GENETIC DISORDERS (JF CUBELLS AND EB BINDER, SECTION EDITORS)
The Genetics of Autism Spectrum Disorders – A Guide
Karsten M. Heil & Christian P. Schaaf
# Springer Science+Business Media New York 2012
Abstract Recent advances in genetic testing technology
have made chromosome microarray analysis (CMA) a
first-tier clinical diagnostic test for Autism Spectrum
Disorders (ASDs). Two main types of microarrays are
available, single nucleotide polymorphism (SNP) arrays
and array comparative genomic hybridization (aCGH),
each with its own advantages and disadvantages in ASDs
testing. Rare genetic variants, and copy number variants
(CNVs) in particular, have been shown to play a major
role in ASDs. More than 200 autism susceptibility genes
have been identified to date, and complex patterns of
inheritance, such as oligogenic heterozygosity, appear
to contribute to the etiopathogenesis of ASDs. Incom-
plete penetrance and variable expressivity represent par-
ticular challenges in the interpretation of CMA testing
of autistic individuals. This review aims to provide an
overview of autism genetics for the practicing physi-
cian and gives hands-on advice on how to follow-up on
abnormal CMA findings in individuals with neuropsychiatric
Keywords Autism spectrum disorders.ASDs.Autism
While there is uncertainty about the exact magnitude, a high
heritability for Autism Spectrum Disorders (ASDs) has long
been recognized because of the results of twin studies .
Despite this, genetic testing – and chromosomal testing
more specifically – have not played prominent roles in the
evaluation of individuals with ASDs until very recently. The
advances of technological sophistication in genetic testing
in the past decade have fundamentally changed the utiliza-
tion of genetic diagnostic tools in the clinical assessment of
ASDs. Today, chromosome microarray analysis (CMA)
technology is recommended as a first-tier clinical diagnostic
test for ASDs [2•].
Very importantly, it should be pointed out from the onset
that genetic testing does not serve the purpose of making the
initial diagnosis of ASDs. Unlike most classic heritable
disorders, the etiopathogenesis of ASDs, and psychiatric
conditions in general, is too multifactorial and our knowl-
edge of the genetic mechanisms involved still too limited for
genetic testing to effectively aid in making the initial diag-
nosis. The diagnosis of ASDs, therefore, remains a clinical
Once the diagnosis has been established, however, the
question about the underlying cause typically arises. The use
This article is part of the Topical Collection on Genetic Disorders
K. M. Heil
Faculty of Medicine, University of Heidelberg,
Im Neuenheimer Feld 134b,
69120 Heidelberg, Germany
C. P. Schaaf (*)
Department of Molecular and Human Genetics,
Baylor College of Medicine,
Jan and Dan Duncan Neurological Research Institute,
1250 Moursund St, Suite 1350,
Houston, TX 77030, USA
Curr Psychiatry Rep (2013) 15:334
of genetic testing now helps to elucidate some of that,
helping both the treating physicians and caregivers to better
understand the etiology unique to a particular patient. Mak-
ing a “genetic” diagnosis in an individual with ASDs can
serve some of the following purposes: (1) Having a genetic
diagnosis can help families gain access to additional resour-
ces through their insurance company or school district, (2)
connect with other families affected by the same genetic
disorder via family support groups and online communities,
(3) decrease parental feelings of guilt, (4) provide better
anticipatory guidance based on literature for a specific un-
derlying genetic condition, (5) better estimate the chance of
recurrence for future children and other family members, (6)
take account of specific management guidelines, if avail-
able, and (7) access specific treatment considerations, if
available. For example, in cases of syndromic autism other
organ systems are commonly affected. Prior knowledge of
the location and nature of expected problems can increase
the chance of effective early treatment .
First, this review aims to provide a non-technical over-
view of (1) the current genetic testing technologies used to
clinically investigate the causes of ASDs, (2) important
findings these investigations have yielded so far, and (3)
their importance for our understanding of the etiogenesis of
ASDs. Second, we aim to offer hands-on advice for the
practicing clinician on how to best conduct and interpret
genetic testing for their patients with ASDs.
