers (Koivunen et al., 2012) (Figure 1).
The surprising discoveries about IDH1/IDH2 have helped to
spark enormous interest in cancer metabolism. They have also
spawned new areas for cancer drug discovery that had little
precedent prior to these cancer genome studies.
Lineage Survival Oncogene Transcription Factors
Another important discovery concerned ‘‘master’’ lineage-
specific transcription factors (TFs). Because such TFs are
typically involved in terminal differentiation of cell types, the pre-
vailing hypothesis was that overexpression would suppress
cancer by promoting lineage maturation and cell-cycle arrest.
Surprisingly, however, an integrative analysis of genome-wide
copy number and transcription showed that MITF, which
encodes the master transcription factor that regulates melano-
cyte survival and differentiation, underwent gene amplification
in a subset of metastatic melanomas (Garraway et al., 2005).
MITF thus served as a prototype for a new category of cancer
genes termed ‘‘lineage survival’’ oncogenes.
Systematic genomic studies subsequently uncovered several
additional lineage survival oncogene TFs. Examples include
NKX2.1 in lung adenocarcinoma, SOX2 in esophageal cancer,
and CDX2 in colorectal cancer (Bass et al., 2009; Salari et al.,
2012; Weir et al., 2007). In hindsight, these TFs are analogous
to the androgen receptor (AR), a nuclear hormone TF that plays
crucial roles in proliferation and survival of normal and malignant
prostate epithelia and is frequently amplified or mutated during
tumor progression (Taplin et al., 1995). Exome-sequencing
studies of castration-resistant prostate cancer have recently
identified somatic mutations in both AR and several key coregu-
lators (Grasso et al., 2012).
One of the most far-reaching discoveries from genomic studies
has been the critical role of epigenomic changes in tumorigen-
esis, which in turn has unleashed a torrent of hypothesis-driven
studies and drug discovery efforts. Abnormal DNA methylation
and chromatin structure were known to be common in cancers,
but it was unclear whether these epigenomic changes played
a causal role in cancer or were simply a noncausal correlate of
the cancerous state. The question was settled with the recogni-
tion that ?40 genes encoding epigenomic regulators show
highly recurrent somatic alterations across a wide range of
cancer types (reviewed in Dawson and Kouzarides, 2012). Muta-
tions that affect the epigenome would seem like a highly efficient
mechanism to rewire cellular circuitry because they provide
a way to affect multiple target genes simultaneously. The next
several sections discuss various epigenomic processes related
in cancer (Figure 2).
Chromatin: Histone Modification
Genomic studies have provided clear genetic evidence that dys-
regulation of chromatin modifiers drives many types of cancer.
Recurrent mutations were found in genes encoding enzymes
Figure 1. Somatic IDH1/2 Mutations Produce the Oncometabolite
Oncogenic effects of 2HG include generation of a CIMP-like phenotype and
inhibition of a-ketoglutarate-dependent enzymes such as histone methyl-
transferases (KMT), histone demethylases (KDM), and prolyl hydroxylases
(EGLN). TET2 mutations are mutually exclusive with IDH1/2 mutations in leuke-
mias and may exert common downstream effects on DNA methylation. Mutant
IDH1/2 proteins are the targets of emerging drug discovery efforts (boxed).
22 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
that add, subtract, or interpret posttranslational modifications to
histone H3. These enzymes include histone (lysine) methyltrans-
ferases (KMTs) and histone (lysine) demethylases (KDM), which
activate or repress genes by modifying specific lysine residues;
histone acetyltransferases (HATs), which regulate transcription
by adding acetyl groups to the histone H3 tail; and histone
readers, which bind various histone modifications and recruit
additional protein complexes to carry out specific effector func-
tions (Figure 2). Among the KMTs, mutations affect the MLL
subfamily, which acts on lysine 4 of H3 (e.g., H3K4); the NSD
subfamily, which acts on H3K36 (Dolnik et al., 2012); and
EZH2, which methylates H3K27 (Morin et al., 2010). Among the
KDMs, mutations affect JARID1A and UTX, which demethylate
H3K4 and H3K27, respectively. Among the HATs, mutations
affect CREBP and EP300 (Gui et al., 2011; Morin et al., 2011;
Peifer et al., 2012).
The genes encoding histone-modifying enzymes typically
exhibit lineage-restricted mutational patterns. For example, the
NSD1 and NSD3 KMTs have so far been found rearranged
only in AML (Jaju et al., 2001; Rosati et al., 2002), and mutations
affecting the histone demethylases (HDMs) KDM5A and KDM5C
appear to occur exclusively in AML and renal cell cancer,
respectively. However, somegenes show a muchbroader muta-
a particular leukemia subtype (mixed lineage leukemia), cancer
sequencing studies have found recurrent MLL gene mutations
in a variety of hematologic and solid tumors, including small-
cell lung cancer (Peifer et al., 2012), lung squamous cancer
(Hammerman et al., 2012), gastric cancer (Zang et al., 2012),
head and neck cancer (Stransky et al., 2011), and prostate
cancer (Barbieri et al., 2012; Grasso et al., 2012). Interestingly,
the DOT1L KMT, which methylates H3K79, is not itself mutated
but becomes essential in MLL-translocated leukemias (Bernt
et al., 2011). HAT mutations are found in B cell lymphomas (Pas-
qualucci et al., 2011a), small-cell lung cancers (Peifer et al.,
2012), and medulloblastoma (Robinson et al., 2012). These
distinctive patterns suggest that mutations affecting chro-
matin-modifying enzymes contribute to cancer by disrupting
expression of specific target genes that play critical roles in
particular cell types. However, the identities of these target
genes remain unknown, and we lack systematic methods for
Affected tumors are typically heterozygous for apparent loss-
of-function alleles, indicating that haploinsufficiency for these
complete loss is cell lethal. (An exception is EZH2, in which
gain-of-function mutants are observed in follicular lymphoma
[Morin et al., 2010], whereas loss-of-function events are seen
in myeloid cancers [Jankowska et al., 2011; Makishima et al.,
2010]). This makes chromatin-modifying enzymes attractive
targets for anticancer drugs because cancer cells carrying only
one functional gene could be uniquely sensitive to inhibitors
(a synergy termed ‘‘synthetic lethality’’). Vigorous drug discovery
efforts are currently underway against many of the enzymes. So
far, the only drugs targeting histone-modifying enzymes in clin-
ical use are the histone deacetylase (HDAC) inhibitors. Ironically,
the HDAC inhibitor vorinostat is approved for treatment of mye-
lodysplastic syndromes and cutaneous T cell lymphomas,
although HDAC genes have not been found mutated in these
(or any other) malignancies.
Chromatin: Nucleosome Remodeling
Another major mutational target affecting chromatin biology is
the SWI/SNF complexes (Figure 2), which regulate chromatin
structure through ATP-dependent nucleosome remodeling (for
a recent review, see Wilson and Roberts ). The importance
of these complexes in tumor biology was initially suggested by
the discovery of biallelic deletions involving SNF5 (a core SWI/
SNF protein) in malignant rhabdoid tumors (an aggressive pedi-
atric cancer). Multiple cancer sequencing surveys then revealed
that the class of genes encoding SWI/SNF factors is one of the
most commonly mutated targets in cancer. In renal cell cancer,
41% of tumors harbor mutations in PBRM1, which encodes
BAF180, a histone acetylation reader and integral component
of the so-called ‘‘BAF’’ SWI/SNF complex) (Varela et al., 2011);
Figure2. GenesEncodingEpigenetic andChromatinRegulatorsAre
Frequent Targets of Mutations in Cancer
The enzymes DNMT3A and TET2 regulate 5-methylcytosine and 5-hydrox-
ymethylcytosine production in genomic DNA; the genes encoding these
enzymes are frequently mutated in leukemias. The histone H3 component of
the nucleosome undergoes extensive modifications involving its lysine (K)-rich
tail. Genes encoding enzymes that read, produce, or interpret these modifi-
cations are frequently mutated in cancer. Examples include histone lysine
methyltransferases (KMTs), histone lysine demethylases (KDMs), and histone
acetyltransferases (HATs). Genes encoding components of the SWI/SNF
chromatin-remodeling complex are also recurrently mutated in cancer. Novel
therapeutics targeting chromatin and epigenetic mechanisms have entered
clinical use or are in development (boxed).
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 23
only the VHL tumor suppressor is mutated more commonly in
carry inactivating mutations in ARID1A, which encodes another
BAF protein (Jones et al., 2010; Wiegand et al., 2010). Frequent
ARID1A mutations have since been observed in many other
cancer types, including up to 30% of hepatocellular carcinomas
(Fujimoto et al., 2012; Huang et al., 2012a), 34% of bladder
cancers, and 21% of endometrioid cancers. Its homologs
ARID1B or ARID2 (a component of the ‘‘PBAF’’ SWI/SNF
complex) harbor recurrent mutations in melanoma (Hodis et al.,
2012; Krauthammer et al., 2012), hepatocellular (Fujimoto
et al., 2012; Li et al., 2011), and pancreatic cancers (Biankin
et al., 2012). As with other histone-modifying proteins, the
SWI/SNF gene mutations are typically loss-of-function alleles;
they often exhibit biallelic inactivation or loss of protein expres-
sion, consistent with a tumor suppressor mechanism.
Another unexpected chromatin-related target is the chromodo-
main-helicase-DNA-binding (CHD) gene family. CHD proteins
regulate chromatin compaction during stem cell differentiation
and may also promote genome stability (Ho and Crabtree,
2010) (Figure 2). Inactivating CHD1 mutations and deletions
comprise likely founder events (together with SPOP mutations)
in a newly recognized ‘‘ETS-negative’’ genetic subtype of pros-
tate cancer (Barbieri et al., 2012), where they appear to confer
distinct patterns of genome derangement (Huang et al.,
2012b). The homolog CHD4 is frequently deleted in endometrial
cancers (Le Gallo et al., 2012). Histone H3.3 itself contains highly
recurrent hot spot mutations in pediatric astrocytoma and
a subtype of medulloblastoma (Robinson et al., 2012; Schwart-
zentruber et al., 2012). Overall, the discovery of extensive chro-
matin and epigenetic mutations by unbiased cancer genome
Systematic surveys have revealed that DNA methylation also
plays a critical role in shaping the cancer genome (Figure 2). In
particular, some cancers show a clear CpG island methylator
phenotype (CIMP). The notion that DNA hypermethylation might
define a biologically important cancer subtype in colorectal
(Toyota et al., 1999), but other reports challenged its existence—
or at least its biological relevance. Systematic interrogation of all
available methylation markers (at that time) across >100 CRC
samples provided definitive evidence for CIMP in CRC (‘‘CRC-
CIMP’’) (Weisenberger et al., 2006). Most CRC-CIMP tumors
show high microsatellite instability (MSI) (Ogino et al., 2006; Wei-
senberger et al., 2006); this is likely due to the fact that such
tumors typically have hypermethylation (and hence repression)
of the MLH locus, whose loss of expression results in MSI. The
etiology of CIMP in CRC remains mysterious, with these tumors
showing few mutations in the DNA methylation machinery.
Subsets of glioblastoma and AML were also found to have
CIMP-like patterns (as described above). In these cases, the
phenomenon islikely due, in part,to 2HGgenerated frommutant
IDH1/2 proteins (Noushmehr et al., 2010), as described above.
DNA hypomethylation also plays an important role in some
cancers. A whole-genome sequencing survey revealed that
?25% of AMLs carry inactivating mutations in DNMT3A (Ley
et al., 2010), an enzyme that catalyzes the addition of methyl
show reduced DNA methylation at the promoter of many genes
involved in cancer (Ha ´jkova ´ et al., 2012); these mutations corre-
late with poorer overall survival (Ley et al., 2010). Subsequently,
recurrent DNMT3A mutations were also found in the myelodys-
plastic syndrome (MDS) (Walter et al., 2011), a neoplastic condi-
tion that often progresses to AML.
The recognition of the key role of DNA methylation has galva-
nized interest in drugs that inhibit this process, such as 5-azaci-
tidine and decitabine. Conceivably, these drugs may act in
a synthetic lethal manner against tumors carrying mutations in
DNMT3A and other genes affecting DNA methylation. Azaciti-
dine has proved especially intriguing: it is the first drug to
improve the survival of patients with myelodysplastic syndrome
(MDS) and has also shown promising efficacy in AML (reviewed
in Estey ). DNMT3A mutations or other altered methylation
phenotypes may define leukemic patient subpopulations that
are more likely to benefit from these drugs (Marcucci et al.,
2012). As with chromatin dysregulation, the critical genes
affected by aberrant DNA methylation remain unclear.
DNA Hydroxyl Methylation
Genomic studies have uncovered a link between a novel epige-
netic modification and cancer. In 2009, biochemical studies
identified a new type of DNA modification: the conversion of
5-methylcytosine (5mC) at CpG islands to a hydroxylated
variant called 5-hydroxymethylcytosine (5hmC) by the ten/
eleven translocation (TET) family of DNA hydroxylases (Kriaucio-
nis and Heintz, 2009; Tahiliani et al., 2009) (Figure 2). Soon
thereafter, genomic surveys found that a family member TET2
shows recurrent inactivating mutations in AML, MDS, and other
myeloproliferative disorders (Delhommeau et al., 2009; Lange-
meijer et al., 2009). As noted above, the TET enzymes require
a-ketoglutarate for their activity and are inhibited by the 2HG on-
cometabolite product of mutant IDH1/2. TET2 and IDH1/2 muta-
tions thus act, at least in part, through a common mechanism;
as would be expected, these mutations rarely co-occur in
AML. Interestingly, however, TET2 and DNMT3A mutations
frequently co-occur in MDS, pointing to an as-yet unexplained
cooperativity between dysregulation of 5mC and 5hmC in leuke-
Complementing the targets above affecting RNA transcription,
cancer sequencing uncovered other important targets involved
in RNA splicing (Figure 3). Though it had long been known that
cancers showed aberrant splicing patterns, it was impossible
to know whether these events played a causal role in cancer or
were simply an effect of cancer.
