Decoding the genome beyond sequencing: The new phase of genomic research
Henry H.Q. Henga,b,c,⁎, Guo Liua, Joshua B. Stevensa, Steven W. Bremera, Karen J. Yea, Batoul Y. Abdallaha,
Steven D. Hornea, Christine J. Yed
aCenter for Molecular Medicine and Genetics, Wayne State University School of Medicine, Detroit, Michigan 48201, USA
bKarmanos Cancer Institute, Wayne State University School of Medicine, Detroit, Michigan 48201, USA
cDepartment of Pathology, Wayne State University School of Medicine, Detroit, Michigan 48201, USA
dDepartment of Internal Medicine, Wayne State University School of Medicine, Detroit, Michigan 48201, USA
a b s t r a c ta r t i c l ei n f o
Received 4 March 2011
Accepted 18 May 2011
Available online 26 May 2011
Non clonal chromosome aberrations (NCCAs)
Clonal chromosome aberrations (CCAs)
Punctuated cancer evolution
While our understanding of gene-based biology has greatly improved, it is clear that the function of the
genome and most diseases cannot be fully explained by genes and other regulatory elements. Genes and the
genome represent distinct levels of genetic organization with their own coding systems; Genes code parts like
protein and RNA, but the genome codes the structure of genetic networks, which are defined by the whole set
of genes, chromosomes and their topological interactions within a cell. Accordingly, the genetic code of DNA
offers limited understanding of genome functions. In this perspective, we introduce the genome theory which
calls for the departure of gene-centric genomic research. To make this transition for the next phase of genomic
research, it is essential to acknowledge the importance of new genome-based biological concepts and to
establish new technology platforms to decode the genome beyond sequencing.
© 2011 Elsevier Inc. All rights reserved.
developmentof various genomic technologies has greatly advanced the
science of genomics. However, despite cutting edge technologies
including whole genome scanning , global gene expression profiling
, copy number variation analysis  and massive parallel sequencing
, the understanding of the human genome and the mechanism of
human diseases remains a challenging process [5–7]. These powerful
technologies have generated scores of data, which paradoxically
challenge the framework of current genomics and gene based concepts
of common disease, including the rationale of analyzing large numbers
of diverse samples with the highest resolution possible. Many diseases
are in fact system diseases where sundry genetic variations can be
involved in a seemingly stochastic fashion. Furthermore, heterogeneity
just “genetic noise”), which cannot be addressed simply by sequencing
DNA and increasing the sample size . This issue represents the very
common diseases including cancer despite extensive large scale
sequencing and whole genome scanning .
There are two obvious but somewhat contrary options that can be
undertaken to move the field forward. One popular option is to
continuously push the limits of technology by increasing the
resolution and speeds while lowering costs in order to analyze more
samples [4,9,10]. It is believed that studying a larger number of
samples will yield the long anticipated genetic patterns of disease by
the elimination of ‘noise’. Unfortunately, many initial reports of this
approach have generated contradictory conclusions, revealing enor-
mous diversity rather than the expected reduction in diversity, and
that high levels of genetic heterogeneity seem to be the general rule
[5,11,12]. Questions are now being raised about whether data from
large scale genomic studies will ever prove to be of promised clinical
value, even if each personal genome or “cancer genome” is sequenced
[13,14]. The extreme complexity of disease heterogeneity, encom-
passes the following: low penetrance of specific gene mutations
within patient populations; multiple genetic–epigenetic and environ-
mental interactions; and the influences of stochastic evolutionary
processes, render most individual molecular mechanisms less than
useful for clinical prediction.
is a departurefromthetraditionalgeneticframeworkandwill provide
answers from a different perspective or level of genetic organization
rather than mainly focusing on DNA and RNA sequences. The basis for
this option is that genome alterations are more common and
profound than individual gene mutations in most human disease
conditions. This new conceptual framework based on the genome
Genomics 98 (2011) 242–252
⁎ Corresponding author at: Center for Molecular Medicine & Genetics, Wayne State
UniversitySchool ofMedicine,540 E.Canfield,3226Scott Hall, Detroit,MI48201.Fax:+1
313 577 5218.
E-mail address: email@example.com (H.H.Q. Heng).
0888-7543/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ygeno
theory calls for a redirect of our efforts to systematically decode
genetic information stored at the genome level [7,15,16]. Initial
calls for just such a change have emanated from the field of cancer
We strongly support the option of refocusingon the entire genome
(not just the sequence of), not only because this approach has been
overlooked, but also because of its ultimate importance in under-
standing how the entire genome functions and in the underpinnings
of the general mechanism of many common complex diseases . In
this perspective, we will briefly discuss the key differences between
genes and the genome by walking readers through our own
experience of making the transition from the gene centric view to
the genome theory . In particular, to convince readers of the
importance of this issue, we would like to point out that current
genome study efforts only decode parts of the genome and do not
address the key issue of decoding the genomeas a whole system. Even
for the ENCODE project (the encyclopedia of DNA elements) and
Human Epigenome project (as well as many other ‘omics’ projects)
[28,29], the conceptual framework is still the gene theory. At a
fundamental level, DNA sequences (including their chemical modifi-
cations) and the genome represent distinctively different levels of
coding and system control, and future genomic research must directly
address the issue of genome coding, genome system control, and how
interactions work across different genetic and epigenetic levels .
Equally important, new technical platforms are urgently needed to
synthesize information at the higher levels and integrate them with
the genome system. Emergent properties of the genome suggest that
system information at the genome level is not a simple summary of
gene sequence information. New technologies also need to integrate
other key features of normal and diseased genomes such as:
heterogeneity at multiple levels; differences between the system
status (such physiological and pathological conditions where patho-
logical conditions often involve genome alterations) and stochasticity
of somatic evolution.
2. Gene vs. genome
2.1. The genome is not equal to the sum of all genes or its entire sequence
The genome is the entity containing an organism's hereditary
information and the main evolutionary selection platform [27,30].
Traditionally, heritable information has been thought to be encoded
exclusively in DNA and RNA sequences. The current, popular concept
of the genome where the collection of all genes and non-coding
sequences explain a given species has been influenced by the gene
centric concept. A key unique feature of the genome however, is the
genomic topology (a multi-dimensional interactive relationship that
exists between genes and is the physical basis of genome architec-
ture) and the emergent properties that exist at this higher level which
have been largely ignored. As a result, the terminology, the Human
Genome Project, as used by sequencing consortiums, implies that
decoding DNA is equal to decoding the genome. This has spawned
many popular but incorrect analogies, including considering the
genome to be a book, where each chromosome represents a chapter.
The problem with this metaphor is that one cannot simply read the
basics of each chapter and comprehend it without including the
multi-dimensional interactions within the system. A chromosome
does not stand on its own as a biological entity and therefore there are
no meaningful messages based on individual chromosomes. To put it
succinctly, a simple parts list does not give a clue as to the assembly
instructions. Similarly, the conventional statement that the sequenc-
ing of the human genomehasprovideda roadmap (or the foundation)
of modern biomedical research is flawed. It is flawed particularly with
regard to nonlinear systems as most complex systems have multiple
levels of organization and the nonlinear relationship between them is
connected through emergent properties, which is difficult to
understand by only summarizing information of the lower parts.
The reductionist tradition of understanding the “parts” first before
understanding the “whole” system is only effective in a linear
Clearly, considering the relationship of the parts (genes) versus
the whole (genome), where the whole is more than the sum of all the
parts, and includes the 3-D interactive structure of the genome within
nuclei, the current reductionist approach of treating the genome as a
“bag of genes” or a collection of linear DNA structures does not mirror
the complexity of the genome system. To illustrate this point, we have
introduced the term, ‘genome context’, to differentiate the gene and
the genome [21,27]. Genome context (the DNA sequence plus the
genomic topology), rather than gene content, defines the structure of
a genetic networkand is the total interactive packagethat functions in
organismal and somatic cell evolution. The revelation of the genome
context results in the need for additional,crucial questionsto be asked
within the genomic community. These include: When a genome is
altered, does the same gene mutation have the same biological
meaning as it does in the original, unaltered genome? If the genome's
key properties are emergent from the DNA level and are fundamen-
tally different from DNA itself, is DNA sequencing crucial to
understand the function of a eukaryotic genome?
In the real world the difference between a parts list and the
blueprint of parts assembly is clear. In genome research, however, this
key difference is often forgotten. It is thus necessary to separate genes
and the genome as these are two different entities in genomic
research. The difference between them is not just a difference of
quantity but also a difference of the level of organization. One
interesting analogy is to consider genes as building material, parts or
even tools, and the genome as the architecture. The same materials
can be used to construct different architectures with distinct functions
. As pointed out by Barbara McClintock, the genome is the
organization that is responsible for activating and reconstructing in
response to environmental challenges .
It is true that increasing attention has been given to the “gene
context” rather than genes alone. However, without the conceptual
framework of the genome theory, isolated approaches will not solve
most paradoxes that the gene theory created .