Genetic Testing Technologies Used to Investigate ASDs
In order to appreciate the rapid progress in genetic testing
technologies over the past decades, it is useful to compare
their resolutions. G-banded karyotypes yield only about
1,400 to 1,700 bands per diploid genome . Modern
microarrays yield anywhere from hundreds of thousands
to – when used in combination – over 150 million data
points [6, 7]. The diploid human genome has roughly 6
billion base pairs , all potentially readable via whole-
genome sequencing. The differences in resolution between
the technologies largely determine their scope and limita-
tions. Although microarrays have a much higher resolution
than light microscopic karyotypes, there are a few genomic
aberrations (e.g., balanced translocations, low-level mosai-
cism) that are generally not detectable by array. These are
very infrequent causes of intellectual disability or ASDs,
The varying levels of resolution possible with different
technologies can be difficult to explain to non-scientifically
trained patients. Clinical geneticists and genetic counselors
have found that it helps to use analogies to aid in explana-
tion (e.g., the genome viewed as a book with two volumes,
many chapters, pages, and letters) .
There are several kinds of microarrays and a multitude of
sequencing methods, which makes the situation even more
complex. As microarray analysis is currently the clinically
most useful tool for genetic testing in ASDs, it will be the
focus of the following discussion. The prospective future
role of sequencing will be addressed in our outlook at the
end of this review.
Currently, the most important distinction among genomic
microarrays is that between array comparative genomic
hybridization (aCGH) and single-nucleotide polymorphism
In aCGH, patient DNA and a control DNA sample are
compared to each other (“comparative”). The array consists
of a compartmentalized panel with specific oligonucleotide
probes attached to each of the thousands to millions of
compartments. Each oligonucleotide is representative of a
particular section of the human genome. Ideally, the sum of
all of the oligonucleotides on the array represents the entire
genome. In practice, many arrays contain representative
portions of the genomic sequence, for example, probe
sequences spaced evenly across the genome or arranged so
that gene-rich regions are more densely represented. First,
the patient and control DNAs are labeled with fluorescent
dyes of different colors (red and green in most cases).
Second, equal concentrations of the two samples are hybrid-
ized simultaneously to the array. If patient and control DNA
have the same copy number (same amount of genomic
material) at a specific locus, their fluorescent signals will
be equally strong. Whenever there are missing or extra
copies in the patient DNA, this will be detected as an
imbalance in the fluorescent signal. Because the oligonu-
cleotides can be made small enough to represent individual
exons in genes, aCGH can achieve what is referred to as
exon-by-exon coverage (SNP arrays cannot always achieve
this level of resolution). So, unlike in traditional karyotype
analysis, it is not only possible to investigate deletions or
duplications of large segments of chromosomes, but even
changes on a sub-gene level are within the scope of aCGH
. The normal copy number is two for chromosomes 1-22,
two for the X-chromosome in females, and one each for the
X and Y chromosome in males. Any deviation from that
“normal” copy number (normal in a statistical sense) would
be considered a copy number variation or copy number
variant (CNV). Loss of one copy is considered a deletion,
and gain of one copy would be considered a duplication
(and sometimes there are greater numbers of extra copies of
genomic regions: triplications, etc). As the next section of
this review will aim to convey, this difference in dosage –
rather than sequence – has recently been shown to often
have profound consequences and play an important role in
the understanding of neuropsychiatric disorders.
The second type of array technology is the SNP array. In
SNP arrays only one DNA sample – that of the patient – is
334, Page 2 of 8Curr Psychiatry Rep (2013) 15:334
used. As in aCGH, SNP arrays work by hybridizing DNA to
an array of compartmentalized oligonucleotides. However,
in SNP arrays, each compartment contains two specific
oligonucleotide sequences (instead of one). Depending on
the nucleotide present at the specific location of the sample
DNA, the section of DNAwill hybridize to one or the other
oligonucleotide sequence in the compartment. Again the
resulting fluorescent signature of the compartment yields
the result. This time the comparison is not made between a
standardized genome and the patient’s genome, but between
two possible sequences of a precise section of the genome.
Sequence here only means the single varying “letter” (SNP)
at this specific genomic location and must not be confused
with the sequence of an entire exon or even gene. SNP
arrays can only detect single-base variations included in
the array. Hence, their possible presence must be known in
advance. More than 150 million SNP locations have been
documented for the human genome so far . In principle,
each one of these SNPs could be tested for with a SNP array.
In practice, each array has a limited number of SNPs it can
test for, though this number can be in the millions.