The answer became clear with exome-sequencing studies in
chronic myelogenous leukemia (CLL) and myelodysplastic
syndromes (MDS). In CLL, the spliceosome gene SF3B1 is
mutated in 10%–15% of cases, and other spliceosomal genes,
such as SFRS1, SFRS7, and U2AF2, are also mutated at lower
frequencies (Puente et al., 2011; Quesada et al., 2012; Wang
et al., 2011a). In MDS, the spectrum is even more striking:
45%–85% of cases harbor mutations in a spliceosome gene,
with SF3B1 and U2AF1 being the most common and other
24 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
genes (such as SF3A1, ZRSR2, SRSF2, and U2AF2) occurring
at lower frequencies (Papaemmanuil et al., 2011; Yoshida et al.,
2011). Spliceosomal genes have also been found significantly
mutated in solid tumors—most notably, U2AF1, SF3B1,
U2AF2, and PRPF40B mutations in lung adenocarcinomas
(Imielinski et al., 2012). SF3B1 is also recurrently mutated in
breast cancer (Ellis et al., 2012) and pancreatic cancer (Biankin
et al., 2012).
The pattern of mutations in the spliceosomal genes contains
important clues about their function. First, the mutations tend
to occur in a mutually exclusive fashion in all tumor types exam-
ined, suggesting that they play similar roles and are thus func-
tionally redundant with respect to causing cancer (for a recent
review, see Lindsley and Ebert ). Second, several of the
genes carry heterozygous missense mutations affecting specific
protein domains, suggesting that they confer a gain of function.
SF3B1 (encoding a member of the splicing factor 3b complex,
which interacts with SF3A proteins and a snRNA species to
form the U2 small nuclear ribonucleoprotein [snRNP]) has muta-
tions affecting the carboxy-terminal HEAT domains. U2AF1
(encoding a member of the U2 snRNP auxiliary factor, a spliceo-
somal component that binds the 30splice acceptor site within
target pre-mRNAs) has mutations affecting conserved zinc
finger domains. SRSF2 (a serine-arginine-rich protein that medi-
ates U2 snRNP assembly through binding of exon-splicing
enhancer elements within pre-mRNA species) also has distinct
codon localizations (Yoshida et al., 2011). In contrast, the
ZRSR2 gene (encoding a spliceosomal adaptor protein) has
mutations distributed throughout its open reading frame and
has frequent nonsense mutations; the pattern is indicative of
loss-of-function mutations. What is missing, of course, is knowl-
edge of the specific aberrant cancer-related splicing events
caused by these mutations.
Genotype-phenotype connections offer some additional
clues. In MDS, SF3B1 mutations occur primarily in subtypes
associated with ring sideroblasts (Papaemmanuil et al., 2011;
Yoshida et al., 2011), whose presence signifies defective eryth-
rocyte maturation. This observation raises the possibility that
mutated SF3B1 may cause ring sideroblast formation, at least
in some MDS subtypes, by governing splicing of a key erythroid
lineage differentiation factor.
Mutations in several splicing factors carry prognostic informa-
tion that might influence clinical management. For example,
U2AF1 mutations have been linked to increased progression
from MDS to AML, and SRSF2 mutations correlate with the so-
called chronic myelomonocytic leukemia (CMML) subtype of
MDS. In CLL, SF3B1 mutations correlate with more rapid
disease progression and lower overall survival (Quesada et al.,
2012; Wang et al., 2011a). U2AF1 mutations were associated
with poor progression-free survival in lung adenocarcinoma
(Imielinski et al., 2012).
Splicing factors were not previously considered attractive
(Figure 3). Indeed, several small molecules and natural products
known to target the spliceosome have been reported, including
spliceostatin A (SSA), a metabolite derived from Pseudomonas
that inhibits the SF3b complex and suppresses splicing
invitro, and pladienolide,
Streptomyces platensis that inhibits the SF3B1 protein directly
(Kaida et al., 2007). A derivative of pladienolide called E7107
has entered phase I clinical trials and shows moderate activity
in thyroid cancer (Folco et al., 2011).
Genome-wideandexome-wide sequencinginmultiple myeloma
(MM) suggested an unexpected (and still unexplained) role of
proteinsynthesisand degradation. InMM,mutationswere found
2011a). DIS3 is an RNA exonuclease that regulates RNA abun-
dance through the exosome complex. FAM46C is a protein
whose function remains unknown but whose expression pattern
is nearly perfectly correlated with that of genes encoding ribo-
gation factors (Chapman et al., 2011a). XBP1 encodes a factor
involved in the unfolded protein response; mutations in the
mouse homolog cause a myeloma-like condition.
Unbiased genomic studies have also uncovered unexpected
roles for the ubiquitination machinery in cancer. In prostate
and endometrial cancers, mutations in SPOP have been
observed in 8%–14% of cases (Barbieri et al., 2012; Kan et al.,
2010; Le Gallo et al., 2012). SPOP encodes the substrate recog-
nition component of an E3 ubiquitin ligase complex. In prostate
cancer, the SPOP mutations affect highly conserved amino
acid residues situated within the substrate-binding motif
(MATH domain), suggesting that they abrogate normal ligase/
substrate interactions. These mutations are mutually exclusive
with ETS rearrangements, thereby defining a distinct genetic
subtype of prostate cancer (Barbieri et al., 2012). In endometrial
cancer, SPOP mutations also occur in the MATH domain but
involve different amino acid residues than those seen in prostate
cancer (Le Gallo et al., 2012). The distinct pattern of mutations in
these two cancers suggests loss of recognition for distinct
substrates, leading to their accumulation.
endometrial, head/neck, bladder, and GI cancers but only rarely
shows recurrent mutations in prostate cancer. In contrast to
SPOP, the mutations appear to be simple loss-of-function
events. The ubiquitin ligase gene WWP1 thus far only shows
recurrent mutations in liver cancer (Fujimoto et al., 2012). The
a compoundproduced by
Figure 3. Cancer-Associated Mutations in
the RNA-Splicing Machinery
Genes encoding spliceosomal components are
recurrently mutated in both hematologic malig-
nancies and solid tumors. Drugs that target SF3B1
have entered clinical trials.
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 25
distinct spectra of cancers seem likely to result from insufficient
degradation of different proteins that are critical for different cell
types. Finding the protein targets is a high priority.
Exome-sequencing studies in head and neck squamous cell
carcinoma (HNSCC) revealed unexpected roles for pathways
involved in squamous cell differentiation (Agrawal et al., 2011;
Stransky et al., 2011). The studies found mutations in NOTCH1
in ?15% of cases, as well as mutations and focal copy number
alterations of NOTCH2 and NOTCH3 in an additional 11%
(Stransky et al., 2011) (Figure 4). Whereas activating NOTCH1/
2 mutations had been reported in various blood cancers (Lohr
et al., 2012; Pasqualucci et al., 2011b; Puente et al., 2011;
Weng et al., 2004), the NOTCH mutations in HNSCC were
clearly loss-of-function events. Parallel studies in myeloid
tions (Klinakis et al., 2011).
The NOTCH mutations turned out to be just a part of the story.
A more sophisticated analysis (of gene sets with recurrent muta-
tions) pointed to genes known to be involved in epidermal devel-
opment and squamous differentiation in HNSCC (Stransky et al.,
2011) (Figure 4). Additional genes mutated in HNSCC (such as
RIPK2, EZH2, and DICER1) were linked to the squamous differ-
entiation program based on results from genetically engineered
mice. Two further genes (SYNE1 and SYNE2, mutated in 20%
and 8% of cases, respectively) were also implicated; these
genes encode proteins that control nuclear polarity and spindle
orientation, which stand upstream of NOTCH signaling in squa-
mous lineage development (Williams et al., 2011). In all, nearly
one-third of HNSCC tumors appeared to harbor at least one
mutation predicted to affect squamous differentiation.
Comprehensive genomic studies soon demonstrated the
importance of dysregulated squamous differentiation in other
tumor types. For example, inactivating NOTCH1/2 mutations
occur in >75% of cutaneous squamous cell carcinomas (Wang
et al., 2011b). Moreover, a study of squamous lung cancer
revealed that 44% of cases harbored mutations in genes that
regulate squamous differentiation (Hammerman et al., 2012).
The loss of function in squamous differentiation contrasts with
the SOX2 lineage survival TF oncogene, which undergoes
frequent amplification in squamous lung cancer, HNSCC,
and cervical squamous cancers (http://www.cbioportal.org/
Connecting the Dots: From Cancer Genes to Cancer
Hanahan and Weinberg have proposed ‘‘hallmark’’ processes
(Hanahan and Weinberg, 2000, 2011). These processes include
genome instability, unlimited cell division, sustained proliferative
signaling, evasion of growth suppression, cellular energetics,
and resisting apoptosis. Many classical cancer genes encode
proteins that mediate or control such processes: for example,
mutations in receptor tyrosine kinases or cell-cycle inhibitors
can be directly understood in terms of ‘‘jamming the accelerator
pedal’’ or ‘‘eliminating the brakes’’ on cell growth.
By contrast, many of the newly discovered cancer genes
affect global processes whose precise connection to cancer
remains obscure. These cancer genes act by deranging gene
expression (through changes to chromatin and DNA methyla-
tion), RNA splicing, protein synthesis and degradation, and
cellular metabolism (Figure 5). Presumably, these global
changes propel cancer by affecting one or more specific targets
involved in cancer processes—activating or repressing specific
Figure 4. Genetic Alterations Affecting Lineage Specification Are
Common in Squamous Tumors
NOTCH and several other lineage regulatory factors are disrupted by genomic
alterations in lung, cervical, head/neck, and cutaneous squamous carci-
nomas. SOX2 is also a lineage survival oncogene that regulates squamous
maturation. Genes that encode proteins shaded in red undergo mutational
activation or amplification; those shaded gray undergo mutational inactivation
Figure 5. Genetic Alterations Disrupt Multiple Cellular Processes
Alterations in a range of cellular processes presumably contribute to cancer
through their action on one or more target genes, mRNAs, or proteins,
although the precise targets remain unknown in many cases (illustrated by
shaded ovals). Even in advance of such knowledge, many mutations suggest
potential targets for therapeutic development and allow stratification for
clinical trials of targeted drugs.
26 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
genes, altering the isoforms of specific mRNAs, and increasing
or decreasing steady-state levels of specific proteins. The key
targets are likely cell type specific, accounting for the presence
of specific subsets of driver genes in particular cancer types.
For the most part, we are ignorant of the precise targets—or
whether we are looking for single targets or multiple targets.
Indeed, mutations affecting global processes seemingly provide
an efficient mechanism by which multiple coregulated targets
might be affected. In some respects, the situation may be
analogous to amplification and deletion of chromosome arms,
which may provide a similarly efficient means to dysregulate
multiple targets. In each case, identifying the full range of target
genes will likely require unbiased genomic surveys at the DNA,
RNA, and protein levels to generate hypotheses, as well as
focused experiments to prove them.
Connecting the new cancer genes to known (or as-yet
unknown) cancer processes will surely accelerate efforts to
understand and treat cancer. Of course, therapeutic progress
can be made even without a full understanding of their action.
For example, inhibitors of the neomorphic IDH1/2 enzymes or
perturbed splicing factors may prove valuable even without
understanding the full range of enzymes affected by 2HG or
SF3B1. Moreover, the set of cancer genes mutated in a tumor
provides a powerful classification tool, identifying natural
subtypes that can be studied in both preclinical and clinical
investigation to detect distinct vulnerabilities and correlate
Completing the Picture: Long Tails, Dark Matter,
Heterogeneity, and Heredity
Genomic studies have definitely shown that our previous inven-
tory of cancer genes was far from complete. The question now is
do we finally have a near-comprehensive catalog? The honest
answer: we don’t know.
high frequency, but many more cancer-related genes are found
mutated at much lower frequencies. For example, a recent
genomic study of breast cancer reported 40 loci that were
mutated at statistically significant rates (Stephens et al., 2012);
of these, 53% of the apparent driver mutations or focal copy
number alterations were concentrated in six genes (TP53,
PIK3CA, ERBB2, FGFR1/ZNF703, and GATA3), and the
remainder were dispersed across 34 genes. Only eight of the
genes were mutated in at least 10% of breast cancers. Many
tumor types exhibit similar ‘‘long tail’’ distributions.
Some of the genes found mutated at low frequencies in some
cancers are more commonly (and significantly) mutated in other
cancers. In the breast cancer example mentioned above, ‘‘long
tail’’genes thataresignificantly mutatedin other cancers include
the SWI/SNF complex genes ARID1A and ARID1B, the
KMT-encoding genes MLL2 and MLL3, and KRAS. This finding
might suggest that the discovery of new driver genes is ap-
proaching a plateau. On the other hand, the fact that so many
driver genes occur at lower frequencies raises the possibility
that many such genes may yet remain undiscovered. Moreover,
some tumors (e.g., some primary prostate cancers) appear to
lack even a single mutation in a proven driver gene.
The problem is due, in part, to the fact that most studies to
date have been insufficiently powered—lacking adequate
sample size to detect low-frequency events and/or adequate
depth of sequence coverage to overcome impurity due to
stromal contamination. Fortunately, it should be feasible to
enumerate all genes carrying nonsynonymous coding mutations
in at least 2% of tumors of every cancer type by sequencing
a sufficiently large number of tumor-normal pairs. (Roughly 950
pairs will be needed per tumor type if the background mutation
rate in the cancer is 2 mutations per Mb and 2,500 pairs if the
rate is 10/Mb.) This scale seems readily achievable for many
tumor types over the next several years.