2.2. Different genetic coding and control systems
If the above analogy that the gene/genome relationship represents
a part versus the whole relationship is correct, then it is obvious that
genes and genomes represent different levels of the system with
different mechanisms for coding genetic information. Understanding
how DNA codes RNA and proteins has been a major achievement of
molecular biology , however, the codes for producing the parts
using DNA andthe codes for assembling parts using the packageof the
genome are very different. Current systems biology suggests that the
function of the genome depends on the genetic network, without
knowing what defines the genetic network structure in the first place.
In contrast, genome theory states that the genome context defines the
structure of a genetic networkby creating the structural or topological
basis of genetic interactions [21,27], thus a specific network is the
emergent property of a given genome (Fig. 1).
According to the genome theory, both genes and the genomic
topology are the key to maintaining the genome system. Therefore,
new systems can be effectively created naturally either by creating
many novel genes, or by re-organizing the existing genome through
changing the genomic topology rather than gene content. Many less
complex species like sponge and water fleas have comparable or even
more genes than humans (comprised of a simple body plan without
organs, muscles or nerve cells, the sea sponge has 18,000 genes
compared to 21,000 genes for humans, while Daphinia water fleas
have over 30,000 genes) [34,35], and most mammals share similar
genes but have very different chromosomal arrangement and
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
composition (genome topology) [36–38]. We therefore conclude that
reorganizing the genome is more influential than inventing new
genes in terms of creating new systems.
It has been puzzling how the system preserves or codes genetic
information at levels above DNA such as the structure of a genetic
network. A recent effort to re-interpret the main function of sex has
unexpectedly provided the answer to this dilemma. It turns out that
both gene content and genome topology are ensured by sexual
reproduction, where the identical genome (reflected as karyotype) is
a key and any significantly altered genome will be eliminated by the
sexual filter [39,40]. As the same karyotype ensures the same genome
topology, and sexual reproduction maintains the same genome
(karyotype), sex in fact preserves the structure of a genetic network
by maintaining the genome context. Now we understand why in all
eukaryotes genetic information is transmitted from cell to cell by
chromosomes and that the genome context defines a species. In
asexual species lacking the sexual filter genome dynamics are
extremelyhigh, resulting in constant alteration of the genome system.
Similarly, very rapid karyotypic evolution can occur during somatic
cell evolution when there is no sexual filter to preserve the genome,
which might be the reason why cancer evolution can be achieved in a
relatively short time window within the human body while
organismal evolution with species stasis requires millions of years
. The important message here is that the genetic information of
system dynamics is not entirely encoded within the DNA. The coding
information at the gene level is conserved across species, but the
coding for gene interaction at the genome level is highly dynamic
among species while it is highly conserved within the same species.
As for the issue of whether genes/genomes represent different
control systems, the answer is obvious. No gene is an island and most
genes are not independent information units. Despite each gene
encoding a specific protein or RNA, global interactions are ultimately
controlled by the higher genome level, and the genome is a package
where emergent properties rule by balancing out contributions of
individual genes. This balance occurs as there are so many genes
contributing to the emergent properties, the influence of individual
genes is limited by the genome context. Interestingly, there are many
examples that illustrate the conflicting relationship between DNA and
the genome level. When identical transgenic DNA molecules are
introduced into mouse meiotic chromosomes, the loop size is
influenced by the integration sites along the chromosome. These
inserts form shorter loops integrating close to the telomeric region
rather than being inserted at the middle of chromosomes, demon-
strating chromosomal position constraints of DNA's behavior .
Similar observations include studies on position effects . Similarly,
when multiple copies of genes are integrated into the mouse genome
in a tandem array, only one copy is expressed possibly due to the
chromatin loop constraint, demonstrating that an organizational role
exists at the level above the gene [43–45].
Perhaps the most striking example of the conflict between the
gene and genome level is the dual functions of meiosis to reduce and
also promote genetic variation. Traditionally, sexual reproduction has
been thought to be a process that increases genetic diversity through
the mixing of genes, an obvious advantage of meiosis. Application of
genome theory, however, has surprisingly demonstrated that the
main function of sexual reproduction is to reduce genetic diversity at
the genome level and maintain system identity, while imparting a
certain degree of diversity at the gene level . This conclusion is
supported particularly through the re-examination of the evolution-
ary function of meiosis . According to Wilkins and Holliday's
recent insightful analysis, “…the conclusion is surprising: the initial
function of chromosome pairing was to limit, not enhance, recombi-
nation”. To further reconcile the field, the main function of sexual
reproduction has been linked to the reduction of genetic diversity in
general and at the genome level in particular . Interestingly, the
gene/genome relationship also might resolve the conflict between
short term evolutionary adaptation and long term system stasis ,
as the genome functions as a main evolutionary constraint (46, Heng,
unpublished data). The concept that different levels of a genetic
system need different coding systems could have broader implica-
tions. It strongly suggests that when considering a multiple level
complex system, one should not rely on coding or information at the
lower level to explain the behavior of the higher level, as distinct
coding and control systems are involved.
2.3. The Genome Theory and its clinical importance
To advance the field, the genome theory of organismal and somatic
cell evolution has been proposed based on key differences between
genes and the genome [24,27] (Fig. 1 and Table 1). The key message of
Fig. 1. A diagram demonstrating how changes of genome topology alter the structure of protein networks (diagram modified from Heng ). For simplicity, only two chromosomes
are drawn within the nucleus representing the genome. Genes are designated A, B, C, D, and E within the chromosomes. When a chromosomal translocation occurs, the genome
topology is altered affecting the physical relationship between chromatin domains, which changes the overall genetic network structure. As a result, the protein network changes
illustrated by the changed relationship between proteins A, B, C, D, and E. DNA codes the individual genes or parts and the interactive relationship among all genes and non-coding
sequences are achieved by the 3-D genome topology which serves as the platform or matrix where the self-organization principle works to form the genetic network.
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
the genome theory is that future genomic research should refocus
on/at the genome level, as the genome package defines the genetic
network and its potential response towards environmental stresses
[27,32].Thegoal ofsequencingthegenomehasbeen achieved anditis
now time to move to the next level of understanding of the genome
system rather than to continue to compile data on its parts list. A
switch to genome based approaches is essential to apply basic
genomic research to the clinic, as genome level changes often drive
DNA level changes.
Recent studies demonstrate that a few chromosomal transloca-
tions often couple with hundreds of rearrangements at the DNA level
. We have compared the copy number variations and karyotypic
dynamics (in the form of frequencies of NCCAs) during spontaneous
cellular transformation in mouse cells, and the highest level of copy
number variation occurs in association with the highest level of non-
clonal chromosome aberrations (NCCAs). This suggests that when the
system is unstable, elevated numbers of genomic alterations can be
detected across different levels of the system, though change at the
karyotype level typically reigns supreme (unpublished data). In fact,
the de novo locus-specific rate of genomic rearrangement is at least
100- to 10,000-fold greater than the rate of point mutations  and
there are an increasing number of reports linking many common
diseases to genome alterations [7,17,21,50]. Altogether, this mounting
evidence calls for a re-examination of the link between stochastic
genome alterations and common diseases especially cancer [7,24].
Linking more diseases to genome alterations rather than specific gene
mutations will ultimately challenge the validity of the current
sequencing efforts being used in an attempt to understand the
general mechanism of common diseases such as cancer. The influence
of different gene level alterations that can lead to the same disease
reduces the importance of monitoring any particular individual gene.
Thus, new genomic based strategies are clearly needed to predict
likely disease potential.
3. Gene and genome based technologies
3.1. Most current genomic technologies are based on the concept
of average profiles
As illustrated by this special issue, the vast majority of molecular
profiling methods are based on studying an average profile of cells. In
addition to the rationale of using the average to wash out ‘noise’,
enabling biological pattern recognition, technical limitations are
another reason why mixed cells are often used in molecular analysis.
If cell populations were truly homogeneous this method would be
justified however, for most disease conditions, compromise of
homeostasis of the system, leads to high heterogeneity. As heteroge-
neity is not insignificant noise, but a key feature of many diseases the
use of profile averaging is misleading [8,31]. For example, an average
profile tells little about potential drug resistance, because heteroge-
neity predominates in these situations making the use of averaging
technologies erroneous . Consider two populations of cells
descendent from healthy cells with a normal chromosome content.
Each of the populations undergoes chromosome gain, and thus differs
from the normal parental cells. In the first homogeneous population,
every cell contains the exact same chromosomal change, so that the
population average is equal to any individual cell. In the second highly
heterogeneous population changes occur that result in a population
average that is the same as the first population, however each
individual cell has a unique number of chromosomes, and thus
different genome system. A drug could be designed to effectively
eliminate the average population, and this drug would work well in
the case of the first population, but it may have little or no effect on
the second heterogeneous population. Similarly, when a given drug is
used, the average effect of the drug can be monitored using a Western
blot analysis that detects the proteins involved in cell death pathways
and measures average pathway response. However, when individual
cells are analyzed, there often are some exceptions noted including
cells that display even better growth. This type of heterogeneity
certainly contributes to drug resistance.