Despite this limitation, there are millions of SNPs that
SNP arrays do already include. The pattern in which
these SNPs occur in a patient’s genome can yield useful
information. As in aCGH, SNP arrays can detect CNVs
(although not always exon-by-exon). Additionally, they
can detect uniparental disomy (UPD) and identity by descent
If the SNP pattern on a chromosome – or smaller segment
of the genome – exactly matches that of the homologous
chromosome or segment, it is termed absence of heterozy-
gosity (AOH). This can occur because of both homologous
segments having been inherited from the same parent (uni-
parental disomy, UPD) or because of shared ancestry or
identity by descent. When large fractions of the genome
display AOH, a close familial relationship between the
individual’s parents has to be considered. The degree of that
relationship can be estimated (e.g., offspring of a mating
between first degree relatives would lead to AOH in approx-
imately one quarter of the entire genome) .
Both array technologies are limited in their ability to
define repetitive areas of the genome as well as pseudo-
genes. And while these areas make up a large part of the
genome, little is known to date about their functional im-
portance . Further advances in technology are necessary
to uncover their true role, as well as their variability among
In summary, aCGH has the advantage that it can be
designed to cover the coding genome on an exon-by-exon
level, while only SNP arrays can detect UPD or identity by
descent. In many settings, the patient’s insurance policy will
only cover a certain type of array sent to a specific labora-
tory. When several options are available, it is advisable to
choose the array with the highest resolution, and the labo-
ratory that provides the most comprehensive interpretation,
ideally helping the ordering physician to put molecular
findings into a clinical context [2•].
Some advice on how to interpret the results of chromo-
some microarrays in the context of ASDs will be provided
later in this review.
The Role of CNVs and Other Rare Variants in ASDs
In classic monogenic disorders (e.g., Huntington disease)
usually both the variant that causes the disorder as well as
the disorder itself are rare. Psychiatric disorders, like schizo-
phrenia, bipolar disorder, and ASDs, are not only more
common, but also genetically more complex, which makes
causation far more difficult to disentangle. When consider-
ing the genetic basis of any complex disorder, it is useful to
distinguish between two models of causation: the common
variant common disease (CVCD) and rare variant common
disease (RVCD) models. There has been considerable de-
bate about the relative merits of these two models [12, 13].
Although the models are not necessarily mutually exclusive
(a common disorder could be influenced by common and
rare variants at the same time), the approach used to inves-
tigate still depends on the assumed model. In this review we
focus on rare variants, which does not mean to argue that
common variants do not play a significant role in the etio-
genesis of ASDs.
Recent microarray studies suggest that 7–20 % of idio-
pathic cases of ASDs are due to CNVs identifiable with
these arrays [14•]. Some important examples include:
Jacquemont et al.  found clinically relevant CNVs in
28 % (8/29) of patients that were previously found to
have a normal G-banded karyotype.
Sebat et al.  found de novo CNVs in 10 % (12/118) of
children from simplex families, 3 % (2/77) in multiplex
families and 1 % (2/196) in controls.
Marshall et al.  found de novo CNVs in 7 % (4/57)
of children with sporadic autism.
Glessner et al.  identified several new susceptibility
genes involved in neuronal cell adhesion and ubiquitin
pathways with high statistical significance in a group of
859 cases and 1,409 healthy controls.
Shen et al.  found clinically significant CNVs in 7 %
(59/848) of cases.
In an extensive literature review of 33 studies including
22,698 patients overall, the International Standard Cyto-
genomic Array Consortium found that CMA offered a
diagnostic yield of 15-20 % versus 3 % for G-banded
karyotypes in patients with idiopathic ASDs or intellec-
tual disability (ID) [2•].
Curr Psychiatry Rep (2013) 15:334 Page 3 of 8, 334
Sanders et al.  found de novo CNVs in 6 % (62/1124)
of simplex cases versus 2 % (15/872) in unaffected
Levy et al.  found de novo CNVs in 8 % (68/858) of
simplex cases versus 2 % (17/863) in unaffected siblings.