In contrast to point mutations in coding regions, our ability to
discover and understand other types of driver mutations is still
distressingly limited. Many more important cancer drivers may
be lurking in the places that we cannot currently interpret. These
include copy number alterations, chromosomal rearrangements,
and noncoding regions.
As noted above, gains and losses spanning whole chromo-
some arms occur commonly in most types of cancer, but it is
difficult to pinpoint the key genes for which the presence of
a few extra copies contributes to cancer. Even for focal amplifi-
cations or deletions, finding the target genes can be difficult. A
study of copy number alterations across cancer types found
that proven cancer genes were known for less than half of recur-
rent focal amplifications and aneven smaller proportion of recur-
rent focal deletions (Beroukhim et al., 2010). Incorporating
sample-matched data sets can help to suggest candidates for
functional validation. For example, a study in glioblastoma
showed that gain of extra copies of chromosome 7 was associ-
ated with dysregulation of the HGF-MET axis (Beroukhim et al.,
2007); pharmacologic experiments showed that cell lines
carrying nonfocal chromosome 7 gains together with HGF and
MET overexpression were preferentially dependent on MET
Chromosomal rearrangements are also pervasive in many
cancers, but our ability to characterize and interpret their impact
has been limited. Whereas basic cancer genome analyses can
be accomplished by mapping short DNA sequences to a fixed
reference sequence, comprehensive study of rearrangements
requires obtaining larger-scale ‘‘linking’’ information to recon-
struct unexpected genomic junctions and performing transcrip-
tome sequencing to detect expressed fusion genes. These
efforts have been aided by recent computational advances,
such as algorithms that reconstruct transcriptomes without the
need for an underlying reference genome (Grabherr et al.,
2011). Most rearrangements may be random passenger events,
but some clearly disrupt cancer genes by creating fusion
proteins or by subjecting a gene to new regulation. Genome
analysis has identified several new fusions involving known
cancer genes, including RAS, RAF, ERG, and PTEN, in prostate
cancer and in other malignancies (L.A.G. and E.S.L, unpublished
data; Palanisamy et al., 2010; Wang et al., 2011c) and NOTCH
genes in breast cancers (Robinson et al., 2011). Although rela-
tively few instances of recurrent rearrangements implicating
new cancer genes have emerged (possibly owing to limited
sample sizes and the challenge of interpreting these events),
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 27
those that have been discovered may implicate new biological
processes. Examples include MAST kinases in breast cancer
(Robinson et al., 2011) and R-spondin family members in 10%
of colon cancers (Seshagiri et al., 2012).
The great uncharted frontier is the >98% of the human
genome that does not encode proteins. Our ignorance is due
to two factors. First, we have lacked adequate data because
cancer genome studies to date have largely focused on the
exome rather than on the whole genome for reasons of cost.
Second, we lack adequate analytical techniques to recognize
recurrent mutations in nongenic territory. To detect a cancer-
associated target, one must aggregate mutations across
a defined region to test whether the rate is sufficiently elevated
above background. This is straightforward for protein-coding
regions, where one can aggregate nonsynonymous mutations
across thousands of bases. But it is more challenging for the
of small regions across the genome to find those with an unusu-
ally high mutation rate. If one searches with a small window, the
mutational signal will be weak (unless the mutation frequency is
noise in the surroundings. At present, the best approach may be
to focus on regions defined by features corresponding to known
biological functions, such as promoters, evolutionary conserva-
tion, and epigenomic modification.
A recent study of regulatory regions in melanoma has
confirmed that important mutations may be lurking in noncoding
regions (Huang et al., 2013). Whole-genome sequencing re-
vealed the presence of highly recurrent somatic mutations at
two specific nucleotides situated within the promoter of the
TERT gene, which encodes a reverse transcriptase component
of the telomerase enzyme. Both of these mutations are cyti-
dine-to-thymidine transitions that generate a de novo binding
site for the ETStranscription factor. Thesesitesincrease expres-
sion from the TERT promoter in reporter assays. The mutations
occur in >70% of melanomas and ?16% of other tumor types
examined, including bladder and hepatocellular carcinomas.
Cancer genome analyses have largely focused on tumors as
a whole. Yet it has been clear for decades that tumors show
extensive cellular and molecular heterogeneity. Indeed, hetero-
geneity was inherent in Nowell’s original clonal model for tumor
evolution (Nowell, 1976). Some early genomic studies have
begun to come to grips with tumor heterogeneity. Initial forays
have documented subclonal
geographic regions of a primary tumor (Gerlinger et al., 2012)
and within hematopoietic malignant populations (Ding et al.,
Studies of heterogeneity are beginning to provide fascinating
glimpses into paths of tumor evolution. For example, a study of
21 breast cancers showed that the most recent common
ancestral tumor cell—which contains the full complement of
mutations common to all tumor cells—arose remarkably early
in ‘‘molecular time’’ (Nik-Zainal et al., 2012b). The precursor
cell typically gives rise to a dominant subclone that represents
at least 50% of all cells in the primary tumor. Knowledge of in-
tratumoral heterogeneity has also revealed instances of conver-
gent evolution. In one study of renal cancer, only 30%–35% of
somatic mutations were concordant across multiple primary
and metastatic sites sampled (Gerlinger et al., 2012); however,
several cancer genes contained distinct genomic alterations
that had arisen in geographically disparate regions of the
primary tumor. This observation thus revealed a remarkable
mutational consolidation that engaged critical pathways linked
to chromatin regulation (SETD2, KDM5C) or signal transduction
Tumor heterogeneity could have important implications for
‘‘precision’’ cancer medicine. Some subclones may contain
pre-existing mutations that confer drug resistance or accelerate
tumor relapse in cancers that show poor clinical responses to
targeted inhibitors. Studies that seek to stratify patients for clin-
ical trials of targeted agents based on specific ‘‘actionable’’
mutations may be confounded if a biopsy sample is not repre-
sentative of the whole tumor. On the other hand, the ability to
identify driver or resistance mutations within subclonal popula-
tions may allow improved prediction of clinical outcomes
(Landau et al., 2013). The growing understanding of intratumoral
heterogeneity may inform the design of clinical studies that
account for this process (e.g., by following the therapeutic
response of the biopsied lesion in addition to the overall tumor
burden) and circumvent its subversive effects (e.g., by devel-
oping therapeutic combinations directed against major and
Studies of cancer heterogeneity will be accelerated by recent
genomic advances enabling single-cell sequencing (Navin et al.,
2011). Whole-exome sequences have been produced from
single cells in both hematologic neoplasms and solid tumors
(Hou et al., 2012; Xu et al., 2012). Moreover, new protocols
that yield more uniform and accurate whole-genome amplifica-
tion have been developed (Zong et al., 2012). Single-cell anal-
yses have already provided new insights into the evolutionary
history of tumors within individual patients and have revealed
functional differences across individual tumor cells (Kreso
et al., 2013). In the future, these advances may enable detailed
genomic studies of circulating tumor cells, thereby providing
high-resolution monitoringof therapeuticresponses or emerging
resistance mechanisms and facilitating detection of aggressive
Although many of the genetic factors that drive a cancer are
acquired through somatic mutation, some are inherited at birth.
Epidemiological studies have long noted an increased risk of
cancer in relatives of affected individuals (Pomerantz and
Freedman, 2011). Genomics has revealed many genesthat influ-
ence predisposition to cancer, although the picture remains far
but we briefly summarize the current state of progress for in-
herited variation (see recent reviews by Hindorff et al. 
and Chung and Chanock ).
One method to identify genes that confer predisposition to
cancer is to study rare, highly penetrant Mendelian cancer
syndromes. These syndromes arise when mutant alleles confer
such a high increased risk (>10-fold) that it is straightforward
to trace their transmission in families by linkage analysis. More
than 100 genes underlying such cancer syndromes have been
28 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
identified, including those underlying retinoblastoma (RB1),
breast cancer (BRCA1, BRCA2), and colon cancer (APC, MU-
TYH and the mismatch repair genes MLH1, MSH1, MSH6, and
PMS2). Such genes have been deeply informative about cancer
biology but together account for <5% of the estimated herita-
bility of cancer (Cazier and Tomlinson, 2010).
To identify cancer genes that confer more modest risks, it is
necessary to use population-based association studies rather
than family-based linkage studies. The methodology for associ-
ation studies depends on whether one wishes to study
‘‘common’’ (>1%) or ‘‘rare’’ (<1%) variants. Common variants
are frequent enough that they can be tested for their individual
effects on cancer risk by genotyping of millions of variants in
cases and controls in ‘‘genome-wide association studies’’
(GWAS) (Altshuler et al., 2008). Rare variants must be combined
together for analysis: studies examine the aggregate frequency
of rare coding variants in each gene to look for an elevated
frequency in cases versus controls. More than 150 cancer risk
loci have been identified thus far, with most having been found
through GWAS (Chung and Chanock, 2011; Hindorff et al.,
2011). The common alleles appear to include many regulatory
variants and to confer a lower increased risk (<30%), whereas
the rare alleles affect coding regions of known cancer genes
(such as ATM, BRIP1, CHEK2, PALB2, and RAD51C in breast
cancer) and tend to have higher risk (2- to 3-fold). The relative
roles of the two classes vary among cancer types. Importantly,
the risk factors identified to date explain only a fraction of the
heritability of cancer (Hindorff et al., 2011). Genomic studies
with much larger samples willbe needed to obtain a fuller picture
of the inherited basis of cancer risk.
Understanding the mechanisms by which common inherited
genetic variants predispose to cancer will require integrative
genomic analysis, which will likely yield important biological
insights. One instructive case is found in a 500 kb ‘‘gene desert’’
in chromosome 8q24. Whereas most cancer-associated loci are
tumor-type specific, this region contains variants that affect risk
of prostate, colon, esophagus, head/neck, breast, and pancreas
cancers (reviewed in Hindorff et al. ). Epigenetic and chro-
mosome conformation studies in human and genetic engi-
neering studies in mouse suggest that the variants alter distal
regulatory sequences controlling the MYC locus, which lies telo-
meric to the region (Ahmadiyeh et al., 2010; Pomerantz et al.,
2009; Sur et al., 2012; Tuupanen et al., 2009). A similar situation
occurs in a 500 kb region in 9p21, where different variants affect
multiple types of cancer (including breast cancer, melanoma,
glioma, and leukemia) as well as noncancer-related diseases
likely alter regulation of the cell-cycle genes CDKN2A/CDKN2B.
The observation that a number of additional cancer-related loci
also affect diabetes (or insulin dysregulation) suggests an impor-
tant role for metabolic processes in cancer (Dupuis et al., 2010;
Pal et al., 2012).
some disparities among ethnic groups. For example, a propor-
tion of the higher risk of prostate cancer in African Americans
and other men of African descent may be due, in part, to allele
frequency differences at chromosome 8q24 (Haiman et al.,
2011; Murphy et al., 2012).
Applying the Knowledge: Diagnostics and Therapeutics
The ultimate test of cancer genomics will be its ability to improve
diagnostics and therapeutics. Academic centers are already
to guide cancer treatment (Dias-Santagata et al., 2010;
MacConaill et al., 2009; Thomas et al., 2007; Wagle et al.,
2012) These platforms involve testing a few hundred specific
cancer-associated mutations or performing full sequencing of
a limited set of cancer-associated genes (Lipson et al., 2012;
Wagle et al., 2012). The early returns suggest that, in ?40%–
60% of cases for many common solid tumors, the information
points to at least one alteration that might influence therapeutic
decision-making or might suggest enrollment in a particular clin-
ical trial (Beltran et al., 2012; Hammerman et al., 2012; CGAN,
2012). As sequencing costs fall, diagnostics may move to
whole-exome or whole-genome sequencing. The challenge will
be to filter and annotate the results for oncologists, based on
a constantly changing landscape of scientific knowledge. Even-
tually, genomic analysis will likely become part of the standard of
care for cancer patients.
Cancer genomics will also become a key component in the
design, execution, and interpretation of clinical trials. Investiga-
tors arealready using genomic information forretrospectiveclin-
ical analyses that correlate treatment response with specific
genomic features. There is growing interest in using deep
genomic characterization of ‘‘exceptional cases,’’ such as rare
tumors that show a complete clinical response to a particular
anticancer regimen. For example, a recent tumor genome-
sequencing study of a bladder cancer patient who experienced
a complete response to a TOR inhibitor (everolimus) identified
two distinct cancer gene mutations (TSC1 and NF2) predicted
to affect oncogenic TOR signaling (Iyer et al., 2012). Sequencing
of additional everolimus-treated tumors confirmed that TSC1
mutations correlate with clinical response.
The prospective use of genomic information may substantially
transform trial design. Cancer trials have traditionally selected
patients based on histologic tumor subtypes. However, it makes
more sense to test targeted therapeutics on the subset of
patients carrying the relevant genetic lesions; by selecting the
patients most likely to benefit, one decreases sample size,
cost, and unjustified harm. In some cases, it will make sense to
enroll patients carrying the same genetic alteration across
a wide range of tumor types (for example, a trial led by investiga-
tors at Memorial Sloan-Kettering Cancer Center in which
BRAFV600mutant tumors from colon, thyroid, lung, and other
organ sites are treated with a selective RAF or MEK inhibitor).
Moreover, novel designs are becoming possible in which one
simultaneously tests multiple drugs or drug combinations. In
these ‘‘basket trials,’’ patients are assigned to different thera-
peutic regimens based on the specific genetic profiles in their
tumor (Kim et al., 2011). Basket trials may employ an ‘‘adaptive
design,’’ allowing ‘‘real-time’’ adjustments if hints of specific
genotype-driven responses are detected (Berry, 2011). Trials
can further be shaped by genomic analysis from serial biopsies
to assess pharmacodynamics response and to characterize
the presence of resistance mechanisms. It may be useful to
create a worldwide ‘‘clearinghouse’’ mechanism that connects
patients to trials based on their genotype, especially to obtain
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 29
a large enough sample to evaluate responses in tumors with
rarer genetic features.