The idea of averaging profiles has been problematic for genetic
validation as well. If there are no average mutations, how can we
validate different mutations? Similarly, if the majority of cells are
different in terms of a gene mutation profile, what is the profile
average we are measuring? This point has been forcefully illustrated
time and time again in cancer research where a high degree of
heterogeneity is a common feature at the genome level [18–20].
Similarly, at the gene level, methods of direct mutation detection
rather than profile averaging have shown that there are large
numbers of random mutations within each tumor, the majority of
which are not recurrent (thus not detected by averaging methods)
. Indications are that the development and application of single
cell profiling methods used in genomic studies are increasing, which
represents a positive future direction.
3.2. How can stochastic genome alterations be monitored?
To understand the dynamic regulation of a genetic network, at
least three types of players need to be analyzed including genes, the
genomic topology and how the self-organization principle actually
works within the nucleus. Currently, most efforts have been focused
on the mechanism of network rewiring, without considering the
genomic topology and how system emergence occurs from lower
level components . Therefore, we need to be aware that there
might be a huge difference between the current knowledge of genetic
networks and the genomic reality. Even with simplified versions of
current analysis, there are two main aspects of network dynamics:
one is rewiring through small scale modifications such as gene
mutation,epigenetic regulationand copynumbervariation.This is the
main focus of current systems biology applied to analyzing network
regulation [53,54]. Another is system recreation through reorganiza-
tion of the genome using the same genes, a key topic that is still
ignored [8,21,25]. In organismal evolution, both processes occur with
genome reorganization being the more dominant for speciation.
Based on the genome theory, most diseases will involve the latter
category as pathological conditions are often caused by large scale
Key differences between genes and the genome.
Coded by DNA/RNACoded by entire set of chromosomes
Information preserved by the
same DNA sequences
Information preserved by the genome
Decoding by sequencing DNA Decoding by illustrating 3-D
interaction of parts
Information on the parts/tools
Information for system assembly/
Status change may lead to
Alteration creates a new genome
(changes are often neutral)
Usually involves macro-evolution
Modifies a system with new
Forms specific RNA/protein
Conserved across species
Defines a genetic network
Conservation of a species
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
these issues in a systematic way. There however have been some
important developments, which have started to move in this
direction. The following are some interesting examples.
3.2.1. Mapping and sequencing interaction sites
The importance of chromosomal and genome topology has long
been appreciated by molecular cytogenomics [17,21,55] and includes
the use of chromatin loop domains to study the distribution pattern of
specific regions of chromosomes within a nucleus , and pinpoint
neighboring chromosomes that are likely to involve translocations
. It is now clear that chromosomes are arranged differently in
different cell types, and that the 3-D chromosomal position influences
the gene activities . However, such important messages have
failed to capture the interest of main stream genome research, possibly
due to limited resolution and inability to applythis type of research in a
high throughput manner . Recently, using proximity-based ligation
(cross-linking to capture the information of 3-D interaction among
chromatin domains) and massive parallel sequencing, the interaction
sites of the genome were studied. Long-range interactions revealed
folding principles of the human genome. Interestingly, this study
confirms many theories generated from traditional cytogenetic studies,
such as the presence of chromosome territories and the spatial
proximity of small, gene-rich chromosomes within the nucleus, and
further illustrates an additional level of genome organization that is
form two genome-wide compartments . Various high-C analyses
have been performed , for example, a 3-D model of a 500-kb region
of human chromosome 16 has been built . We hope that such an
approach will soon capture more interest in the field .
different species. Future genomic comparisons are needed to study the
conservation of interactive relationships among networks by compar-
ing the genome context. Among the potential limitations of imparting
gene theory to explain genome issues which involve a different level of
the system. For example, when studying specific chromosomal trans-
locations in cancer, focus has centered on identifying genes altered by
translocation (dysfunctional genes or newly formed fusion genes),
without realizing that in addition to these directly changed genes, the
more dominant effect might come from the alteration of the entire
genomic topology defined network structure. The identification of a
fusion gene might only represent part of the story even though the
fusion gene is what we can easily trace based on current technologies.
Considering that there is a high degree of heterogeneity between
detectable fusion genes and cancer phenotypes in most cancers and
progression is diverse in patients with the same translocation, more
emphasis should be placed on the overall genome alteration.
Accordingly, utilization of the genome based framework needs to be
sped up to help guide and benefit future technological developments.
3.2.2. Using non clonal chromosome aberrations to study genome
Karyotype analysis has been extensively utilized in cancer and
human genetics research. Cytogenetic markers not only have helped
to identify disease specific genes, but have also been used in clinical
diagnosis [55,63]. Traditionally, however, only clonal chromosome
aberrations are reported, particularly when commonly shared among
patients. In contrast, detected NCCAs have been considered to be
insignificant genetic noise, and thus ignored despite occurring
commonly [17–20]. According to the genome theory, the level of
seemingly random genome alterations is not noise, but represents
system heterogeneity. Using cancer evolution as a model, we have
linked increased frequencies of NCCAs to system dynamics and
population diversity. Interestingly, the level of NCCAs can directly
reflect system instability, regardless of the causes of the NCCAs.
There is a seemingly unlimited list of factors that can lead to
system instability under certain conditions. In addition to genes that
are known to maintain genome integrity, many onco-proteins,
carcinogenic treatments, viral infections, inflammation, the aging
process, various disease conditions and even general physical stresses
can contribute to system instability that is linked with elevated
frequencies of NCCAs . Furthermore, following the comparison of
different cancer models with different molecular mechanisms, we
found that the general mechanism can be explained by the direct
relationship between stress induced system dynamics (illustrated by
elevated NCCAs) and tumorigenicity . By connecting the dots, we
have finally established the evolutionary mechanism of cancer.
Evolutionary mechanism≥∑all individual molecular mechanisms
The evolutionary mechanism of cancer can be explained by
three simple key components, namely, stress induced system
dynamics, population diversity and genome (not gene) mediated
system macro-evolution. To state that the evolutionary mechanism
of cancer is equal to the collection of all individual molecular
mechanisms, puts individual mechanisms that most researchers
are working on into perspective, as each specific mechanism can
explain some but not all cases [24,64]. Since there are countless
possible stochastic interactions that could occur between known
molecular mechanisms, and as evolution is an emergent property,
the evolutionary mechanism is actually much larger than the
collection of all molecular mechanisms and this emphasizes that
new strategies to monitor evolutionary potential are now needed
rather than strategies that focus only on individual molecular
mechanisms. In fact, the initial efforts of The Cancer Genome Atlas
have revealed that most key cancer gene mutations have low
penetration among patients. An even greater challenge is not just
characterizing each genetic aberration's defined pathway that can
contribute to cancer (this can be achieved by sequencing a great
number of samples), but more important is predicting the
pathway switching that occurs during treatment. This can further
reduce the clinical significance of individual genetic aberrations
when the penetration is very low within a patient population.
Interestingly, NCCAs are elevated in diseases other than cancer. For
example, we have detected much higher frequencies of NCCAs in a
limited number of Gulf War Syndrome patients (unpublished data).
Similarly, various mouse models of metabolic diseases have been
shown to exhibit increased NCCAs. It is thus necessary to evaluate
more patient samples to investigate the connection between system
instability and disease [7,31].
3.2.3. Analyzing genome chaos
To illustrate how drastically the genome can be re-organized
under stress, forming a totally new system, discussion of the
phenomenon of genome chaos (or karyotypic chaos) is required.
Our in vitro immortalization model has illustrated that the genome
displays massive genome alterations (including numerical and
structural alterations) and this process is essential for cellular
immortalization. Many translocation events occur in each genome,
andaseverycell is different,this phenomenonprovidesa large variety
of combinations of genetic material for genome evolution. Genome
chaos can be observed from an array of experimental and pathological
conditions . Fig. 2 illustrates one example of genome chaos
induced by drug treatment. SKY technologyhas been used to illustrate
the complexity of the translocations, as each color represents one
chromosomal origin. Many translocated chromosomes can be rapidly
formed from fragments of other chromosomes. It is worth noting that
SKY or multiple color FISH methods are essential to study these
massively altered karyotypes. Initially these technologies were
developed as a powerful method to identify the pattern of
karyotype abnormalities in cancer [64–66], and now our research
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
has demonstrated its ultimate usefulness in the study of stochastic
genome alterations such as NCCAs.
That so many genome level alterations can be generated so rapidly
when a cell endures stress is very important to understand the
mechanism of genome reshuffling and its evolutionary implications,
particularly in understanding the mechanism of drug resistance. It is
possible that chromosome fragmentation, which is a key mechanism
of mitotic cell death, plays a role in generating genome chaos.