These studies clearly suggest an important role of CNVs
in the genetic etiology of ASDs. For several loci, the over-
representation of CNVs in autistic individuals vs. controls
has been replicated by multiple studies. The clinical pheno-
types and genotype-phenotype correlations are being char-
acterized. Table 1 provides an overview of some of the
CNVs most strongly associated with ASDs in the literature
Each individual CNV is rare and accounts for fewer than
1 % of all cases of ASDs. In addition, most of these CNVs
are not found in individuals with ASDs exclusively, but also
occur in unaffected controls, albeit with much lower fre-
quency. Hence, any single one of them is not a fully pene-
trant ASDs-causing variant in itself (unlike the completely
penetrant mutation causing Huntington disease, for instance,
where everyone carrying the mutation will get the disease if
he or she lives long enough). This phenomenon of incom-
plete penetrance is typical for genetically heterogeneous
complex disorders. A better analogy than Huntington dis-
ease would therefore be cancer-predisposing genetic
changes – e.g., BRCA1 or BRCA2 mutations in breast
cancer – where the genetic variant confers a predisposition
that significantly changes probabilities but does not determine
disease development all by itself. Here, the genetic variant is
onlythe firsthitina twoormultiple hitprocessthatultimately
leads to disease. In a similar way, the current thinking on the
role of genetics in ASDs is that the degree of mutational
burden (in copy number or sequence) is an important deter-
minant of the risk that a given individual carries .
How do copy number and sequence variants interact
with other loci in the genome and the environment to
produce the phenotype seen in the disorder? There is a
growing body of evidence suggesting that a multiple hit
model involving a relatively small number of genes in any
given patient might explain a substantial part of the dis-
order in that patient [23–25]. These results support the
previously made suggestion that autism is a complex
genetic disorder resulting from simultaneous genetic var-
iations in a few, several, or even multiple genes .
However, little is known about the actual molecular
underpinnings of these oligogenic events. Future work
will have to investigate how oligogenic variants interact
to produce the actual clinical and behavioral phenotype.
An outlook on this avenue of research is provided in the
last section of this review.
Another important concept in the context of the genetics
of ASDs, and psychiatric genetics more generally, is vari-
able expressivity. In the field of genetics, expressivity refers
to the variations in the phenotype among individuals with a
given genotype. While penetrance relates to the probability
of an individual with a given genotype to manifest clinically
at all, expressivity refers to the variability in the spectrum of
symptoms. In the field of neuropsychiatric genetics, there is
large variability in expressivity for most CNVs associated
with ASDs to date. Even within the same family, a CNV
may manifest with ASD in one individual, schizophrenia in
another, and yet bipolar disorder and anxiety in a third. As
seen in Table 1, most CNVs associated with ASDs to date
also associate with other cognitive and psychiatric pheno-
types [26, 27, 28•].
Table 1 CNVs associated with Autism Spectrum Disorders
Locus ASDs candidate
Dup/Del Odds ratio for
ASDs (95 % CI)
Additional clinical features reported
in some affected individuals*
8.0 (3.5 – 18.4)
High for exonic
30.0 (1.9 – 480.4)
ID, SCZ, BD
Macrocephaly, dysmorphic facial features
Hypotonia, seizures, macrocephaly
[20, 30, 48, 49]
[18, 30, 50]NRXN1
3q29VIPR2Del ID, SCZ, BDTapering fingers, mild dysmorphic
Obesity, seizures, macrocephaly
Growth retardation, hypotonia
Congenital heart disease, palatal
[20, 30, 51]
30.7 (3.4 – 275.1)
42.6 (15.7 – 115.5)
10.8 (3.5 – 33.1)
11.8 (6.1 – 22.7)
9.5 (5.2 – 17.4)
3.3 (1.6 – 6.6)
High (not seen in
ID, SCZ, BD
ID, SCZ, BD
[20, 30, 52]
[18, 20, 30]
[18, 20, 30, 53]
[18, 20, 30, 53]
[18, 20, 30]
[20, 30, 48]
*As cited in www.ncbi.nlm.nih.gov/sites/GeneTests/review or www.orpha.net. Otherwise specified
CI 0 confidence interval, Dup 0 duplication, Del 0 deletion, ID 0 intellectual disability, SCZ 0 schizophrenia, BD 0 bipolar disorder
334, Page 4 of 8Curr Psychiatry Rep (2013) 15:334
Due to all of the above-mentioned considerations, there
is a good case for the outstanding role of rare genetic
variants in the etiogenesis of ASDs. Because of the indi-
vidual rarity of each of these variants, a truly comprehen-
sive evaluation of their role is limited by the ability to
investigate such variants in both a systematic way and on
a very large scale, ideally involving tens of thousands of
patients and controls.