Cancer drug discovery efforts are already being shaped by the
findings fromgenomic studies.Insomecases,the productofthe
mutated gene may be an appropriate drug target. In many other
cases, mutations may confer specific vulnerabilities on the
cancer cell that can be discovered through functional genomic
studies, such as comprehensive gene inhibition screens with
RNA interference across large numbers of cancer cell lines
with varying genotypes.
Beyond the development of specific drugs, knowledge of the
cancer genome will be critical to design combination therapies,
which willbe essential for conqueringcancer.Mosttumors even-
tually develop resistance to single-agent therapeutics (reviewed
in Garraway and Ja ¨nne ). For example, the use of RAF
ular responses, but tumors reappear within a year (Chapman
et al., 2011b; Flaherty et al., 2012a, 2012b; Sosman et al., 2012).
Multiple genetic mechanisms of resistance have been described
cation, activating MEK1 mutations, and NF1 loss), each of which
produces sustained MAP kinase (ERK) activity in the presence of
2010; Poulikakos et al., 2011; Wagle et al., 2011; Whittaker et al.,
2013). These findings raise the possibility that adding an ERK
inhibitor to existing RAF/MEK inhibitor regimens could provide
an additional clinical benefit (Whittaker et al., 2013).
Systematic preclinical studies may make it possible to antici-
pate the mechanisms of resistance, allowing therapeutic scien-
tists to plan for resistance long before it arises in the clinic. For
example, a recent study performed large-scale screens (using
RNAi knockdown and ORF overexpression) to identify genes
whose loss or amplification can confer resistance to RAF inhibi-
taker et al., 2013). Another group systematically screened
stromal cell lines to identify those that secrete factors that confer
resistance on adjacent cancer cells; the screen revealed that
hepatocyte growth factor confers resistance to RAF inhibition
(Straussman et al., 2012). Such approaches may make it
possible to formulate rational combination therapies even before
the results of single-agent clinical trials are known.
In the end, combination therapy depends on shifting the odds
of resistance. There is cause for optimism: mathematical
modeling suggests that resistance may often be due to pre-ex-
isting mutations in the tumor cell population (Michor et al.,
2005). If so, it should be possible to prevent recurrence by treat-
ing simultaneously with drugs directed against several indepen-
dent targets so that the chance of a single cell carrying all the
necessary resistance mutations is vanishingly small. This is, of
course, the basis for the successful triple-drug combinations
against HIV. Ultimately, cancer genomics should aim to provide
a comprehensive roadmap for selecting rational, multidrug
combinations for anticancer therapy.
Next Steps for Cancer Genomics
The early fruits of cancer genome studies have confirmed
Renato Dulbecco’s prediction about the value of complement-
ing ‘‘piecemeal approaches’’ with systematic genome-wide
studies. The results have already opened new frontiers in basic,
translational, and clinical investigation. Still, current studies
have only scratched the surface of what can be learned from
comprehensive study of the cancer genome. Cancer genomics
has largely focused on documenting the mutations in primary
tumors. Over the coming years, the field should expand
its focus to gather systematic information to inform a wider
range of biological and clinical questions. Below, we suggest
four important components for the next phase of cancer
Complete the Mutational Atlas of Primary Tumors
A straightforward but critical component is to finish compiling
the catalog of significantly mutated genes in primary tumors of
every feasible cancer type. Given the long tail of cancer genes
and the variable background mutation rates, such studies will
require thousands of tumor-normal pairs. Why bother to press
for completeness? Scientifically, because the low-frequency
drivers may in aggregate make a substantial contribution and
because they are likely to harbor further surprises. Medically,
because physicians will want to be able to recognize all driver
mutations in each patient to optimize therapy. Fortunately, these
efforts should become increasingly feasible and affordable given
the decreasing costs of sequencing and the increasing ability to
analyze small amounts of starting material from formalin-fixed,
paraffin-embedded archival samples. The analysis must expand
beyond the exome to include the whole genome (including long-
range links to detect translocations), the transcriptome, and the
epigenome (at least the methylome and key chromatin modifica-
tions). Improved laboratory and analytical methods will be
needed to discern the targets of nonfocal chromosome copy
number aberrations, epigenomic modifications, and nongenic
translocations. In addition, the genomic information should be
thoroughly mined to identify germline variants that contribute
to cancer risk.
Expand the Mutational Atlas beyond Primary Tumors
The second component is to systematically expand the atlas
beyond primary tumors to include the natural history of human
cancer, aswell as the homology to cancer in key model systems.
A mutational atlas of the natural history of cancer would involve
comprehensive genomic analysis of preneoplastic lesions,
metastases from various organ sites, and tumors that show
different types of responses to therapies, including extreme
response, intrinsic resistance, and acquired resistance. Ideally,
all clinical trials in oncology would be subject to such analysis.
Genomic characterization should also be applied to animal
models of cancer so that we can better connect these to human
cancers based on mechanism. In addition to genetically engi-
neered mouse models, intensive studies of naturally occurring
cancers in large animals, especially dogs (Karlsson and
Lindblad-Toh, 2008), may provide both insights and important
preclinical models for drug testing.
Create a Functional Encyclopedia of Altered Pathways
and Acquired Vulnerabilities
Though a mutational catalog will provide a comprehensive
picture of cancer genomes, this catalog alone is not enough.
We need to produce a functional encyclopedia of altered cellular
pathways and acquired vulnerabilities that correspond to each
30 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
cancer genome. Genomic approaches can propel systematic
functional studies, just as they have propelled comprehensive
structural studies. Building a functional encyclopedia will involve
(1) creating tractable models representing the full range of
cancer genotypes and (2) characterizing these models with
respect to their genomic alterations, essential pathways, and
therapeutic vulnerabilities. Already, ongoing projects are assem-
bling large collections of cancer cell lines; defining their genomic
changes; characterizing their cellular states at the RNA, protein,
andposttranslational levels; and determining their sensitivities to
anticancer drugs, RNAi-based inhibition of every gene, and
microenvironmental interactions (Barretina et al., 2012; Garnett
et al., 2012). With a sufficiently large collection of cell models,
one can correlate pathways and vulnerabilities with specific
markers for patient selection in clinical trials, and potential new
targets for cancer drug development.
One limitation has been thatcurrent cancer cell lines represent
a biased sampling of cancer and cancer genotypes, owing to
differences in the ability to derive cell lines. However, new
methods (such as Rho kinase inhibitor-treated feeder layers
and ‘‘organoid’’ culture systems) appear poised to greatly
expand the repertoire of available cancer models (Huch et al.,
2013; Liu et al., 2012). Patient-derived xenografts can also play
a key role in preclinical studies of new therapeutics.
Enable and Promote Sharing of Cancer Genomic
Finally, there is one critical component that is an essential foun-
dation for the others: widespread information sharing. Cancer
genome information will grow exponentially in the years ahead
as genome analysis moves from the research lab to routine clin-
ical care for millions of patients around the world. If it were
possible to share and analyze thistorrent of genomic information
together with associated clinical outcome data, it could signifi-
cantly accelerate the understanding and treatment of cancer.
Theinformation wouldspeednot onlythe identification ofcancer
genes, but also the correlation of therapeutic responses to spe-
cific tumor genotype, including dramatic responses to new tar-
to different regimens. In effect, it would connect cancer care
around the world into a laboratory for continuous improvement.
Making this world a reality will require coordinated efforts by
researchers, hospitals, and patient groups to accomplish two
goals: (1) creating the computational infrastructure to enable
sharing and (2) promoting a culture of sharing. It is easy to
imagine an alternative future in which cancer genomic informa-
tion cannot be aggregated because it is stored in inaccessible
sites and incompatible formats, much as is the case with elec-
tronic medical records in the U.S. To avoid this outcome, it will
be necessary to have common or interoperable standards for
data and analysis, cloud-based storage solutions to ensure
data security, and rigorous systems to enforce patients’ instruc-
tions concerning their data. But technology platforms alone will
not suffice. Clinicians, hospitals, and healthcare networks will
need to become engaged in collecting and sharing clinical
outcome data. Pharmaceutical companies and others will need
to share data from completed clinical trials. Ultimately, patient
advocacy groups may provide the impetus for cultural change,
as happened with AIDS. Though it must be up to each patient
to decide whether to share his or her data, we suspect that
most cancer patients will actively want to allow their information
to be appropriately aggregated and shared (with appropriate
rules and technology to protect privacy) to accelerate progress
for this and future generations of patients. We must ensure
that patients have the right and abilityto contribute their informa-
tion to a global fight against cancer.
Genomics has become a powerful tool for cancer research,
yielding important biological surprises and enabling systematic
classification based on cellular mechanism. Cancer genomics
is just now emerging from its first phase, which has been largely
To fulfill its full promise, the field will need to deepen the struc-
tural characterization of cancer genomes, complement it with
comprehensive functional characterization of cancer cells, and
enable and promote information sharing across the world. Ulti-
mately, cancer genomics is about fully knowing the enemy.
While not alone a guarantee of victory, it is an essential part of
any overall plan of attack.
We thank Todd Golub, Matthew Meyerson, and Michael Stratton for reviewing
the manuscript and for valuable comments. We thank Gad Getz, Charles
Roberts, and Matthew Freedman for helpful discussions. The design of figures
was aided by materials from ScienceSlides (http://www.visiscience.com).
L.A.G. and E.S.L. are supported by the U.S. National Human Genome
Research Institute, the U.S. National Cancer Institute, the Slim Initiative for
Genomic Medicine, and the Broad Institute. L.A.G. also receives support
from the Department of Defense, the Starr Cancer Consortium, the Prostate
Cancer Foundation, the Melanoma Research Alliance, and the Adelson
Medical Research Foundation. L.A.G. and E.S.L. are consultants and equity
holders in Foundation Medicine, Inc.
Agrawal, N., Frederick, M.J., Pickering, C.R., Bettegowda, C., Chang, K., Li,
R.J., Fakhry, C., Xie, T.X., Zhang, J., Wang, J.,et al. (2011). Exome sequencing
of head and neck squamous cell carcinoma reveals inactivating mutations in
NOTCH1. Science 333, 1154–1157.
Ahmadiyeh, N., Pomerantz, M.M.,Grisanzio, C.,Herman,P.,Jia,L.,Almendro,
V., He, H.H., Brown, M., Liu, X.S., Davis, M., et al. (2010). 8q24 prostate,
breast, and colon cancer risk loci show tissue-specific long-range interaction
with MYC. Proc. Natl. Acad. Sci. USA 107, 9742–9746.
Altshuler, D., Daly, M.J., and Lander, E.S. (2008). Genetic mapping in human
disease. Science 322, 881–888.
Baca, S.C., Prandi, D., Lawrence, M.S., Mosquera, J.M., Romanel, A., Drier,
Y., Park, K., Kitabayashi, N., MacDonald, T.Y., Ghandi, M., et al. (2013).
Punctuated evolution of prostate cancer genomes. Cell 153. Published online
April 25, 2013. http://dx.doi.org/10.1016/j.cell.2013.03.021.
Banerji, S., Cibulskis, K., Rangel-Escareno, C., Brown, K.K., Carter, S.L.,
Frederick, A.M., Lawrence, M.S., Sivachenko, A.Y., Sougnez, C., Zou, L.,
et al. (2012). Sequence analysis of mutations and translocations across breast
cancer subtypes. Nature 486, 405–409.
Barbieri, C.E., Baca, S.C., Lawrence, M.S., Demichelis, F., Blattner, M.,
Theurillat, J.P., White, T.A., Stojanov, P., Van Allen, E., Stransky, N., et al.
(2012). Exome sequencing identifies recurrent SPOP, FOXA1 and MED12
mutations in prostate cancer. Nat. Genet. 44, 685–689.
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 31
Barretina, J., Caponigro, G., Stransky, N., Venkatesan,K., Margolin,A.A., Kim,
S., Wilson, C.J., Leha ´r, J., Kryukov, G.V., Sonkin, D., et al. (2012). The Cancer
Cell Line Encyclopedia enables predictive modelling of anticancer drug sensi-
tivity. Nature 483, 603–607.
Bass,A.J., Watanabe,H.,Mermel,C.H., Yu,S.,Perner,S.,Verhaak,R.G.,Kim,
S.Y., Wardwell, L., Tamayo, P., Gat-Viks, I., et al. (2009). SOX2 is an amplified
lineage-survival oncogene in lung and esophageal squamous cell carcinomas.
Nat. Genet. 41, 1238–1242.
Beltran, H., Yelensky, R., Frampton, G.M., Park, K., Downing, S.R., Macdon-
ald, T.Y., Jarosz, M., Lipson, D., Tagawa, S.T., Nanus, D.M., et al. (2012). Tar-
geted next-generation sequencing of advanced prostate cancer identifies
potential therapeutic targets and disease heterogeneity. Eur. Urol. Published
online September 5, 2012. http://dx.doi.org/10.1016/j.eururo.2012.08.053.
Bentley, D.R., Balasubramanian, S., Swerdlow, H.P., Smith, G.P., Milton, J.,
Brown, C.G., Hall, K.P., Evers, D.J., Barnes, C.L., Bignell, H.R., et al. (2008).
Accuratewholehumangenome sequencing usingreversible terminator chem-
istry. Nature 456, 53–59.
Berger, M.F., Lawrence, M.S., Demichelis, F., Drier, Y., Cibulskis, K.,
Sivachenko, A.Y., Sboner, A., Esgueva, R., Pflueger, D., Sougnez, C., et al.
(2011). The genomic complexity of primary human prostate cancer. Nature
Berger, M.F., Hodis, E., Heffernan, T.P., Deribe, Y.L., Lawrence, M.S.,
Protopopov, A., Ivanova, E., Watson, I.R., Nickerson, E., Ghosh, P., et al.