Chromosome fragmentation is different from apoptosis and repre-
sents a powerful means to cut chromosomes into different pieces
(Fig. 3). Our hypothesis is that, following incomplete chromosomal
fragmentation [17,21,67,68], there is a new mechanism to rejoin the
massive amount of fragments to form chaotic chromosomes within a
few generations. Due to this massive re-organization occurring in
such a short time, some unknown mechanism rather than conven-
tional DNA repair must be involved. Interestingly, similar phenomena
have been reported in clinical samples where massive genome
alterations were detected which involved a limited number of
chromosomes. The term “chromothripsis” has been used to describe
this specific form of genome chaos [69,70]. Systematic analysis is
needed to correlate which massively altered genomes (different
systems) are associated with which different network structures.
Sequencing the newly formed rejoined chromosomes is also of
Fig. 2. SKY (spectral karyotype) image of genome chaos where there are massive translocations among chromosomes. A. In a normal mouse genome, each chromosome is “painted”
by one color. B. The reverse DAPI image of the same mitotic figure in A. C. There are massive translocation events detected within a chaotic genome following short drug treatment.
These newly formed giant chromosomes are possibly derived from complex chromosomal fusion, with each color representing their chromosomal origin. D. The reverse DAPI image
of the same mitotic figure in C. Drastically altered genomes represent a new genome system and understanding the mechanism generating genome chaos and genome evolution can
provide insight into genome re-organization.
Fig. 3. Images of chromosome fragmentation where mitotic chromosomes are cut into small pieces. A. Reverse DAPI image of a normal mouse karyotype. B. Example of chromosome
fragmentation where both chromosomes and fragments are detected. C. An example of late stage chromosome fragmentation where there is minimal visible chromosomal
morphology. Chromosome fragmentation is a newly identified form of mitotic celldeath. It is possible that chromosomefragmentation contributes to the formation of genome chaos.
Chromosome fragmentation also represents another example of previously disregarded non clonal chromosome aberrations.
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
4. Evolutionary and system approaches
New technologies need to be built around both holistic system
theresult of genomic evolution, and most diseases are a system problem
and disease initiation, progression and drug response/resistance all
represent typical evolutionary processes. Without the evolutionary
context, much of our genomic knowledge does not make sense.
For example, when discussing the issue of how much of the
genome is functional and why there are so many non-coding
sequences or non-essential genes, the historical evolutionary process of
genome reorganization and the dynamic of environmental variation
must be considered. The genomic evolution process is mainly based on
genome reshuffling rather than “useful gene accumulation” and
genomes are products that reflect their historical formation and their
mechanism of evolutionary selection. This point has often been
forgotten when searching for essential genes. The yeast model
illustrates this, as about 70% of yeast genes can be deleted under
experimental conditions without any apparent repercussion. This
approach is not representative of real world situations and thus is
fundamentally limited. In order to properly consider future challenges,
the following issues are worth discussing.
4.1. Which level should be the priority of a study?
There are many genetic and epigenetic levels that can be linked to
disease. At the gene level, gene mutations, splicing variations, and
epigenetic regulation can be linked to gene functions. At the
chromatin level, epigenetic regulation, domain configuration and 3-
D interactions are important. At the genome level, there is massive
copy number variation, and a great number of cytogenetic aberrations
that contribute to a given phenotype. How can an experimental
system be established to combine all of this information together?
How can these levels be prioritized when their information is
conflicting? Which levels of information are more important in
terms of clinical implications? Should we pay more attention to global
approaches or to reductionist approaches? Is it possible to ignore
some high resolution data at the lower system levels and focus on
higher system levels or even the overall system behavior?
To address these issues, one must consider both information and
evolutionary theories. According to the information theory, regarding
multiple levels of a system, lower level information often has less to
do with system control than higher levels. Consideration of the
recently established relationship between genes and the genome, and
between micro and macro evolution , enables classification of
diseases into different categories for future studies. If the issue is
driven by macro-evolution such as cancer where genome system
replacement is the key, the research focus must be on genome
alterations.Incontrast,if the researchinvolvesgenemutationwithout
genome level changes, such as developmental processes that are not a
result of changing the genome, then gene level analysis is appropriate.
Therefore, the first question to consider with regard to specific diseases
is determining whether macro or micro-evolution is involved.
To judge whether macro-evolution is involved, the process
dynamics need to be analyzed (of a cell culture process or animal
model) by comparing karyotypes. This can be achieved by spectral
karyotype (SKY) analysis. Different observation time points during
the process are selected and 50–100 mitotic figures are analyzed to
record the patterns of NCCAs and CCAs (clonal chromosome
aberrations). If there is a high level of genome turnover during the
monitoring process, thenmacro-cellular evolutionis likely involved. It
should be pointed out that, there are many types of chromosomal
aberrations, and many such as NCCAs have been traditionally ignored.
For example, defective mitotic figures and chromosome fragmenta-
tion were regarded as slide preparation artifacts until our recent
studies [17,21,71,72]. Despite our series of studies demonstrating the
importance of NCCAs [18–22], the research community has been slow
to accept this important development. Systematic studies are needed
to characterize these ignored chromosomal aberrations. By monitor-
ing NCCAs, it will be much easier to link many diseases with system
instability and the macro-evolutionary process.
4.2. Is there a disease genome?
Diseases can be divided into four groups: based on their genetic
penetration and whether or not genome instability is involved. In
types A and B, genetic factors are commonly shared in patient
populations with stable (A) or unstable genomes (B). For types C and
D genetic factors are rare in patient populations with stable (C) or
unstable (D) genomes. Many common diseases belong to the type D
group . If the genome is stable and there are overwhelming
numbers of normal karyotypes across various stages of the disease
condition, such a disease should be considered a gene based disease. If
stable but abnormal karyotypes dominate during different stages of
the disease progression (like Down's syndrome), then this genome
would be considered to be a “disease genome”. If however, the vast
majority of patients display drastically different karyotypes (such as
cancer patients), there is no cancer genome. Unfortunately, to date,
most of the cancer genome sequencing projects have ignored the
differences between genomes by focusing on known genes.
When genome-defined systems are completely different, what is
the meaning of comparing the same gene mutation? Recent reports of
various cancer genome sequencing projects have revealed high levels
of genome alterations with hundreds of rearrangements in individual
tumors. However, the majority of analyses still focus on the
association of known cancer genes without illustrating key genome
level contributions. Clearly, these massive genome alterations change
the meaning of individual cancer gene mutations as they have created
totally new genome systems [5,21,27].
The concept that specific functions of individual genes can be
altered by newly formed genomes has been elegantly confirmed by
yeast evolutionary studies . The MYO1 gene encodes proteins
essential for cytokinesis. Following MYO1 deletion, cytokinesis
function was restored in some cells through extensive genome
alteration. When MYO1 was re-introduced back to these altered
genomes, the gene was no longer relevant to these new systems and
no longer participated in the original functions. Similar thinking and
approaches need to be applied to other common diseases as well,
despite the difference of the degree of genome alteration between
cancer and other common diseases.
4.3. Two phases of evolutionary dynamics
Cancer evolution can be dissected into a repeating series of two
phases of evolutionary events . In the punctuated phase, macro-
evolution dominates while in the stepwise phase, there is a mixture of
macro-and microevolution occurring. Knowing the interaction of the
two phases of somatic cell evolution is useful to monitor system
evolution and adopt suitable monitoring technologies accordingly. In
stable systems, evolution can be triggered by either micro- or macro-
evolutionary events, so both gene level changes and genome level
changes need to be analyzed. In the dominant macro-evolution phase,
gene studies are of very limited value. Significantly, cancer evolution
can be observed as multiple cycles of NCCAs/CCAs which transition
during key status changes such as immortalization, transformation,
andacquisitionofdrug resistance.Therefore,theevolutionary journey
throughout each cancer event (from initiation through later clinically
significant stages), is not achieved by one genome system that
remains through all these stages. In contrast, it is achieved by a series
process (reflected by a succession of NCCAs/CCAs cycles) [19–21]. This
key point has been ignored by the current gene mutation based cancer
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
theory. In order to correctly understand the mechanism of cancer, the
importance of the two phases of evolutionary dynamics must be
appreciated, as the two phases of dynamics should be detectable at
levels below (such as the gene level) and above the genome . It
would be interesting to investigate what patterns of karyotype
evolution occur in other common diseases. Based on our studies on
different stress conditions inducing elevated levels of genome alter-
ation, it is reasonable to link genome alterations to many common
diseases. Increasing numbers of reports are pointing in this direction
[7,74,75]. However, there needs to be a greater emphasis on stochastic
genome alterations rather than clonal somatic genome variations.
4.4. Integrating the time factor in genomic research
The evolutionary approach in genomic studies must integrate the
time factor as time is a key aspect of the evolutionary process. Such
studies require sampling the process from multiple time windows in
order to watch evolutionin action. As evolution is a stochastic process,
and each tumor represents one run of independent evolution, the
molecular pathways involved in any given tumor might be drastically
different from one another. This fact might be the greatest challenge
when trying to apply individual pathway information to patient care.