Potential Pitfalls and Practical Advice
In Fig. 1, we suggest a testing algorithm that we hope will be
useful in clinical practice. Once a clinical diagnosis of
autism spectrum disorder is established, genetic testing
should be initiated. If CNVs are found on the array test,
it is critical to determine whether the deleted or duplicated
locus is associated with ASDs or not. One of the best and
most comprehensive online resources is provided by the
Navigating to “CNV” and then “Explore Copy Number Var-
iants” yields a long searchable list of documented CNVs in-
volved in ASDs. The “# of studies” column provides a first
rough clue on the amount of evidence that this CNV might be
involved in ASDs. For each CNV detailed information is
available, including a list of all studies that referenced the
CNV as well as different representations of data gathered on
CNVs that have not been previously associated with ASDs
(unclassified variants, UVs) are challenging to interpret.
There are some general conceptsto assess their pathogenicity:
a) If the CNV is de novo (not present in either parent) or
only occurs in affected individuals within the family, a
causal relationship is more likely. A CNV that also
occurs in healthy parents is thought of as less likely to
be causally linked to the condition of the patient (al-
though that conclusion begs the questions of variable
penetrance or expressivity we discussed above—a real
challenge for very rare variants).
b) The larger the CNV and the more genes included in
the deleted or duplicated segment, the more likely it is
to lead to phenotypic consequences. The test report
should indicate which and how many genes are locat-
ed in the interval of the CNV. If only coordinates are
provided, the ambitious provider may manually check,
using free online tools, e.g., UCSC Genome Browser
c) In general, and for most loci, deletions are thought to
have stronger effects than their reciprocal duplications.
The identificationof CNVs ofuncertain significance can be
a diagnostic and counseling challenge. When communicating
Fig. 1 Proposed algorithm for the genetic evaluation of individuals with Autism Spectrum Disorders
Curr Psychiatry Rep (2013) 15:334 Page 5 of 8, 334
should be pointed out that to a certain degree genetic variation
is normal and part of human diversity.
Whenever the ordering provider is uncertain as to how to
proceed at any step of the testing process or on how to
interpret the results, we suggest two possible ways to solicit
help: (1) liaising with the laboratory that conducted the test
and/or (2) referral to a clinical geneticist. One can search for
board-certified clinical geneticists in the US and Canada by
location on the American Board of Medical Genetics web-
As illustrated above, a substantial body of evidence links
specific rare genetic variants to ASDs. Where will the
field go from here? The search for more associations as
well as more and ever more rare variants will continue.
Yet, this approach is inherently limited to the discovery of
correlations while the nature of causation remains obscure.
Elucidating the mechanisms involved will require different
Bioinformatic approaches are beginning to put genetic
findings into a more functional context. Gilman et al. used
network-based analysis of genetic associations and identi-
fied a large biological network of genes affected by rare de
novo CNVs in autism [29••]. The genes forming the net-
work were not functionally independent, but related to cer-
tain biological functions and pathways, such as synapse
development, axon guidance, and neuron motility. This
suggests that not only the dose matters, but also the specific
combination of rare variants and how they relate to certain
specific fundamental molecular processes. This result pro-
vides further evidence for the involvement of specific neuro-
developmental pathways in the pathogenesis of ASDs that
had also been found in previous studies [18, 30, 31].
In particular, there is accumulating evidence for an im-
portant role of synaptic genes and proteins in the pathophys-
iology of ASDs. Huda Zoghbi first proposed in 2003 that
autism resulted from disruption of postnatal or experience-
dependent synaptic plasticity, based on findings in Rett
syndrome, which can manifest autism spectrum disorder as
one of its phenotypes, and some rare cases of autism caused
by mutations in the gene encoding neuroligin-4 . The
hypothesis was subsequently supported by the identification
of mutations (both CNVs and point mutations) in multiple
genes coding for synaptic cell adhesion molecules, in par-
ticular the presynaptic neurexins (NRXN1 , NRXN2
, NRXN3 ) and their postsynaptic ligands, the neuro-
ligins (NLGN1 , NLGN 3, NLGN4X , and NLGN4Y
), as well as the Shank family of scaffolding proteins at
the postsynaptic density (SHANK1 , SHANK2 ,
SHANK3 ). Functional studies, including the generation
of animal models, have shown that altering the levels of
these proteins alters morphology, function, and plasticity of
synapses [41, 42]. Importantly, it appears that many of the
associated phenotypes can be reversed once the level of the
respective protein has been restored [43••, 44], providing
hopefor the potential treatabilityofautism spectrumdisorders
Outlook and Conclusions
In the final part of this review, we want to focus on possible
future developments in genetic testing technologies and
their implications for rare variant testing in ASDs.