(2012). Melanoma genome sequencing reveals frequent PREX2 mutations.
Nature 485, 502–506.
Bernt, K.M., Zhu, N., Sinha, A.U., Vempati, S., Faber, J., Krivtsov, A.V., Feng,
Z., Punt, N., Daigle, A., Bullinger, L., et al. (2011). MLL-rearranged leukemia is
dependent on aberrant H3K79 methylation by DOT1L. Cancer Cell 20, 66–78.
Beroukhim, R., Getz, G., Nghiemphu, L., Barretina, J., Hsueh, T., Linhart, D.,
Vivanco, I., Lee, J.C., Huang, J.H., Alexander, S., et al. (2007). Assessing the
significance of chromosomal aberrations in cancer: methodology and applica-
tion to glioma. Proc. Natl. Acad. Sci. USA 104, 20007–20012.
Beroukhim, R., Mermel, C.H., Porter, D., Wei, G., Raychaudhuri, S., Donovan,
J., Barretina, J.,Boehm, J.S.,Dobson, J.,Urashima, M.,et al.(2010).The land-
scape of somatic copy-number alteration across human cancers. Nature 463,
Berry, D.A. (2011). Adaptive clinical trials in oncology. Nat. Rev. Clin. Oncol. 9,
Biankin, A.V., Waddell, N., Kassahn, K.S., Gingras, M.C., Muthuswamy, L.B.,
Johns, A.L., Miller, D.K., Wilson, P.J., Patch, A.M., Wu, J., et al.; Australian
Pancreatic Cancer Genome Initiative. (2012). Pancreatic cancer genomes
reveal aberrations in axon guidance pathway genes. Nature 491, 399–405.
Boveri, T. (2008). Concerning the origin of malignant tumours by Theodor Bo-
veri. Translated and annotated by Henry Harris. J. Cell Sci. 121(Suppl 1), 1–84.
CGAN (Cancer Genome Atlas Network). (2012). Comprehensive molecular
portraits of human breast tumours. Nature 490, 61–70.
CGARN (Cancer Genome Atlas Research Network). (2008). Comprehensive
genomic characterization defines human glioblastoma genes and core path-
ways. Nature 455, 1061–1068.
Carter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird,
P.W., Onofrio, R.C., Winckler, W., Weir, B.A., et al. (2012). Absolute quantifica-
to identify human cancer predisposition genes. J. Pathol. 220, 255–262.
Chapman, M.A., Lawrence, M.S., Keats, J.J., Cibulskis, K., Sougnez, C.,
Schinzel, A.C., Harview, C.L., Brunet, J.P., Ahmann, G.J., Adli, M., et al.
(2011a). Initial genome sequencing and analysis of multiple myeloma. Nature
Chapman, P.B., Hauschild, A., Robert, C., Haanen, J.B., Ascierto, P., Larkin,
J., Dummer, R., Garbe, C., Testori, A., Maio, M., et al.; BRIM-3 Study Group.
(2011b). Improved survival with vemurafenib in melanoma with BRAF V600E
mutation. N. Engl. J. Med. 364, 2507–2516.
Chen, K., Wallis, J.W., McLellan, M.D., Larson, D.E., Kalicki, J.M., Pohl, C.S.,
McGrath, S.D., Wendl, M.C., Zhang, Q., Locke, D.P., et al. (2009).
BreakDancer: an algorithm for high-resolution mapping of genomic structural
variation. Nat. Methods 6, 677–681.
Chin, K., DeVries, S., Fridlyand, J., Spellman, P.T., Roydasgupta, R., Kuo,
W.L., Lapuk, A., Neve, R.M., Qian, Z., Ryder, T., et al. (2006). Genomic and
transcriptional aberrations linked to breast cancer pathophysiologies. Cancer
Cell 10, 529–541.
Chung, C.C., and Chanock, S.J. (2011). Current status of genome-wide asso-
ciation studies in cancer. Hum. Genet. 130, 59–78.
Cibulskis, K., Lawrence,M.S., Carter, S.L., Sivachenko, A.,Jaffe, D.,Sougnez,
C., Gabriel, S., Meyerson, M., Lander, E.S., and Getz, G. (2013). Sensitive
detection of somatic point mutations in impure and heterogeneous cancer
samples. Nat. Biotechnol. Published online February 20, 2013. http://dx.doi.
Crasta, K., Ganem, N.J., Dagher, R., Lantermann, A.B., Ivanova, E.V., Pan, Y.,
Nezi, L., Protopopov, A., Chowdhury, D., and Pellman, D. (2012). DNA breaks
and chromosome pulverization from errors in mitosis. Nature 482, 53–58.
Dang,L., White,D.W., Gross, S.,Bennett, B.D., Bittinger, M.A., Driggers, E.M.,
IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–744.
Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague,
J., Woffendin, H., Garnett, M.J., Bottomley, W., et al. (2002). Mutations of the
BRAF gene in human cancer. Nature 417, 949–954.
Dawson, M.A., and Kouzarides, T. (2012). Cancer epigenetics: from mecha-
nism to therapy. Cell 150, 12–27.
Dees, N.D., Zhang, Q., Kandoth, C., Wendl, M.C., Schierding, W., Koboldt,
D.C., Mooney, T.B., Callaway, M.B., Dooling, D., Mardis, E.R., et al. (2012).
MuSiC: identifying mutational significance in cancer genomes. Genome Res.
Kosmider, O.,LeCouedic, J.P., Robert, F.,Alberdi, A.,et al.(2009).Mutationin
TET2 in myeloid cancers. N. Engl. J. Med. 360, 2289–2301.
Demetri, G.D., von Mehren, M., Blanke, C.D., Van den Abbeele, A.D.,
Eisenberg, B., Roberts, P.J., Heinrich, M.C., Tuveson, D.A., Singer, S.,
Janicek, M., et al. (2002). Efficacy and safety of imatinib mesylate in advanced
gastrointestinal stromal tumors. N. Engl. J. Med. 347, 472–480.
Dias-Santagata, D., Akhavanfard, S., David, S.S., Vernovsky, K., Kuhlmann,
G., Boisvert, S.L., Stubbs, H., McDermott, U., Settleman, J., Kwak, E.L.,
et al. (2010). Rapid targeted mutational analysis of human tumours: a clinical
platform toguide personalizedcancer medicine.EMBOMol. Med.2,146–158.
Ding, L., Getz, G., Wheeler, D.A., Mardis, E.R., McLellan, M.D., Cibulskis, K.,
Sougnez, C., Greulich, H., Muzny, D.M., Morgan, M.B., et al. (2008). Somatic
mutations affect key pathways in lung adenocarcinoma. Nature 455, 1069–
Ding, L., Ley, T.J., Larson, D.E., Miller, C.A., Koboldt, D.C., Welch, J.S.,
Ritchey, J.K., Young, M.A., Lamprecht, T., McLellan, M.D., et al. (2012). Clonal
evolution in relapsed acute myeloid leukaemia revealed by whole-genome
sequencing. Nature 481, 506–510.
Dolnik, A., Engelmann, J.C., Scharfenberger-Schmeer, M., Mauch, J.,
Kelkenberg-Schade, S., Haldemann, B., Fries, T., Kro ¨nke, J., Ku ¨hn, M.W.,
Paschka, P.,et al.(2012). Commonlyaltered genomic regions inacute myeloid
leukemia are enriched for somatic mutations involved in chromatin remodeling
and splicing. Blood 120, e83–e92.
Druker, B.J., Talpaz, M., Resta, D.J., Peng, B., Buchdunger, E., Ford, J.M.,
Lydon, N.B., Kantarjian, H., Capdeville, R., Ohno-Jones, S., and Sawyers,
C.L. (2001). Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine
kinase in chronic myeloid leukemia. N. Engl. J. Med. 344, 1031–1037.
Dulbecco, R.(1986).Aturning point incancer research: sequencing thehuman
genome. Science 231, 1055–1056.
Dupuis, J., Langenberg, C., Prokopenko, I., Saxena, R., Soranzo, N., Jackson,
A.U., Wheeler, E., Glazer, N.L., Bouatia-Naji, N., Gloyn, A.L., et al.; DIAGRAM
Consortium; GIANT Consortium; Global BPgen Consortium; Anders Hamsten
32 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
on behalf of Procardis Consortium; MAGIC investigators. (2010). New genetic
loci implicated in fasting glucose homeostasis and their impact on type 2 dia-
betes risk. Nat. Genet. 42, 105–116.
Dutt, A., Salvesen, H.B., Chen, T.H., Ramos, A.H., Onofrio, R.C., Hatton, C.,
Nicoletti, R., Winckler, W., Grewal, R., Hanna, M., et al. (2008). Drug-sensitive
FGFR2 mutations in endometrial carcinoma. Proc. Natl. Acad. Sci. USA 105,
Ellis, M.J., Ding, L., Shen, D., Luo, J., Suman, V.J., Wallis, J.W., Van Tine, B.A.,
Hoog, J., Goiffon, R.J., Goldstein, T.C., et al. (2012). Whole-genome analysis
informs breast cancer response to aromatase inhibition. Nature 486, 353–360.
Emery, C.M., Vijayendran, K.G., Zipser, M.C., Sawyer, A.M., Niu, L., Kim, J.J.,
Hatton, C., Chopra, R., Oberholzer, P.A., Karpova, M.B., et al. (2009). MEK1
mutations confer resistance to MEK and B-RAF inhibition. Proc. Natl. Acad.
Sci. USA 106, 20411–20416.
Estey, E. (2007). Acute myeloid leukemia and myelodysplastic syndromes in
older patients. J. Clin. Oncol. 25, 1908–1915.
Figueroa, M.E., Abdel-Wahab, O., Lu, C., Ward, P.S., Patel, J., Shih, A., Li, Y.,
Bhagwat, N., Vasanthakumar, A., Fernandez, H.F., et al. (2010). Leukemic
IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt
TET2 function, and impair hematopoietic differentiation. Cancer Cell 18,
Flaherty, K.T., Infante, J.R., Daud, A., Gonzalez, R., Kefford, R.F., Sosman, J.,
Hamid, O., Schuchter, L., Cebon, J., Ibrahim, N., et al. (2012a). Combined
BRAF and MEK inhibition in melanoma with BRAF V600 mutations. N. Engl.
J. Med. 367, 1694–1703.
Flaherty,K.T., Robert, C.,Hersey,P.,Nathan,P.,Garbe,C., Milhem,M.,Demi-
dov, L.V., Hassel, J.C., Rutkowski, P., Mohr, P., et al.; METRIC Study Group.
(2012b). Improved survival with MEK inhibition in BRAF-mutated melanoma.
N. Engl. J. Med. 367, 107–114.
Folco, E.G., Coil, K.E., and Reed,R. (2011). The anti-tumor drug E7107 reveals
anessentialrolefor SF3binremodeling U2snRNPtoexpose thebranch point-
binding region. Genes Dev. 25, 440–444.
Fujimoto, A., Totoki, Y., Abe, T., Boroevich, K.A., Hosoda, F., Nguyen, H.H.,
Aoki, M., Hosono, N., Kubo, M., Miya, F., et al. (2012). Whole-genome
sequencing of liver cancers identifies etiological influences on mutation
patterns and recurrent mutations in chromatin regulators. Nat. Genet. 44,
Gabor Miklos, G.L. (2005). The human cancer genome project—one more
misstep in the war on cancer. Nat. Biotechnol. 23, 535–537.
Garnett, M.J., Edelman, E.J., Heidorn, S.J., Greenman, C.D., Dastur, A., Lau,
K.W., Greninger, P., Thompson, I.R., Luo, X., Soares, J., et al. (2012). System-
atic identificationof genomic markersof drug sensitivityin cancer cells. Nature
Garraway, L.A., and Ja ¨nne, P.A. (2012). Circumventing cancer drug resistance
in the era of personalized medicine. Cancer Discov. 2, 214–226.
Garraway, L.A., Widlund, H.R., Rubin, M.A., Getz, G., Berger, A.J.,
Ramaswamy, S., Beroukhim, R., Milner, D.A., Granter, S.R., Du, J., et al.
(2005). Integrative genomic analyses identify MITF as a lineage survival onco-
gene amplified in malignant melanoma. Nature 436, 117–122.
Gerlinger, M., Rowan, A.J., Horswell, S., Larkin, J., Endesfelder, D., Gronroos,
E., Martinez, P., Matthews, N., Stewart, A., Tarpey, P., et al. (2012). Intratumor
heterogeneity and branched evolution revealed by multiregion sequencing. N.
Engl. J. Med. 366, 883–892.
Gnirke, A., Melnikov, A., Maguire, J., Rogov, P., LeProust, E.M., Brockman,
W., Fennell, T., Giannoukos, G., Fisher, S., Russ, C., et al. (2009). Solution
hybridselection withultra-long oligonucleotidesformassivelyparalleltargeted
sequencing. Nat. Biotechnol. 27, 182–189.
Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, D.A., Amit, I.,
Adiconis, X., Fan, L., Raychowdhury, R., Zeng, Q., et al. (2011). Full-length
transcriptome assembly from RNA-Seq data without a reference genome.
Nat. Biotechnol. 29, 644–652.
Grasso, C.S., Wu, Y.M., Robinson, D.R., Cao, X., Dhanasekaran, S.M., Khan,
A.P., Quist, M.J., Jing, X., Lonigro, R.J., Brenner, J.C., et al. (2012). The muta-
tional landscape of lethal castration-resistant prostate cancer. Nature 487,
Greenman, C., Stephens, P., Smith, R., Dalgliesh, G.L., Hunter, C., Bignell, G.,
Davies, H., Teague,J., Butler, A., Stevens, C., et al. (2007). Patterns of somatic
mutation in human cancer genomes. Nature 446, 153–158.