However, by applying evolutionary principles, the evolutionary
potential may be predictable if based on the overall system stability.
For example, by ignoring individual specific molecular mechanisms,
we have focused on measuring the frequencies of NCCAs, which is an
indicator of system dynamics, to study the evolutionary potential of
cancer. Regardless of what the contributing factors are, as soon as the
frequencies of NCCAs reached a certain level, cancer formation
happened in due time. Further research needs to integrate the degree
of system instability and its progression over time. There are some
studies that address this issue [76,77], however they have been
focused on gene mutations rather than genome alterations.
5. Conclusion and future perspective
Due to the confusion between the genome context and the gene
content of the genome, a fashionable trend of pushing genomic
research “beyond the genome” has started, following the sequencing
phase of the genome project. Beyond the genome has become a topic
for genomic conferences and popular science. The truth of the matter
is that we have not yet decoded the genome and do not understand
how the genome defines the genetic network. Sequencing the whole
genome is just the first step to decoding the genome's DNA, and the
genome's crucial topological role and the mechanism of how the
different parts interact and create emergent functional properties is
basically unknown. It is important to realize that various networks
represent the emergent properties of the whole genome rather than
limited genes. Equally important, altered genomes lead to altered
network structures. It is thus less useful to study network dynamics
without monitoring the genome status. The real task for the research
community is to begin to really study the genome rather than look
beyond it. To achieve the goal of decoding the genome, new
fundamental principles of biology must be formed to provide a
conceptual framework and the needed guidance to develop new
technological platforms. Only these new genome based concepts and
methodologies (rather than the DNA technologies that we have used
so far) can effectively study genome level dynamics and alterations as
well as implications to human diseases.
This year marks the tenth anniversary of the completion of the
sequencing of the human genome (Human Genome Project) [4,9,10].
Since the launching of this project in 1990, both basic knowledge and
genomic technologies have drastically grown (e.g. from Sanger-based
capillary sequencing to massive parallel sequencing; from monitoring
the expression of a couple of genes to global expression analysis).
There is now an increased appreciation of important features of the
human genome, including the number of genes, AT/GC-content, the
high degree of alternative splicing (many more proteins than genes),
gene order/cluster, conservation of genes and non-coding sequences,
the similarity of the proteome across placental mammals (where the
creation of fundamental new proteins is rare), the significance of
transposons and non-coding RNA, copy number variations, linkage
disequilibrium, syntenic genomic blocks among species, codon usage
bias, and many others. These new discoveries/technologies have also
been applied to human diseases studies, where a total of 2850 genes
contributing to Mendelian diseases have been identified, and as more
than 1100 loci affecting more than 165 diseases have been associated
with common diseases, this illustrates the high levels of genetic
heterogeneity among patients . Paradoxically, the success of
identifying large numbers of genetic factors poses the greatest
challenge for their application to clinical use. For instance, there are
over 10,000 different genetic variants that have been associated with
Schizophrenia. Each of them is relatively rare and responsible for a
tiny increase in disease risk . What is the clinical significance of
such diverse genetic variants? Despite these impressive findings, little
is known on how to decode the genome as we have discussed. For
example, how does self-organization act upon the genome context?
How can the genomic topology be revealed from sequencing data?
How do altered genomes affect genetic networks and what is the
relationship between genome reorganization and re-wiring? What is
the relationship among different levels of genetic systems and the
implication to human diseases? Are genome level alterations
responsible for the missing heritability? Can massive amounts of
genomic data be applied to clinical settings? Despite many positive
comments regarding the impact of the sequencing project, a key
conclusion that is emerging is that just sequencing DNA will not
revealthe mystery of life, and a new way of thinkingthat departsfrom
the gene theory is now urgently needed [7,13,15,27,31].
The first and hardest step is accepting the fact that decoding the
genome and sequencing DNA are significantly different, and the
genomic landscape will not be revealed simply by sequencing more
DNA samples . It is time for the research community to develop
new concepts and technologies, rather than to continue to pile on
more of what we already know. We have recently introduced the
genome theory as a departure from the gene theory [21,24,27]. There
has also been an increasing call in chromosomal studies that “a DNA
sequence isn't enough: to understandthe workingsof the genome,we
must study chromosome structure” .
To advance the transition from gene oriented research to genome
based research, the following avenues need to be explored. First,
technical platforms are required to study system behavior (e.g.
evolutionary dynamics) without falling back on traditional approaches
genome level heterogeneity rather than monitoring each specific
individual genetic defect or pathway, as there are potentially so many
defects that attempt to analyze them all results in fuzzy data sets with
little or no meaning. By focusing on a different level, the lower level of
level of simplicity .
Second, systematic comparisons of the relationships of the genome,
transcriptome and proteome are needed [81,82]. Much as the
complement of all genes is not equal to the genome, the proteins of a
given genome are not the entire proteome without the essential
components of interaction dynamics (where the cellular topology is
important and is related to genome topology). Similarly, the viewpoint
proteins than genes) is not accurate. Again, it is not only the number of
protein species that matter, but the interaction potential that exists
which is defined by the genome. Sequencing the genome only reveals
the code of the parts list. The majority of current proteomics research
analyzes only the parts and its potential interactions are artificially
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
topology, the averaged interaction map is artificial and not truly
value. Thus, the transcriptome and proteome should be considered as
emergent properties of the genome. The understanding of the
convergent and divergent relationship among them is of importance,
especially the similarity and difference between typical Mendelian
and medical interventional conditions.
Third, there are many fascinating studies on network re-wiring
when altering master regulators [83,84]. Further analyses are needed
to integrate genome topology and compare network re-wiring within
a givengenomeandtodeterminehowtheformationof a newgenome
re-creates a new boundary for the new network, as they involve
different patterns of evolutionary dynamics. Clinically useful quanti-
tative models are needed with prediction value. So far, some big
challenges for system biology (where the gene theory dominates)
include dealingwithmultiple levels of heterogeneity that occurin real
biological systems (where the genome theory is essential) and
applying them to design practical useful models.
Fourth, genome chaos represents a powerful model system to
understand the gene and genome relationship, and their contri-
butions to human diseases [18,31]. With various experimental
manipulations, the evolutionary process of a chaotic genome can
be traced using single cell genomic analysis to study both the
micro and macro phases of evolution. In addition, networks can be
systematically compared within physiological and pathological
conditions, including with or without drug treatments. More
importantly, it should finally be determined whether quantitative
features from the gene level can be used to predict behavior
occurring at the genome level, which could also offer insight on
how information converges and/or diverges at different levels.
Fifth, cytogenomic analysis should once again be restored as a
level and is capable of studying individual cells as well as cell
populations. We are not recommending the current trend of the
molecularization of current cytogenetic analysis, as copy number
variation testing cannot be used to replace karyotype analysis (not
only do karyoytypic and sub-chromosomal changes represent different
levels of a system, but current molecular technologies profile the
average, favoring the biased clonal populations while eliminating
important heterogeneity). Similarly, to focus solely on individual
chromatin domains is not sufficient to study genome topology, the
domain view . Clearly, these excitingtechnologiesand observations
from chromatinloop domainsneed tobeintegrated intothelevelof the
whole genome package [43–45,86–88]. Similarly, chromosome based
genes but should also consider the genome perspective. For example,
the three dimensional synteny among different species needs to be
analyzed. Another important area to investigate is the linkage between
karyotypes and species to determine whether species display different
genome context and thereby establish the fact that it is the genome
rather genes that defines a species. Such an analysis should also be
carried out in cancer research to illustrate that the key feature of cancer
is non-shared random genome alterations.
Expectations of future genomic research need to be adjusted as
well, as there has been a genomic bubble in terms of its clinical
promise . Due to the high degree of complexity and heterogeneity
of the multiple levels of the genetic system, one cannot always
determine a distinct causative relationship. Our recent data illustrates
that there might be no clear causative relationship at the molecular
level within a complex biosystem (31, unpublished data). Common
diseases might not have dominant common patterns similar to those
identified in typical Mendelian disease . It is likely that stochastic
genome alterations are involved in common diseases and are the
general rule rather than the exception . Our studies have made the
case in cancer research and now it is time to apply the genome theory
to studies of other common diseases [5,7,24,27,75].
Another important realization is that communication with the
non-scientific community about the clinical implications of
genomic information needs to be more realistic. Yes, we can
sequence tens of thousands of cancer samples and establish the
most comprehensive genomic catalogs with highly heterogeneous
details, but this information might not be useful clinically for
prediction and treatment of individual patients, as “there are likely
to be fundamental limits on precise prediction due to the complex
architecture of common traits” . If high heterogeneity of the
genome is a key for many common diseases, sequencing DNA will
not deliver the promise of major medical advancement. Another
important note to large scale genomic research is that in addition
to paying attention to statistical significance, there also is a huge
need to determine the biological relevance.