One parameter of constant improvement is the resolution
of genetic tests. Microarrays have already drastically im-
proved their resolution and continue to do so. At the end of
this process, whole genome sequencing will produce a level
of resolution down to the single base pair level.
In 2011, whole exome sequencing was introduced as a
clinical test in the US. The exome is defined as that part of
the genome that is formed by exons, the portions of genes
that encode messenger RNA (most of which then is trans-
lated into protein). It makes up only about 1 % of the entire
genome. Currently, the main advantage of whole-exome
over whole-genome sequencing is lower cost. As the cost
of next-generation sequencing techniques continues to drop,
whole-genome sequencing will become affordable and is
expected to replace whole exome sequencing within the
next few years.
The upside of this process is that variants can be identified
that were too small or too rare to be identified by previous
technologies. On the downsidethiswill– particularlyatfirst –
also add more and more unclassified variants. Genetic testing
is progressing at a rate that far outstrips the collection of
clinical data that can be meaningfully connected to the enor-
mous amounts of raw genomic data produced. Even when
larger amounts of clinical data are available, many more
professionals trained in bioinformatics than are currently
available will be needed before these clinical data can be
integrated into practice [45–47].
It is therefore of utmost importance that clinicians –
particularly in the field of psychiatry – are involved in the
gathering of solid, comprehensive, and well-structured clin-
ical data. Without this essential cooperation the advances in
technology will produce a better understanding of the struc-
ture and variation of the human genome, but its functional
and pathophysiological implications will necessarily remain
concealed. As a first step to help better integrate clinical and
molecular data, we urge clinicians ordering genetic diagnos-
tic tests to be very diligent about providing a specific and
detailed indication for testing. The more clinical information
334, Page 6 of 8 Curr Psychiatry Rep (2013) 15:334
that is conveyed to the laboratory, the better they will be
able to interpret their findings and the more they can help
putting these back into a clinical context.
Ultimately, it will take very large-scale projects to help us
fully understand the complexity of the genetics of ASDs.
Then, functional studies need to translate genomic findings
into molecular mechanisms. Fortunately, in the case of ASDs
there is hope and even emerging evidence that this will result
in more molecular targets for drugs, better diagnostics, and
personalized treatments for the benefit of our patients.
the Joan and Stanford Alexander family and the Ting Tsung and Wei
Fong Chao Foundation. Dr. Schaaf is the recipient of a Doris Duke
Clinical Scientist Development Award.
Dr. Schaaf ’s work is generously supported by
Department of Molecular and Human Genetics at Baylor College of
Medicine, which derives revenue from chromosomal microarray analysis
K.M. Heil: none; C.P. Schaaf is a faculty member of the
Papers of particular interest, published recently, have been
• Of importance
•• Of major importance
1. Smalley SL, Asarnow RF, Spence MA. Autism and genetics. Arch
Gen Psychiatry. 1988;45:953–61.
2. • Miller DT, Adam MP, Aradhya S, et al. Consensus statement:
chromosomal microarray is a first-tier clinical diagnostic test for
individuals with developmental disabilities or congenital anoma-
lies. Am J Hum Genet. 2010;86(5):749–64. A meta-analysis of
chromosome microarray testing in individuals with intellectual
disability, congenital anomalies, and autism spectrum disorders.
3. El-Fishawy P, State MW. The genetics of autism: key issues, recent
findings, and clinical implications. Psychiatr Clin North Am.
4. Bassett AS, McDonald-McGinn DM, Devriendt K, et al. Practical
guidelines for managing patients with 22q11.2 deletion syndrome.
J Pediatr. 2011;159(2):332–9.
5. Bayani J, Squire JA. Traditional banding of chromosomes for cyto-
genetic analysis. Curr Protoc Cell Biol. 2004. Chapter 22:Unit 22.3.
6. Schaaf CP, Wisznewska J, Beaudet AL. Copy Number and SNP
arrays in clinical diagnostics. Annu Rev Genomics Hum Genet.
7. National Center for Biotechnology Information (NCBI): NCBI
dbSNP build 137 for Human. Available at http://www.ncbi.nlm.
Accessed July 2012.
8. National Center for Biotechnology Information (NCBI): Human
genome assembly information for GRCh37.p9. Available at http://
data/index.shtml. Accessed July 2012.