Gui, Y., Guo, G., Huang, Y., Hu, X., Tang, A., Gao, S., Wu, R., Chen, C., Li, X.,
Zhou, L., et al. (2011). Frequent mutations of chromatin remodeling genes in
transitional cell carcinoma of the bladder. Nat. Genet. 43, 875–878.
Haiman, C.A., Chen, G.K., Blot, W.J., Strom, S.S., Berndt, S.I., Kittles, R.A.,
izing genetic risk at known prostate cancer susceptibility loci in African Amer-
icans. PLoS Genet. 7, e1001387.
Ha ´jkova ´, H., Markova ´, J., Ha? skovec, C., Sa ´rova ´, I., Fuchs, O., Koste? cka, A.,
Cetkovsky ´, P., Michalova ´, K., and Schwarz, J. (2012). Decreased DNA meth-
ylation in acute myeloid leukemia patients with DNMT3A mutations and prog-
nostic implications of DNA methylation. Leuk. Res. 36, 1128–1133.
Hammerman, P.S., Hayes, D.N., Wilkerson, M.D., Schultz, N., Bose, R., Chu,
A., Collisson, E.A., Cope, L., Creighton, C.J., Getz, G., et al.; Cancer Genome
Atlas Research Network. (2012). Comprehensive genomic characterization of
squamous cell lung cancers. Nature 489, 519–525.
Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cell 100,
Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next gener-
ation. Cell 144, 646–674.
Hindorff, L.A., Gillanders, E.M., and Manolio, T.A. (2011). Genetic architecture
of cancer and other complex diseases: lessons learned and future directions.
Carcinogenesis 32, 945–954.
Ho, L., and Crabtree, G.R. (2010). Chromatin remodelling during development.
Nature 463, 474–484.
Hodges, E., Xuan, Z., Balija, V., Kramer, M., Molla, M.N., Smith, S.W., Middle,
C.M., Rodesch, M.J., Albert, T.J., Hannon, G.J., and McCombie, W.R. (2007).
Genome-wide in situ exon capture for selective resequencing. Nat. Genet. 39,
Hodis, E., Watson, I.R., Kryukov, G.V., Arold, S.T., Imielinski, M., Theurillat,
J.P., Nickerson, E., Auclair, D., Li, L., Place, C., et al. (2012). A landscape of
driver mutations in melanoma. Cell 150, 251–263.
Holland, A.J., and Cleveland, D.W. (2012). Chromoanagenesis and cancer:
mechanisms and consequences of localized, complex chromosomal rear-
rangements. Nat. Med. 18, 1630–1638.
Hou, Y., Song, L., Zhu, P., Zhang, B., Tao, Y., Xu, X., Li, F., Wu, K., Liang, J.,
Shao, D., et al. (2012). Single-cell exome sequencing and monoclonal evolu-
tion of a JAK2-negative myeloproliferative neoplasm. Cell 148, 873–885.
Huang, J., Deng, Q., Wang, Q., Li, K.Y., Dai, J.H., Li, N., Zhu, Z.D., Zhou, B.,
Liu, X.Y., Liu, R.F., et al. (2012a). Exome sequencing of hepatitis B virus-asso-
ciated hepatocellular carcinoma. Nat. Genet. 44, 1117–1121.
Huang, S., Gulzar, Z.G., Salari, K., Lapointe, J., Brooks, J.D., and Pollack, J.R.
(2012b). Recurrent deletion of CHD1 in prostate cancer with relevance to cell
invasiveness. Oncogene 31, 4164–4170.
Huang, F.W., Hodis, E., Xu, M.J., Kryukov, G.V., Chin, L., and Garraway, L.A.
(2013). Highly recurrent TERT promoter mutations in human melanoma.
Science 339, 957–959.
T., Hamer, K., Sasaki, N., Finegold, M.J., et al. (2013). In vitro expansion of
single Lgr5+ liver stem cells induced by Wnt-driven regeneration. Nature
Hudson, T.J., Anderson, W., Artez, A., Barker, A.D., Bell, C., Bernabe ´, R.R.,
Bhan, M.K., Calvo, F., Eerola, I., Gerhard, D.S., et al.; International Cancer
Genome Consortium. (2010). International network of cancer genome
projects. Nature 464, 993–998.
Imielinski, M., Berger, A.H., Hammerman, P.S., Hernandez, B., Pugh, T.J.,
thehallmarks oflungadenocarcinomawithmassivelyparallel sequencing.Cell
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 33
International Human Genome Sequencing Consortium. (2004). Finishing the
euchromatic sequence of the human genome. Nature 431, 931–945.
Iyer, G., Hanrahan, A.J., Milowsky, M.I., Al-Ahmadie, H., Scott, S.N.,
Janakiraman, M., Pirun, M., Sander, C., Socci, N.D., Ostrovnaya, I., et al.
(2012). Genome sequencing identifies a basis for everolimus sensitivity.
Science 338, 221.
Jaju, R.J., Fidler, C., Haas, O.A., Strickson, A.J., Watkins, F., Clark, K., Cross,
fused to NUP98 in the t(5;11)(q35;p15.5) in de novo childhood acute myeloid
leukemia. Blood 98, 1264–1267.
Jankowska, A.M., Makishima, H., Tiu, R.V., Szpurka, H., Huang, Y., Traina, F.,
Visconte, V., Sugimoto, Y., Prince, C., O’Keefe, C., et al. (2011). Mutational
spectrum analysis of chronic myelomonocytic leukemia includes genes asso-
ciated with epigenetic regulation: UTX, EZH2, and DNMT3A. Blood 118,3932–
Johannessen, C.M., Boehm, J.S., Kim, S.Y., Thomas, S.R., Wardwell, L.,
Johnson, L.A., Emery, C.M., Stransky, N., Cogdill, A.P., Barretina, J., et al.
(2010). COT drives resistance to RAF inhibition through MAP kinase pathway
reactivation. Nature 468, 968–972.
Jones, S., Wang, T.L., Shih, IeM., Mao, T.L., Nakayama, K., Roden, R., Glas,
R., Slamon, D., Diaz, L.A., Jr., Vogelstein, B., et al. (2010). Frequent mutations
Kaida, D., Motoyoshi, H., Tashiro, E., Nojima, T., Hagiwara, M., Ishigami, K.,
Watanabe, H., Kitahara, T., Yoshida, T., Nakajima, H., et al. (2007). Spliceos-
tatin A targets SF3b and inhibits both splicing and nuclear retention of pre-
mRNA. Nat. Chem. Biol. 3, 576–583.
P., Haverty, P.M., Bourgon, R., Zheng, J., et al. (2010). Diverse somatic muta-
tion patterns and pathway alterations in human cancers. Nature 466, 869–873.
in dogs and other model organisms. Nat. Rev. Genet. 9, 713–725.
Kim, D., and Salzberg, S.L. (2011). TopHat-Fusion: an algorithm for discovery
of novel fusion transcripts. Genome Biol. 12, R72.
Kim, Y.H., Kwei, K.A., Girard, L., Salari, K., Kao, J., Pacyna-Gengelbach, M.,
Wang, P., Hernandez-Boussard, T., Gazdar, A.F., Petersen, I., et al. (2010).
Genomic and functional analysis identifies CRKL as an oncogene amplified
in lung cancer. Oncogene 29, 1421–1430.
Kim, E.S., Herbst, R.S., Wistuba, I.I., Lee, J.J., Blumenschein, G.R., Jr., Tsao,
A., Stewart, D.J., Hicks, M.E., Erasmus, J., Jr., Gupta, S., et al. (2011). The
BATTLE trial: personalizing therapy for lung cancer. Cancer Discov. 1, 44–53.
Klinakis,A.,Lobry,C., Abdel-Wahab,O.,Oh,P.,Haeno, H., Buonamici, S.,van
De Walle, I., Cathelin, S., Trimarchi, T., Araldi, E., et al. (2011). A novel tumour-
suppressor function for the Notch pathway in myeloid leukaemia. Nature 473,
Koivunen, P., Lee, S., Duncan, C.G., Lopez, G., Lu, G., Ramkissoon, S.,
Losman, J.A., Joensuu, P., Bergmann, U., Gross, S., et al. (2012). Transforma-
tion by the (R)-enantiomer of 2-hydroxyglutarate linked to EGLN activation.
Nature 483, 484–488.
Kostic, A.D., Ojesina, A.I., Pedamallu, C.S., Jung, J., Verhaak, R.G., Getz, G.,
and Meyerson, M. (2011). PathSeq: software to identify or discover microbes
by deep sequencing of human tissue. Nat. Biotechnol. 29, 393–396.
Kralovics, R., Passamonti, F., Buser, A.S., Teo, S.S., Tiedt, R., Passweg, J.R.,
JAK2 in myeloproliferative disorders. N. Engl. J. Med. 352, 1779–1790.
Krauthammer, M., Kong, Y., Ha, B.H., Evans, P., Bacchiocchi, A., McCusker,
identifies recurrent somatic RAC1 mutations in melanoma. Nat. Genet. 44,
Kreso, A., O’Brien, C.A., van Galen, P., Gan, O.I., Notta, F., Brown, A.M., Ng,
K., Ma, J., Wienholds, E., Dunant, C., et al. (2013). Variable clonal repopulation
dynamics influence chemotherapy response in colorectal cancer. Science
Kriaucionis,S.,and Heintz, N.(2009). ThenuclearDNA base5-hydroxymethyl-
cytosine is present in Purkinje neurons and the brain. Science 324, 929–930.
Landau, D.A., Carter, S.L., Stojanov, P., McKenna, A., Stevenson, K.,
Lawrence, M.S., Sougnez, C., Stewart, C., Sivachenko, A., Wang, L., et al.
(2013). Evolution and impact of subclonal mutations in chronic lymphocytic
leukemia. Cell 152, 714–726.
Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J.,
Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al.; International Human
Genome Sequencing Consortium. (2001). Initial sequencing and analysis of
the human genome. Nature 409, 860–921.
Langemeijer, S.M., Kuiper, R.P., Berends, M., Knops, R., Aslanyan, M.G.,
Massop, M., Stevens-Linders, E., van Hoogen, P., van Kessel, A.G.,
Raymakers, R.A., et al. (2009). Acquired mutations in TET2 are common in
myelodysplastic syndromes. Nat. Genet. 41, 838–842.
Le Gallo, M., O’Hara, A.J., Rudd, M.L., Urick, M.E., Hansen, N.F., O’Neil, N.J.,
Price, J.C., Zhang, S., England, B.M., Godwin, A.K., et al.; NIH Intramural
Sequencing Center (NISC) Comparative Sequencing Program. (2012). Exome
sequencing of serous endometrial tumors identifies recurrent somatic muta-
tionsinchromatin-remodelingandubiquitin ligasecomplex genes.Nat.Genet.
Levine, R.L., Wadleigh, M., Cools, J., Ebert, B.L., Wernig, G., Huntly, B.J.,
Boggon, T.J., Wlodarska, I., Clark, J.J., Moore, S., et al. (2005). Activating
mutation in the tyrosine kinase JAK2 in polycythemia vera, essential thrombo-
cythemia, and myeloid metaplasia with myelofibrosis. Cancer Cell 7, 387–397.
Ley, T.J., Mardis, E.R., Ding, L., Fulton, B., McLellan, M.D., Chen, K., Dooling,
D., Dunford-Shore, B.H., McGrath, S., Hickenbotham, M., et al. (2008). DNA
sequencing of a cytogenetically normal acute myeloid leukaemia genome.
Nature 456, 66–72.
Ley, T.J., Ding, L., Walter, M.J., McLellan, M.D., Lamprecht, T., Larson, D.E.,
Kandoth, C., Payton, J.E., Baty, J., Welch, J., et al. (2010). DNMT3A mutations
in acute myeloid leukemia. N. Engl. J. Med. 363, 2424–2433.
Daniel, H.D., Kannangai, R., Offerhaus, G.J., et al. (2011). Inactivating muta-
tions of the chromatin remodeling gene ARID2 in hepatocellular carcinoma.
Nat. Genet. 43, 828–829.
Lindsley, R.C., and Ebert, B.L. (2013). Molecular Pathophysiology of Myelo-
dysplastic Syndromes. Annu. Rev. Pathol. 8, 21–47. Published online August
28, 2012. http://dx.doi.org/10.1146/annurev-pathol-011811-132436.
J.A., Balasubramanian, S., Bloom, T., Brennan, K.W., et al. (2012). Identifica-
tion of new ALK and RET gene fusions from colorectal and lung cancer biop-
sies. Nat. Med. 18, 382–384.
Dharmadhikari, A.V., Cooper, M.L., Wiszniewska, J., Zhang, F., Withers,
M.A., et al. (2011). Chromosome catastrophes involve replication mechanisms
generating complex genomic rearrangements. Cell 146, 889–903.
P.,Erez, A.,Nagamani,S.C., Dhar,S.U.,Ko1odziejska,K.E.,
Liu, X., Ory, V., Chapman, S., Yuan, H., Albanese, C., Kallakury, B., Timofeeva,
O.A., Nealon, C., Dakic, A., Simic, V., et al. (2012). ROCK inhibitor and feeder
cells induce the conditional reprogramming of epithelial cells. Am. J. Pathol.
Lohr, J.G., Stojanov, P., Lawrence, M.S., Auclair, D., Chapuy, B., Sougnez, C.,
Cruz-Gordillo, P., Knoechel, B., Asmann, Y.W., Slager, S.L., et al. (2012).
Discovery and prioritization of somatic mutations in diffuse large B-cell
lymphoma (DLBCL) by whole-exome sequencing. Proc. Natl. Acad. Sci.
USA 109, 3879–3884.
Lu, C., Ward, P.S., Kapoor, G.S., Rohle, D., Turcan, S., Abdel-Wahab, O.,
Edwards, C.R., Khanin, R., Figueroa, M.E., Melnick, A., et al. (2012). IDH muta-
tion impairs histone demethylation and results in a block to cell differentiation.