A powerful approach for future functional genomics is to apply the
evolutionary principle and methodologies to study the relationship
between genome alteration, copy number variation, gene mutation,
epigenetic variation and network dynamics. We predict that the
identification of genome level alterations is the most important level
of a system, this will lead to the understanding that the overall genome
of that system [27,30,32,90,91,92,93,94]. In particular, a multiple
dimension global interactive map is needed exhibiting important
transitions between germline cells to various types of somatic cells,
from physiological conditions to pathological conditions, and demon-
strating differences between normal and abnormal genome systems.
Decoding the genome is much more difficult than analyzing DNA
as despite our superior capability to sequence DNA, we have not
achieved a fundamental technical platform to understand genome
level coding. However, this difficult transition MUST be achieved in
order to push the field of genomics forward. Welcome to the
challenges of the Genome Age.
We dedicate this perspective to Dr. Lap-Chee Tsui in honor of
his 60th birthday. HH deeply appreciates Dr. Tsui's outstanding
mentorship and influence. This manuscript is part of a series of
studies entitled, “The mechanisms of somatic cell and organismal
evolution”. We would like to thank Gloria Heppner; Gary Stein;
Donald Coffey and O.J. Miller for their continuous support and
interest in this project. Thanks also go to Markku Kurkinen,
Leonard Lipovich, Rafe Furst, Steve Krawetz and Jian-Bing Fan for
discussions. This work was partially supported by grants from the
Susan G. Komen Breast Cancer Foundation, SeeDNA Inc and the
 T.A. Pearson, T.A. Manolio, How to interpret a genome-wide association study,
JAMA 299 (2008) 1335–1344.
 O. Alter, P.O. Brown, D. Botstein, Singular value decomposition for genome-wide
expression data processing and modeling, Proc. Natl. Acad. Sci. U.S.A. 97 (2000)
 L. Feuk, A.R. Carson, S.W. Scherer, Structural variation in the human genome, Nat.
Rev. Genet. 7 (2006) 85–97.
 E.R. Mardis, A decade's perspective on DNA sequencing technology, Nature 470
 H.H. Heng, Cancer genome sequencing: the challenges ahead, Bioessays 29 (2007)
 T.A. Manolio, F.S. Collins, N.J. Cox, D.B. Goldstein, L.A. Hindorff, D.J. Hunter, M.I.
McCarthy, E.M. Ramos, L.R. Cardon, A. Chakravarti, J.H. Cho, A.E. Guttmacher, A.
Kong, L. Kruglyak, E. Mardis, C.N. Rotimi, M. Slatkin, D. Valle, A.S. Whittemore, M.
Boehnke, A.G. Clark, E.E. Eichler, G. Gibson, J.L. Haines, T.F. Mackay, S.A. McCarroll,
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
P.M. Visscher, Finding the missing heritability of complex diseases, Nature 461
 H.H. Heng, Missing heritability and stochastic genome alterations, Nat. Rev. Genet.
11 (2010) 813.
 H.H. Heng, S.W. Bremer, J.B. Stevens, K.J. Ye, G. Liu, C.J. Ye, Genetic and epigenetic
heterogeneity in cancer: a genome-centric perspective, J. Cell. Physiol. 220 (2009)
 E.S. Lander, Initial impact of the sequencing of the human genome, Nature 470
 E.D. Green, M.S. Guyer, Charting a course for genomic medicine from base pairs to
bedside, Nature 470 (2011) 204–213.
 L.D. Wood, D.W. Parsons, S. Jones, J. Lin, T. Sjoblom, R.J. Leary, D. Shen, S.M. Boca, T.
Barber, J. Ptak, N. Silliman, S. Szabo, Z. Dezso, V. Ustyanksky, T. Nikolskaya, Y.
Nikolsky, R. Karchin, P.A. Wilson, J.S. Kaminker, Z. Zhang, R. Croshaw, J. Willis, D.
Dawson, M. Shipitsin, J.K. Willson, S. Sukumar, K. Polyak, B.H. Park, C.L.
Pethiyagoda, P.V. Pant, D.G. Ballinger, A.B. Sparks, J. Hartigan, D.R. Smith, E. Suh,
N. Papadopoulos, P. Buckhaults, S.D. Markowitz, G. Parmigiani, K.W. Kinzler, V.E.
Velculescu, B. Vogelstein, The genomic landscapes of human breast and colorectal
cancers, Science 318 (2007) 1108–1113.
 M.F. Berger, M.S. Lawrence, F. Demichelis, Y. Drier, K. Cibulskis, A.Y. Sivachenko, A.
Sboner, R. Esgueva, D. Pflueger, C. Sougnez, R. Onofrio, S.L. Carter, K. Park, L.
Habegger, L. Ambrogio, T. Fennell, M. Parkin, G. Saksena, D. Voet, A.H. Ramos, T.J.
Pugh, J. Wilkinson, S. Fisher, W. Winckler, S. Mahan, K. Ardlie, J. Baldwin, J.W.
Simons, N. Kitabayashi, T.Y. MacDonald, P.W. Kantoff, L. Chin, S.B. Gabriel, M.B.
Gerstein, T.R. Golub, M. Meyerson, A. Tewari, E.S. Lander, G. Getz, M.A. Rubin, L.A.
Garraway, The genomic complexity of primary human prostate cancer, Nature
470 (2011) 214–220.
 E.E. Eichler, J. Flint, G. Gibson, A. Kong, S.M. Leal, J.H. Moore, J.H. Nadeau, Missing
heritability and strategies for finding the underlying causes of complex disease,
Nat. Rev. Genet. 11 (2010) 446–450.
 J.P. Evans, E.M. Meslin, T.M. Marteau, T. Caulfield, Genomics. Deflating the
genomic bubble, Science 331 (2011) 861–862.
 H.H. Heng, The conflict between complex systems and reductionism, JAMA 300
 H.H. Heng, G. Liu, J.B. Stevens, S.W. Bremer, K.J. Ye, C.J. Ye, Genetic and epigenetic
heterogeneity in cancer: the ultimate challenge for drug therapy, Curr. Drug
Targets 11 (2010) 1304–1316.
 H.H. Heng, J.B. Stevens, G. Liu, S.W. Bremer, C.J. Ye, Imaging genome abnormalities
in cancer research, Cell Chromosome 3 (2004) 1.
 H.H. Heng, J.B. Stevens, G. Liu, S.W. Bremer, K.J. Ye, P.V. Reddy, G.S. Wu, Y.A. Wang,
M.A. Tainsky, C.J. Ye, Stochastic cancer progression driven by non-clonal
chromosome aberrations, J. Cell. Physiol. 208 (2006) 461–472.
 H.H. Heng, G. Liu, S. Bremer, K.J. Ye, J. Stevens, C.J. Ye, Clonal and non-clonal
chromosome aberrations and genome variation and aberration, Genome 49
 H.H. Heng, S.W. Bremer, J. Stevens, K.J. Ye, F. Miller, G. Liu, C.J. Ye, Cancer
progression by non-clonal chromosome aberrations, J. Cell. Biochem. 98 (2006)
 C.J. Ye, G. Liu, S.W. Bremer, H.H. Heng, The dynamics of cancer chromosomes and
genomes, Cytogenet. Genome Res. 118 (2007) 237–246.
 H.H. Heng, J.B. Stevens, L. Lawrenson, G. Liu, K.J. Ye, S.W. Bremer, C.J. Ye, Patterns
of genome dynamics and cancer evolution, Cell. Oncol. 30 (2008) 513–514.
 H.H. Heng, The gene-centric concept: a new liability? Bioessays 30 (2008)
 H.H. Heng, J.B. Stevens, S.W. Bremer, K.J. Ye, G. Liu, C.J. Ye, The evolutionary
mechanism of cancer, J. Cell. Biochem. 109 (2010) 1072–1084.
 P. Duesberg, Chromosomal chaos and cancer, Sci. Am. 296 (2007) 52–59.
 J.M. Nicholson, P. Duesberg, On the karyotypic origin and evolution of cancer cells,
Cancer Genet. Cytogenet. 194 (2009) 96–110.
 H.H. Heng, The genome-centric concept: resynthesis of evolutionary theory,
Bioessays 31 (2009) 512–525.
 American Association for Cancer Research Human Epigenome Task Force,
European Union, Network of Excellence, Scientific Advisory Board,
Moving AHEAD with an international human epigenome project, Nature 454
 P.M. Durand, R.E. Michod, Genomics in the light of evolutionary transitions,
Evolution 64 (2010) 1533–1540.
 H.H.Q. Heng, Bio-complexity: challenging reductionism, Handbook on Systems
and Complexity in Health, (in press).
 B. McClintock, The significance of responses of the genome to challenge, Science
226 (1984) 792–801.
 F. Crick, What mad pursuit: a personal view of scientific discovery, Basic Books,
New York, 1988.