Nature 483, 474–478.
Lynch, T.J., Bell, D.W., Sordella, R., Gurubhagavatula, S., Okimoto, R.A.,
Brannigan, B.W., Harris, P.L., Haserlat, S.M., Supko, J.G., Haluska, F.G.,
et al. (2004). Activating mutations in the epidermal growth factor receptor
34 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
underlying responsiveness of non-small-cell lung cancer to gefitinib. N. Engl.
J. Med. 350, 2129–2139.
Macconaill, L.E., and Garraway, L.A. (2010). Clinical implications of the cancer
genome. J. Clin. Oncol. 28, 5219–5228.
MacConaill, L.E., Campbell, C.D., Kehoe, S.M., Bass, A.J., Hatton, C., Niu, L.,
Davis, M., Yao, K., Hanna, M., Mondal, C., et al. (2009). Profiling critical cancer
gene mutations in clinical tumor samples. PLoS ONE 4, e7887.
Makishima, H., Jankowska, A.M., Tiu, R.V., Szpurka, H., Sugimoto, Y., Hu, Z.,
Saunthararajah, Y., Guinta, K., Keddache, M.A., Putnam, P., et al. (2010).
Novel homo- and hemizygous mutations in EZH2 in myeloid malignancies.
Leukemia: official journal of the Leukemia Society of America. Leukemia 24,
Radmacher, M.D., Kohlschmidt, J., Nicolet, D., Whitman, S.P., et al. (2012).
Age-related prognostic impact of different types of DNMT3A mutations in
adults with primary cytogenetically normal acute myeloid leukemia. J. Clin.
Oncol. 30, 742–750.
Mardis, E.R., Ding, L., Dooling, D.J., Larson, D.E., McLellan, M.D., Chen, K.,
Koboldt, D.C., Fulton, R.S., Delehaunty, K.D., McGrath, S.D., et al. (2009).
N. Engl. J. Med. 361, 1058–1066.
Margulies, M.,Egholm,M.,Altman,W.E.,Attiya, S.,Bader,J.S., Bemben,L.A.,
Berka, J., Braverman, M.S., Chen, Y.J., Chen, Z., et al. (2005). Genome
sequencing in microfabricated high-density picolitre reactors. Nature 437,
Meyerson, M., Gabriel, S., and Getz, G. (2010). Advances in understanding
cancer genomes through second-generation sequencing. Nat. Rev. Genet.
Michor, F., Hughes, T.P., Iwasa, Y., Branford, S., Shah, N.P., Sawyers, C.L.,
and Nowak, M.A. (2005). Dynamics of chronic myeloid leukaemia. Nature
Molenaar, J.J., Koster, J., Zwijnenburg, D.A., van Sluis, P., Valentijn, L.J., van
der Ploeg, I., Hamdi, M., van Nes, J., Westerman, B.A., van Arkel, J., et al.
(2012). Sequencing of neuroblastoma identifies chromothripsis and defects
in neuritogenesis genes. Nature 483, 589–593.
Morin, R.D., Johnson, N.A., Severson, T.M., Mungall, A.J., An, J., Goya, R.,
Paul, J.E., Boyle, M., Woolcock, B.W., Kuchenbauer, F., et al. (2010). Somatic
mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell
lymphomas of germinal-center origin. Nat. Genet. 42, 181–185.
Morin, R.D., Mendez-Lago, M.,Mungall, A.J., Goya,R.,Mungall,K.L.,Corbett,
R.D., Johnson, N.A., Severson, T.M., Chiu, R., Field, M., et al. (2011). Frequent
mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature 476,
Murphy, A.B., Ukoli, F., Freeman, V., Bennett, F., Aiken, W., Tulloch, T., Coard,
K., Angwafo, F., and Kittles, R.A. (2012). 8q24 risk alleles in West African and
Caribbean men. Prostate 72, 1366–1373.
Navin, N., Kendall, J., Troge, J., Andrews, P., Rodgers, L., McIndoo, J., Cook,
K., Stepansky, A., Levy, D., Esposito, D., et al. (2011). Tumour evolution in-
ferred by single-cell sequencing. Nature 472, 90–94.
Nazarian, R., Shi, H., Wang, Q., Kong, X., Koya, R.C., Lee, H., Chen, Z., Lee,
M.K., Attar, N., Sazegar, H., et al. (2010). Melanomas acquire resistance to
B-RAF(V600E) inhibition by RTK or N-RAS upregulation. Nature 468, 973–977.
Ngo, V.N., Young, R.M., Schmitz, R., Jhavar, S., Xiao, W., Lim, K.H., Kohlham-
mer, H., Xu, W., Yang, Y., Zhao, H., et al. (2011). Oncogenically active MYD88
mutations in human lymphoma. Nature 470, 115–119.
Nik-Zainal, S., Alexandrov, L.B., Wedge, D.C., Van Loo, P., Greenman, C.D.,
Raine, K., Jones, D., Hinton, J., Marshall, J., Stebbings, L.A., et al.; Breast
Cancer Working Group of the International Cancer Genome Consortium.
(2012a). Mutational processes molding the genomes of 21 breast cancers.
Cell 149, 979–993.
Nik-Zainal, S., Van Loo, P., Wedge, D.C., Alexandrov, L.B., Greenman, C.D.,
Lau, K.W., Raine, K., Jones, D., Marshall, J., Ramakrishna, M., et al.; Breast
Cancer Working Group of the International Cancer Genome Consortium.
(2012b). The life history of 21 breast cancers. Cell 149, 994–1007.
Noushmehr, H., Weisenberger, D.J., Diefes, K., Phillips, H.S., Pujara, K.,
Berman, B.P., Pan, F., Pelloski, C.E., Sulman, E.P., Bhat, K.P., et al.; Cancer
Genome Atlas Research Network. (2010). Identification of a CpG island meth-
ylator phenotype that defines a distinct subgroup of glioma. Cancer Cell 17,
Nowell, P.C. (1976). The clonal evolution of tumor cell populations. Science
Ogino, S., Cantor, M., Kawasaki, T., Brahmandam, M., Kirkner, G.J.,
Weisenberger, D.J., Campan, M., Laird, P.W., Loda, M., and Fuchs, C.S.
(2006). CpG island methylator phenotype (CIMP) of colorectal cancer is best
characterised by quantitative DNA methylation analysis and prospective
cohort studies. Gut 55, 1000–1006.
Paez, J.G., Ja ¨nne, P.A., Lee, J.C., Tracy, S., Greulich, H., Gabriel, S., Herman,
P., Kaye, F.J., Lindeman, N., Boggon, T.J., et al. (2004). EGFR mutations in
lung cancer: correlation with clinical response to gefitinib therapy. Science
Pal, A., Barber, T.M., Van de Bunt, M., Rudge, S.A., Zhang, Q., Lachlan, K.L.,
Cooper, N.S., Linden, H., Levy, J.C., Wakelam, M.J., et al. (2012). PTEN muta-
tions as a cause of constitutive insulin sensitivity and obesity. N. Engl. J. Med.
Ramnarayanan, K., Shankar, S., Han, B., Cao, Q., Cao, X., Suleman, K.,
et al. (2010). Rearrangements of the RAF kinase pathway in prostate cancer,
gastric cancer and melanoma. Nat. Med. 16, 793–798.
N., Ateeq,B.,Kalyana-Sundaram,S., Pflueger,D.,
Pao, W., Miller, V., Zakowski, M., Doherty, J., Politi, K., Sarkaria, I., Singh, B.,
Heelan, R.,Rusch,V.,Fulton, L.,etal.(2004).EGF receptorgene mutationsare
common in lung cancers from ‘‘never smokers’’and are associated with sensi-
tivity of tumors to gefitinib and erlotinib. Proc. Natl. Acad. Sci. USA 101,
D., Pellagatti, A., Wainscoat, J.S., Hellstrom-Lindberg, E., Gambacorti-Pass-
erini, C., et al.; Chronic Myeloid Disorders Working Group of the International
Cancer Genome Consortium. (2011). Somatic SF3B1 mutation in myelodys-
plasia with ring sideroblasts. N. Engl. J. Med. 365, 1384–1395.
Parsons, D.W., Jones, S., Zhang, X., Lin, J.C., Leary, R.J., Angenendt, P.,
Mankoo, P., Carter, H., Siu, I.M., Gallia, G.L., et al. (2008). An integrated
genomic analysis of human glioblastoma multiforme. Science 321, 1807–
Pasqualucci, L., Dominguez-Sola, D., Chiarenza, A., Fabbri, G., Grunn, A.,
Trifonov, V., Kasper, L.H., Lerach, S., Tang, H., Ma, J., et al. (2011a). Inacti-
vating mutations of acetyltransferase genes in B-cell lymphoma. Nature 471,
Pasqualucci,L., Trifonov,V.,Fabbri, G., Ma,J., Rossi,D., Chiarenza,A.,Wells,
V.A., Grunn, A., Messina, M., Elliot, O., et al. (2011b). Analysis of the coding
genome of diffuse large B-cell lymphoma. Nat. Genet. 43, 830–837.
Peifer, M., Ferna ´ndez-Cuesta, L., Sos, M.L., George, J., Seidel, D., Kasper,
L.H., Plenker, D., Leenders, F., Sun, R., Zander, T., et al. (2012). Integrative
genome analyses identify key somatic driver mutations of small-cell lung
cancer. Nat. Genet. 44, 1104–1110.
Pleasance, E.D., Cheetham, R.K., Stephens, P.J., McBride, D.J., Humphray,
S.J., Greenman, C.D., Varela, I., Lin, M.L., Ordo ´n ˜ez, G.R., Bignell, G.R.,
et al. (2010a). A comprehensive catalogue of somatic mutations from a human
cancer genome. Nature 463, 191–196. Published online December 16, 2009.
Pleasance, E.D., Stephens, P.J., O’Meara, S., McBride, D.J., Meynert, A.,
Jones, D., Lin, M.L., Beare, D., Lau, K.W., Greenman, C., et al. (2010b). A
small-cell lung cancer genome with complex signatures of tobacco exposure.
Nature 463, 184–190. Published online December 16, 2009. http://dx.doi.org/
Pollock, P.M., Gartside, M.G., Dejeza, L.C., Powell, M.A., Mallon, M.A.,
Davies, H., Mohammadi, M., Futreal, P.A., Stratton, M.R., Trent, J.M., and
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 35
Goodfellow, P.J. (2007). Frequent activating FGFR2 mutations in endometrial
carcinomas parallel germline mutations associated with craniosynostosis and
skeletal dysplasia syndromes. Oncogene 26, 7158–7162.
Pomerantz, M.M., and Freedman, M.L. (2011). The genetics of cancer risk.
Cancer J. 17, 416–422.
Pomerantz, M.M., Ahmadiyeh, N., Jia, L., Herman, P., Verzi, M.P.,
Doddapaneni, H., Beckwith, C.A., Chan, J.A., Hills, A., Davis, M., et al.
(2009). The 8q24 cancer risk variant rs6983267 shows long-range interaction
with MYC in colorectal cancer. Nat. Genet. 41, 882–884.
Poulikakos, P.I., Persaud, Y., Janakiraman, M., Kong, X., Ng, C., Moriceau, G.,
Shi, H., Atefi, M., Titz, B., Gabay, M.T., et al. (2011). RAF inhibitor resistance is
mediated by dimerization of aberrantly spliced BRAF(V600E). Nature 480,
Puente, X.S., Pinyol, M., Quesada, V., Conde, L., Ordo ´n ˜ez, G.R., Villamor, N.,
Escaramis, G., Jares, P., Bea `, S., Gonza ´lez-Dı ´az, M., et al. (2011). Whole-
genome sequencing identifies recurrent mutations in chronic lymphocytic
leukaemia. Nature 475, 101–105.
Quesada, V., Conde, L., Villamor, N., Ordo ´n ˜ez, G.R., Jares, P., Bassaganyas,
L., Ramsay, A.J., Bea `, S., Pinyol, M., Martı ´nez-Trillos, A., et al. (2012). Exome
sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in
chronic lymphocytic leukemia. Nat. Genet. 44, 47–52.
Rausch, T., Jones, D.T., Zapatka, M., Stu ¨tz, A.M., Zichner, T., Weischenfeldt,
J., Ja ¨ger, N., Remke, M., Shih, D., Northcott, P.A., et al. (2012). Genome
sequencing of pediatric medulloblastoma links catastrophic DNA rearrange-
ments with TP53 mutations. Cell 148, 59–71.
recurrent rearrangements of the MAST kinase and Notch gene families in
breast cancer. Nat. Med. 17, 1646–1651.
Robinson, G.,Parker, M.,Kranenburg, T.A., Lu, C.,Chen,X.,Ding,L.,Phoenix,
T.N., Hedlund, E., Wei, L., Zhu, X., et al. (2012). Novel mutations target distinct
subgroups of medulloblastoma. Nature 488, 43–48.
Rosati, R., La Starza, R., Veronese, A., Aventin, A., Schwienbacher, C.,
Vallespi, T., Negrini, M., Martelli, M.F., and Mecucci, C. (2002). NUP98 is fused
to theNSD3gene inacute myeloid
t(8;11)(p11.2;p15). Blood 99, 3857–3860.
Salari, K., Spulak, M.E., Cuff, J., Forster, A.D., Giacomini, C.P., Huang, S., Ko,
M.E., Lin, A.Y., van de Rijn, M., and Pollack, J.R. (2012). CDX2 is an amplified
lineage-survival oncogene in colorectal cancer. Proc. Natl. Acad. Sci. USA
Samuels, Y., Wang, Z., Bardelli, A., Silliman, N., Ptak, J., Szabo, S., Yan, H.,
Gazdar, A., Powell, S.M., Riggins, G.J., et al. (2004). High frequency of muta-
tions of the PIK3CA gene in human cancers. Science 304, 554.