 M. Srivastava, O. Simakov, J. Chapman, B. Fahey, M.E. Gauthier, T. Mitros, G.S.
Richards, C. Conaco, M. Dacre, U. Hellsten, C. Larroux, N.H. Putnam, M. Stanke, M.
Adamska, A. Darling, S.M. Degnan, T.H. Oakley, D.C. Plachetzki, Y. Zhai, M.
Adamski, A. Calcino, S.F. Cummins, D.M. Goodstein, C. Harris, D.J. Jackson, S.P. Leys,
S. Shu, B.J. Woodcroft, M. Vervoort, K.S. Kosik, G. Manning, B.M. Degnan, D.S.
Rokhsar, The Amphimedon queenslandica genome and the evolution of animal
complexity, Nature 466 (2010) 720–726.
 J.K. Colbourne, M.E. Pfrender, D. Gilbert, W.K. Thomas, A. Tucker, T.H. Oakley, S.
Tokishita, A. Aerts, G.J. Arnold, M.K. Basu, D.J. Bauer, C.E. Caceres, L. Carmel, C.
Casola, J.H. Choi, J.C. Detter, Q. Dong, S. Dusheyko, B.D. Eads, T. Frohlich, K.A.
Geiler-Samerotte, D. Gerlach, P. Hatcher, S. Jogdeo, J. Krijgsveld, E.V. Kriventseva,
D. Kultz, C. Laforsch, E. Lindquist, J. Lopez, J.R. Manak, J. Muller, J. Pangilinan, R.P.
Patwardhan, S. Pitluck, E.J. Pritham, A. Rechtsteiner, M. Rho, I.B. Rogozin, O.
Sakarya, A. Salamov, S. Schaack, H. Shapiro, Y. Shiga, C. Skalitzky, Z. Smith, A.
Souvorov, W. Sung, Z. Tang, D. Tsuchiya, H. Tu, H. Vos, M. Wang, Y.I. Wolf, H.
Yamagata, T. Yamada, Y. Ye, J.R. Shaw, J. Andrews, T.J. Crease, H. Tang, S.M. Lucas,
H.M. Robertson, P. Bork, E.V. Koonin, E.M. Zdobnov, I.V. Grigoriev, M. Lynch, J.L.
Boore, The ecoresponsive genome of Daphnia pulex, Science 331 (2011) 555–561.
 M. Kohn, J. Hogel, W. Vogel, P. Minich, H. Kehrer-Sawatzki, J.A. Graves, H.
Hameister, Reconstruction of a 450-My-old ancestral vertebrate protokaryotype,
Trends Genet. 22 (2006) 203–210.
 M.A. Ferguson-Smith, V. Trifonov, Mammalian karyotype evolution, Nat. Rev.
Genet. 8 (2007) 950–962.
 W.J. Murphy, D.M. Larkin, A. Everts-van der Wind, G. Bourque, G. Tesler, L. Auvil,
J.E. Beever, B.P. Chowdhary, F. Galibert, L. Gatzke, C. Hitte, S.N. Meyers, D. Milan,
E.A. Ostrander, G. Pape, H.G. Parker, T. Raudsepp, M.B. Rogatcheva, L.B. Schook, L.C.
Skow, M. Welge, J.E. Womack, J. O'Brien S, P.A. Pevzner, H.A. Lewin, Dynamics of
mammalian chromosome evolution inferred from multispecies comparative
maps, Science 309 (2005) 613–617.
 H.H. Heng, Elimination of altered karyotypes by sexual reproduction preserves
species identity, Genome 50 (2007) 517–524.
 A.S. Wilkins, R. Holliday, The evolution of meiosis from mitosis, Genetics 181
 H.H. Heng, J.W. Chamberlain, X.M. Shi, B. Spyropoulos, L.C. Tsui, P.B. Moens,
Regulation of meiotic chromatin loop size by chromosomal position, Proc. Natl.
Acad. Sci. U.S.A. 93 (1996) 2795–2800.
 J.M. Palmer, N.P. Keller, Secondary metabolism in fungi: does chromosomal
location matter? Curr. Opin. Microbiol. 13 (2010) 431–436.
 H.H. Heng, S.A. Krawetz, W. Lu, S. Bremer, G. Liu, C.J. Ye, Re-defining the chromatin
loop domain, Cytogenet. Cell Genet. 93 (2001) 155–161.
 H.H. Heng, S. Goetze, C.J. Ye, G. Liu, J.B. Stevens, S.W. Bremer, S.M. Wykes, J. Bode, S.A.
Krawetz, Chromatin loops are selectively anchored using scaffold/matrix-attachment
regions, J. Cell Sci. 117 (2004) 999–1008.
 J. Bode, S. Goetze, H. Heng, S.A. Krawetz, C. Benham, From DNA structure to gene
expression: mediators of nuclear compartmentalization and dynamics, Chromosome
Res. 11 (2003) 435–445.
 R. Gorelick, H.H. Heng, Sex reduces genetic variation: a multidisciplinary review,
Evolution 65 (2011) 1088–1098.
 D.J. Futuyma, Evolutionary constraint and ecological consequences, Evolution 64
 K. Kitada, A. Taima, K. Ogasawara, S. Metsugi, S. Aikawa, Chromosome-specific
segmentation revealed by structural analysis of individually isolated chromosomes,
Genes Chromosomes Cancer 50 (2011) 217–227.
 J.R. Lupski, Genomic rearrangements and sporadic disease, Nat. Genet. 39 (2007)
 I.Y. Iourov, S.G. Vorsanova, Y.B. Yurov, Chromosomal mosaicism goes global, Mol
Cytogenet 1 (2008) 26.
 J.H. Bielas, K.R. Loeb, B.P. Rubin, L.D. True, L.A. Loeb, Human cancers express a
mutator phenotype, Proc. Natl. Acad. Sci. U.S.A. 103 (2006) 18238–18242.
 D.H. Erwin, E.H. Davidson,The evolution of hierarchical gene regulatory networks,
Nat. Rev. Genet. 10 (2009) 141–148.
 S. Huang, I. Ernberg, S. Kauffman, Cancer attractors: a systems view of tumors
from a gene network dynamics and developmental perspective, Semin. Cell Dev.
Biol. 20 (2009) 869–876.
 P. Ao, Global view of bionetwork dynamics: adaptive landscape, J. Genet.
Genomics 36 (2009) 63–73.
 H.H. Heng, B. Spyropoulos, P.B. Moens, FISH technology in chromosome and
genome research, Bioessays 19 (1997) 75–84.
 C. Lanctot, T. Cheutin, M. Cremer, G. Cavalli, T. Cremer, Dynamic genome
architecture in the nuclear space: regulation of gene expression in three
dimensions, Nat. Rev. Genet. 8 (2007) 104–115.
 J.J. Roix, P.G. McQueen, P.J. Munson, L.A. Parada, T. Misteli, Spatial proximity of
translocation-prone gene loci in human lymphomas, Nat. Genet. 34 (2003)
 T. Misteli, The inner life of the genome, Sci. Am. 304 (2011) 66–73.
 E. Lieberman-Aiden, N.L. van Berkum, L. Williams, M. Imakaev, T. Ragoczy, A.
Telling, I. Amit, B.R. Lajoie, P.J. Sabo, M.O. Dorschner, R. Sandstrom, B. Bernstein,
M.A. Bender, M. Groudine, A. Gnirke, J. Stamatoyannopoulos, L.A. Mirny, E.S.
Lander, J. Dekker, Comprehensive mapping of long-range interactions reveals
folding principles of the human genome, Science 326 (2009) 289–293.
 N.L. van Berkum, E. Lieberman-Aiden, L. Williams, M. Imakaev, A. Gnirke, L.A.
Mirny, J. Dekker, E.S. Lander, Hi-C: a method to study the three-dimensional
architecture of genomes, J. Vis. Exp. (2010).
 D. Bau, A. Sanyal, B.R. Lajoie, E. Capriotti, M. Byron, J.B. Lawrence, J. Dekker, M.A.
Marti-Renom, The three-dimensional folding of the alpha-globin gene domain
reveals formation of chromatin globules, Nat. Struct. Mol. Biol. 18 (2011)
 A.M. Hillmer, F. Yao, K. Inaki, W.H. Lee, P.N. Ariyaratne, A.S. Teo, X.Y. Woo, Z.
Zhang, H. Zhao, L. Ukil, J.P. Chen, F. Zhu, J.B. So, M. Salto-Tellez, W.T. Poh, K.F.
Zawack, N. Nagarajan, S. Gao, G. Li, V. Kumar, H.P. Lim, Y.Y. Sia, C.S. Chan, S.T.
Leong, S.C. Neo, P.S. Choi, H. Thoreau, P.B. Tan, A. Shaha, X. Ruan, J. Bergh, P. Hall, V.
Cacheux-Rataboul, C.L. Wei, K.G. Yeoh, W.K. Sung, G. Bourque, E.T. Liu, Y. Ruan,
Comprehensive long-span paired-end-tag mapping reveals characteristic
patterns of structural variations in epithelial cancer genomes, Genome Res.