Sturm, D., Fontebasso, A.M., Quang, D.A., To ¨njes, M., et al. (2012). Driver
mutations in histone H3.3 and chromatin remodelling genes in paediatric glio-
blastoma. Nature 482, 226–231.
Seshagiri, S., Stawiski, E.W., Durinck, S., Modrusan, Z., Storm, E.E., Conboy,
C.B., Chaudhuri, S., Guan, Y., Janakiraman, V., Jaiswal, B.S., et al. (2012).
Recurrent R-spondin fusions in colon cancer. Nature 488, 660–664.
Shibata, T., Ohta, T., Tong, K.I., Kokubu, A., Odogawa, R., Tsuta, K., Asamura,
H., Yamamoto, M., and Hirohashi, S. (2008). Cancer related mutations inNRF2
impair its recognition by Keap1-Cul3 E3 ligase and promote malignancy. Proc.
Natl. Acad. Sci. USA 105, 13568–13573.
Sjo ¨blom, T., Jones, S., Wood, L.D., Parsons, D.W., Lin, J., Barber, T.D.,
Mandelker, D., Leary, R.J., Ptak, J., Silliman, N., et al. (2006). The consensus
coding sequences of human breast and colorectal cancers. Science 314,
Sosman, J.A., Kim, K.B., Schuchter, L., Gonzalez, R., Pavlick, A.C., Weber,
J.S., McArthur, G.A., Hutson, T.E., Moschos, S.J., Flaherty, K.T., et al.
(2012). Survival in BRAF V600-mutant advanced melanoma treated with
vemurafenib. N. Engl. J. Med. 366, 707–714.
Stehelin, D., Varmus, H.E., Bishop, J.M., and Vogt, P.K. (1976). DNA related to
the transforming gene(s) of avian sarcoma viruses is present in normal avian
DNA. Nature 260, 170–173.
Stephens, P.J., Greenman, C.D., Fu, B., Yang, F., Bignell, G.R., Mudie, L.J.,
Pleasance, E.D., Lau, K.W., Beare, D., Stebbings, L.A., et al. (2011). Massive
genomic rearrangement acquired in a single catastrophic event during cancer
development. Cell 144, 27–40.
Stephens, P.J., Tarpey, P.S., Davies, H., Van Loo, P., Greenman, C., Wedge,
D.C., Nik-Zainal, S., Martin, S., Varela, I., Bignell, G.R., et al.; Oslo Breast
Cancer Consortium (OSBREAC). (2012). The landscape of cancer genes and
mutational processes in breast cancer. Nature 486, 400–404.
Stransky, N., Egloff, A.M., Tward, A.D., Kostic, A.D., Cibulskis,K., Sivachenko,
A., Kryukov, G.V., Lawrence, M.S., Sougnez, C., McKenna, A., et al. (2011).
The mutational landscape of head and neck squamous cell carcinoma.
Science 333, 1157–1160.
Stratton, M.R., Campbell, P.J., and Futreal, P.A. (2009). The cancer genome.
Nature 458, 719–724.
Straussman, R., Morikawa, T., Shee, K., Barzily-Rokni, M., Qian, Z.R., Du, J.,
Davis, A., Mongare, M.M., Gould, J., Frederick, D.T., et al. (2012). Tumour
micro-environment elicits innate resistance to RAF inhibitors through HGF
secretion. Nature 487, 500–504.
Sur, I.K., Hallikas, O., Va ¨ha ¨rautio, A., Yan, J., Turunen, M., Enge, M., Taipale,
M., Karhu, A., Aaltonen, L.A., and Taipale, J. (2012). Mice lacking a Myc
enhancer that includes human SNP rs6983267 are resistant to intestinal
tumors. Science 338, 1360–1363.
Tabin, C.J., Bradley, S.M., Bargmann, C.I., Weinberg, R.A., Papageorge, A.G.,
Scolnick, E.M., Dhar, R., Lowy, D.R., and Chang, E.H. (1982). Mechanism of
activation of a human oncogene. Nature 300, 143–149.
Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y.,
Agarwal, S., Iyer, L.M., Liu, D.R., Aravind, L., et al. (2009). Conversion of
5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL
partner TET1. Science 324, 930–935.
Taplin, M.E., Bubley, G.J., Shuster, T.D., Frantz, M.E., Spooner, A.E., Ogata,
G.K., Keer, H.N., and Balk, S.P. (1995). Mutation of the androgen-receptor
gene in metastatic androgen-independent prostate cancer. N. Engl. J. Med.
Thomas, R.K., Baker, A.C., Debiasi, R.M., Winckler, W., Laframboise, T., Lin,
W.M., Wang, M., Feng, W., Zander, T., MacConaill, L., et al. (2007). High-
throughput oncogene mutation profiling in human cancer. Nat. Genet. 39,
Toyota, M., Ahuja, N., Ohe-Toyota, M., Herman, J.G., Baylin, S.B., and Issa,
J.P. (1999). CpG island methylator phenotype in colorectal cancer. Proc.
Natl. Acad. Sci. USA 96, 8681–8686.
Trapnell, C., Pachter, L., and Salzberg, S.L. (2009). TopHat: discovering splice
junctions with RNA-Seq. Bioinformatics 25, 1105–1111.
Turcan, S., Rohle, D., Goenka, A., Walsh, L.A., Fang, F., Yilmaz, E., Campos,
C., Fabius, A.W., Lu, C., Ward, P.S., et al. (2012). IDH1 mutation is sufficient to
establish the glioma hypermethylator phenotype. Nature 483, 479–483.
Tuupanen, S., Turunen, M., Lehtonen, R., Hallikas, O., Vanharanta, S., Kivioja,
T., Bjo ¨rklund, M., Wei, G., Yan, J., Niittyma ¨ki, I., et al. (2009). The common
colorectal cancer predisposition SNP rs6983267 at chromosome 8q24
confers potential to enhanced Wnt signaling. Nat. Genet. 41, 885–890.
Varela,I., Tarpey, P.,Raine, K.,Huang,D., Ong, C.K.,Stephens,P.,Davies,H.,
Jones, D., Lin, M.L., Teague, J., et al. (2011). Exome sequencing identifies
frequent mutation of the SWI/SNF complex gene PBRM1 in renal carcinoma.
Nature 469, 539–542.
Venter, J.C., Adams, M.D., Myers, E.W., Li, P.W., Mural, R.J., Sutton, G.G.,
Smith, H.O., Yandell, M., Evans, C.A., Holt, R.A., et al. (2001). The sequence
of the human genome. Science 291, 1304–1351.
Wagle, N., Emery, C., Berger, M.F., Davis, M.J., Sawyer, A., Pochanard, P.,
Kehoe, S.M., Johannessen, C.M., Macconaill, L.E., Hahn, W.C., et al. (2011).
Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor
genomic profiling. J. Clin. Oncol. 29, 3085–3096.
36 Cell 153, March 28, 2013 ª2013 Elsevier Inc.
Wagle, N., Berger, M.F., Davis, M.J., Blumenstiel, B., Defelice, M., Pochanard,
P., Ducar, M., Van Hummelen, P., Macconaill, L.E., Hahn, W.C., et al. (2012).
High-throughput detection of actionable genomic alterations in clinical tumor
samples by targeted, massively parallel sequencing. Cancer Discov. 2, 82–93.
Walter, M.J., Ding, L., Shen, D., Shao, J., Grillot, M., McLellan, M., Fulton, R.,
Schmidt, H., Kalicki-Veizer, J., O’Laughlin, M., et al. (2011). Recurrent
DNMT3A mutations in patients with myelodysplastic syndromes. Leukemia:
official journal of the Leukemia Society of America. Leukemia 25, 1153–1158.
Wang, L., Lawrence, M.S., Wan, Y., Stojanov, P., Sougnez, C., Stevenson, K.,
Werner, L., Sivachenko, A., DeLuca, D.S., Zhang, L., et al. (2011a). SF3B1 and
other novel cancer genes in chronic lymphocytic leukemia. N. Engl. J. Med.
Wang, N.J., Sanborn, Z., Arnett, K.L., Bayston, L.J., Liao, W., Proby, C.M.,
Leigh, I.M., Collisson, E.A., Gordon, P.B., Jakkula, L., et al. (2011b). Loss-of-
function mutations in Notch receptors in cutaneous and lung squamous cell
carcinoma. Proc. Natl. Acad. Sci. USA 108, 17761–17766.
Wang, X.S.,Shankar, S., Dhanasekaran, S.M., Ateeq, B.,Sasaki, A.T., Jing, X.,
Robinson, D., Cao, Q., Prensner, J.R., Yocum, A.K., et al. (2011c). Character-
izationofKRAS rearrangements inmetastaticprostate cancer.CancerDiscov.
Ward, P.S., Patel, J., Wise, D.R.,Abdel-Wahab, O., Bennett, B.D., Coller, H.A.,
Cross, J.R., Fantin, V.R., Hedvat, C.V., Perl, A.E., et al. (2010). The common
feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic
enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer
Cell 17, 225–234.
Weinberg, R. (2010). Point: Hypotheses first. Nature 464, 678.
Weir, B.A., Woo, M.S., Getz, G., Perner, S., Ding, L., Beroukhim, R., Lin, W.M.,
Province, M.A., Kraja, A., Johnson, L.A., et al. (2007). Characterizing the
cancer genome in lung adenocarcinoma. Nature 450, 893–898.
Weisenberger, D.J., Siegmund, K.D., Campan, M., Young, J., Long, T.I.,
Faasse, M.A., Kang, G.H., Widschwendter, M., Weener, D., Buchanan, D.,
et al. (2006). CpG island methylator phenotype underlies sporadic microsatel-
lite instability and is tightly associated with BRAF mutation in colorectal
cancer. Nat. Genet. 38, 787–793.
Weiss, J., Sos, M.L., Seidel, D., Peifer, M., Zander, T., Heuckmann, J.M.,
Ullrich, R.T., Menon, R., Maier, S., Soltermann, A., et al. (2010). Frequent
and focal FGFR1 amplification associates with therapeutically tractable
FGFR1 dependency insquamous cell lung cancer. Sci. Transl. Med. 2,62ra93.
Weng, A.P., Ferrando, A.A., Lee, W., Morris, J.P., 4th, Silverman, L.B.,
Sanchez-Irizarry, C., Blacklow, S.C., Look, A.T., and Aster, J.C. (2004). Acti-
vating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia.
Science 306, 269–271.
Whittaker, S.R., Theurillat, J.P., Van Allen, E., Wagle, N., Hsiao, J., Cowley,
G.S., Schadendorf, D., Root, D.E.,and Garraway, L.A. (2013). A genome-scale
RNA interference screen implicates NF1 loss in resistance to RAF inhibition.
Cancer Discov. Published online January 3, 2013. http://dx.doi.org/10.1158/
Wiegand, K.C., Shah,S.P., Al-Agha, O.M., Zhao, Y.,Tse, K., Zeng, T.,Senz, J.,
McConechy, M.K., Anglesio, M.S., Kalloger, S.E., et al. (2010). ARID1A muta-
tions in endometriosis-associated ovarian carcinomas. N. Engl. J. Med. 363,
Williams, S.E., Beronja, S., Pasolli, H.A., and Fuchs, E. (2011). Asymmetric cell
divisions promote Notch-dependent epidermal differentiation. Nature 470,
Wilson, B.G., and Roberts, C.W. (2011). SWI/SNF nucleosome remodellers
and cancer. Nat. Rev. Cancer 11, 481–492.
Wood, L.D., Parsons, D.W., Jones, S., Lin, J., Sjo ¨blom, T., Leary, R.J., Shen,
D., Boca, S.M., Barber, T., Ptak, J., et al. (2007). The genomic landscapes of
human breast and colorectal cancers. Science 318, 1108–1113.
Xu, W., Yang, H., Liu, Y., Wang, P., Kim, S.H., Ito, S., Yang, C., Wang, P., Xiao,
M.T., Liu, L.X., et al. (2011). Oncometabolite 2-hydroxyglutarate is a com-
petitive inhibitor of a-ketoglutarate-dependent dioxygenases. Cancer Cell
Xu, X.,Hou, Y., Yin,X., Bao, L., Tang, A.,Song, L.,Li, F., Tsang, S., Wu,K., Wu,
H., etal.(2012).Single-cellexomesequencing reveals single-nucleotide muta-
tion characteristics of a kidney tumor. Cell 148, 886–895.
Yan, H., Parsons, D.W., Jin, G., McLendon, R., Rasheed, B.A., Yuan, W., Kos,
I., Batinic-Haberle, I., Jones, S., Riggins, G.J., et al. (2009). IDH1 and IDH2
mutations in gliomas. N. Engl. J. Med. 360, 765–773.
Yoshida, K., Sanada, M., Shiraishi, Y., Nowak, D., Nagata, Y., Yamamoto, R.,
Sato, Y., Sato-Otsubo, A., Kon, A., Nagasaki, M., et al. (2011). Frequent
pathway mutations of splicing machinery in myelodysplasia. Nature 478,
Zang, Z.J., Cutcutache, I., Poon, S.L., Zhang, S.L., McPherson, J.R., Tao, J.,
Rajasegaran, V., Heng, H.L., Deng, N., Gan, A., et al. (2012). Exome
sequencing of gastric adenocarcinoma identifies recurrent somatic mutations
in cell adhesion and chromatin remodeling genes. Nat. Genet. 44, 570–574.
Zong, C., Lu, S., Chapman, A.R., and Xie, X.S. (2012). Genome-wide detection
of single-nucleotide and copy-number variations of a single human cell.
Science 338, 1622–1626.
Cell 153, March 28, 2013 ª2013 Elsevier Inc. 37