21 (2011) 665–675.
 H.H. Heng, C.J. Ye, F. Yang, S. Ebrahim, G. Liu, S.W. Bremer, C.M. Thomas, J. Ye, T.J.
Chen, C. Tuck-Muller, J.W. Yu, S.A. Krawetz, A. Johnson, Analysis of marker or
complex chromosomal rearrangements present in pre- and post-natal karyotypes
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252
utilizing a combination of G-banding, spectral karyotyping and fluorescence in Download full-text
situ hybridization, Clin. Genet. 63 (2003) 358–367.
 C.J. Ye, J.B. Stevens, G. Liu, S.W. Bremer, A.S. Jaiswal, K.J. Ye, M.F. Lin, L. Lawrenson,
W.D. Lancaster, M. Kurkinen, J.D. Liao, C.G. Gairola, M.P. Shekhar, S. Narayan, F.R.
Miller, H.H. Heng, Genome based cell population heterogeneity promotes
tumorigenicity: the evolutionary mechanism of cancer, J. Cell. Physiol. 219
 E. Schrock, S. du Manoir, T. Veldman, B. Schoell, J. Wienberg, M.A. Ferguson-Smith,
Y. Ning, D.H. Ledbetter, I. Bar-Am, D. Soenksen, Y. Garini, T. Ried, Multicolor
spectral karyotyping of human chromosomes, Science 273 (1996) 494–497.
 C.J. Ye, W. Lu, G. Liu, S.W. Bremer, Y.A. Wang, P. Moens, M. Hughes, S.A. Krawetz,
H.H. Heng, The combination of SKY and specific loci detection with FISH or
immunostaining, Cytogenet. Cell Genet. 93 (2001) 195–202.
 J.B. Stevens, G. Liu, S.W. Bremer, K.J. Ye, W. Xu, J. Xu, Y. Sun, G.S. Wu, S. Savasan,
S.A. Krawetz, C.J. Ye, H.H. Heng, Mitotic cell death by chromosome fragmentation,
Cancer Res. 67 (2007) 7686–7694.
 J.B. Stevens, B.Y. Abdallah, S.M. Regan, G. Liu, S.W. Bremer, C.J. Ye, H.H. Heng,
Comparison of mitotic cell death by chromosome fragmentation to premature
chromosome condensation, Mol Cytogenet 3 (2010) 20.
 P.J. Stephens, C.D. Greenman, B. Fu, F. Yang, G.R. Bignell, L.J. Mudie, E.D. Pleasance,
K.W. Lau, D. Beare, L.A. Stebbings, S. McLaren, M.L. Lin, D.J. McBride, I. Varela, S.
Nik-Zainal, C. Leroy, M. Jia, A. Menzies, A.P. Butler, J.W. Teague, M.A. Quail, J.
Burton, H. Swerdlow, N.P. Carter, L.A. Morsberger, C. Iacobuzio-Donahue, G.A.
Follows, A.R. Green, A.M. Flanagan, M.R. Stratton, P.A. Futreal, P.J. Campbell,
Massive genomic rearrangement acquired in a single catastrophic event during
cancer development, Cell 144 (2011) 27–40.
 M. Meyerson, D. Pellman, Cancer genomes evolve by pulverizing single
chromosomes, Cell 144 (2011) 9–10.
 H.Q. Heng, W.Y. Chen, Y.C. Wang, Effects of pingyanymycin on chromosomes: a
possible structural basis for chromosome aberration, Mutat. Res. 199 (1988)
 H.H. Heng, J. Squire, L.C. Tsui, High-resolution mapping of mammalian genes by in
situ hybridization to free chromatin, Proc. Natl. Acad. Sci. U.S.A. 89 (1992)
 G. Rancati, N. Pavelka, B. Fleharty, A. Noll, R. Trimble, K. Walton, A. Perera, K.
Staehling-Hampton, C.W. Seidel, R. Li, Aneuploidy underlies rapid adaptive
evolution of yeast cells deprived of a conservedcytokinesis motor, Cell 135 (2008)
 C.L. Smith, A. Bolton, G. Nguyen, Genomic and epigenomic instability, fragile sites,
schizophrenia and autism, Curr Genomics 11 (2010) 447–469.
 I.Y. Iourov, S.G. Vorsanova, Y.B. Yurov, Somatic genome variations in health and
disease, Curr Genomics 11 (2010) 387–396.
 C.C. Maley, P.C. Galipeau, J.C. Finley, V.J. Wongsurawat, X. Li, C.A. Sanchez, T.G.
Paulson, P.L. Blount, R.A. Risques, P.S. Rabinovitch, B.J. Reid, Genetic clonal
diversity predicts progression to esophageal adenocarcinoma, Nat. Genet. 38
 S. Jones, W.D. Chen, G. Parmigiani, F. Diehl, N. Beerenwinkel, T. Antal, A. Traulsen,
M.A. Nowak, C. Siegel, V.E. Velculescu, K.W. Kinzler, B. Vogelstein, J. Willis, S.D.
Markowitz, Comparative lesion sequencing provides insights into tumor
evolution, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 4283–4288.
 N. Wade, Hoopla, and disappointment, in schizophrenia research, The New York
Times, The New York Times Company, New York, 2009.
 H.H. Heng,Joshua B.Stevens,Steve W.Bremer, Guo Liu,BatoulY.Abdallah, Christine
J. Ye, Evolutionary mechanisms and diversity in cancer, Adv Cancer Res (in press).
 M. Baker, Genomics: genomes in three dimensions, Nature 470 (2011) 289–294.
 L. Lipovich, R. Johnson, C.Y. Lin, MacroRNA underdogs in a microRNA world:
evolutionary, regulatory, and biomedical significance of mammalian long non-
protein-coding RNA, Biochim. Biophys. Acta 1799 (2010) 597–615.
 P. Carninci, Y. Hayashizaki, Noncoding RNA transcription beyond annotated
genes, Curr. Opin. Genet. Dev. 17 (2007) 139–144.
 M. Isalan, C. Lemerle, K. Michalodimitrakis, C. Horn, P. Beltrao, E. Raineri, M.
Garriga-Canut, L. Serrano, Evolvability and hierarchy in rewired bacterial gene
networks, Nature 452 (2008) 840–845.
 S.R. Paladugu, S. Zhao, A. Ray, A. Raval, Mining protein networks for synthetic
genetic interactions, BMC Bioinforma. 9 (2008) 426.
 D. Carter, L. Chakalova, C.S. Osborne, Y.F. Dai, P. Fraser, Long-range chromatin
regulatory interactions in vivo, Nat. Genet. 32 (2002) 623–626.
 D. Ottaviani, E. Lever, R. Mitter, T. Jones, T. Forshew, R. Christova, E.M. Tomazou,
V.K. Rakyan, S.A. Krawetz, A.E. Platts, B. Segarane, S. Beck, D. Sheer, Reconfigura-
tion of genomic anchors upon transcriptional activation of the human major
histocompatibility complex, Genome Res. 18 (2008) 1778–1786.
 E. Chevret, E.V. Volpi, D. Sheer, Mini review: form and function in the human
interphase chromosome, Cytogenet. Cell Genet. 90 (2000) 13–21.
 S. Horike, S. Cai, M. Miyano, J.F. Cheng, T. Kohwi-Shigematsu, Loss of silent-
chromatin looping and impaired imprinting of DLX5 in Rett syndrome, Nat. Genet.
37 (2005) 31–40.
 J. McClellan, M.C. King, Genetic heterogeneity in human diseases, Cell 141 (2010)
 R. Gorelick, R.M.D. Laubichler, Genetic=Heritable (Genetic≠DNA), Biol Theory 3
 J.B. Stevens, B.Y. Abdallah, G. Liu, C.J. Ye, S.D. Horne, G. Wang, S. Savasan, M.
Shekhar, S.A. Krawetz, M. Hüttemann, M.A. Tainsky, G.S. Wu, Y. Xie, K. Zhang, H.Q.
Heng, Diverse system stresses: common mechanisms of chromosome fragmen-
tation, Cell Death Disease (in press).
 P.A. Astolfi, F. Salamini, V. Sgaramella, Are we Genomic Mosaics? Variations of the
Genome of Somatic Cells can Contribute to Diversify our Phenotypes, Curr.
Genomics 11 (2010) 379–386.
 M.S. Abu-Asab, M. Chaouchi, S. Alesci, S. Galli, M. Laassri, A.K. Cheema, F. Atouf, J.
VanMeter, H. Amri, Biomarkers in the age of omics: time for a systems biology
approach, Omics 15 (2011) 105–112.
 D.R. Forsdyke, Scherrer and Jost's symposium: the gene concept in 2008, Theory
Biosci. 128 (2009) 157–161.
H.H.Q. Heng et al. / Genomics 98 (2011) 242–252