9. Lubitz RJ, Komaromy M, Crawford B, et al. Development and
pilot evaluation of novel genetic educational materials designed
for an underserved patient population. Genet Test. 2007;11
10. Schaaf CP, Scott DA, Wiszniewska J, Beaudet AL. Identification
of incestuous parental relationships by SNP-based DNA micro-
arrays. Lancet. 2011;377(9765):555–6.
11. Santangelo AM, de Souza FS, Franchini LF, et al. Ancient exapta-
tion of a CORE-SINE retroposon into a highly conserved mam-
malian neuronal enhancer of the proopiomelanocortin gene. PloS
12. McClellan J, King MC. Genetic heterogeneity in human disease.
13. Klein RJ, Xu X, Mukherjee S, et al. Successes of genome-wide
association studies. Cell. 2010;142:350–1. author reply 353-5.
14. • Miles JH. Autism spectrum disorders – a genetics review. Genet
Med. 2011;13(4):278–94. An excellent, up-to-date review on autism
15. Jacquemont ML, Sanlaville D, Redon R, et al. Array-based com-
parative genomic hybridisation identifies high frequency of cryptic
chromosomal rearrangements in patients with syndromic autism
spectrum disorders. J Med Genet. 2006;43:843–9.
16. Sebat J, Lakshmi B, Malhotra D, et al. Strong association of de
novo copy number mutations with autism. Science. 2007;316:445–
17. Marshall CR, Noor A, Vincent JB, et al. Structural variation of
chromosomes in autism spectrum disorder. Am J Hum Genet.
18. Glessner JT, Wang K, Cai G, et al. Autism genome-wide copy
number variation reveals ubiquitin and neuronal genes. Nature.
19. Shen Y, Dies KA, Holm IA, et al. Clinical genetic testing for
patients with autism spectrum disorders. Pediatrics. 2010;125(4):
20. Sanders SJ, Ercan-Sencicek AG, Hus V, et al. Multiple recurrent de
novo CNVs, including duplications of the 7q11.23 Williams syn-
drome region, are strongly associated with autism. Neuron. 2011;70
21. Levy D, Ronemus M, Yamrom B, et al. Rare de novo and trans-
mitted copy-number variation in autistic spectrum disorders. Neu-
22. Welberg L. Autism: The importance of getting the dose right. Nat
Rev Neurosci. 2011;12(8):429.
23. Schaaf CP, Sabo A, Sakai Y. Oligogenic heterozygosity in individ-
uals with high-functioning autism spectrum disorders. Hum Mol
24. Leblond CS, Heinrich J, Delorme R. Genetic and functional anal-
yses of SHANK2 mutations suggest a multiple hit model of autism
spectrum disorders. PLoS Genet. 2012;8(2):e1002521.
25. Girirajan S, Rosenfeld JA, Cooper GM, et al. A recurrent 16p12.1
microdeletion supports a two-hit model for severe developmental
delay. Nat Genet. 2010;42(3):203–9.
26. Sebat J, Levy DL, McCarthy SE. Rare structural variants in
schizophrenia: one disorder, multiple mutations; one mutation,
multiple disorders. Trends Genet. 2009;25(12):528–35.
27. Sullivan PF, Magnusson C, Reichenberg A, et al. Family history of
schizophrenia and bipolar disorder as risk factors for autism. Arch
Gen Psychiatry. 2012;2:1–5.
28. • Malhotra D, Sebat J. CNVs: harbingers of a rare variant revolu-
tion in psychiatric genetics. Cell. 2012;148(6):1223–41. An excel-
lent review on the role of genetics in the etiopathogenesis of
29. •• Gilman SR, Iossifov I, Levy D, et al. Rare de novo variants
associated with autism implicate a large functional network of
genes involved in formation and function of synapses. Neuron.
2011;70(5):898–907. A bioinformatics approach to the complex
landscape of autism genetics.
30. Pinto D, Pagnamenta AT, Klei L, et al. Functional impact of global
rare copy number variation in autism spectrum disorders. Nature.
Curr Psychiatry Rep (2013) 15:334 Page 7 of 8, 334
31. Bourgeron T. A synaptic trek to autism. Curr Opin Neurobiol. Download full-text
32. Zoghbi HY. Postnatal neurodevelopmental disorders: meeting at
the synapse? Science. 2003;302(5646):826–30.
33. Feng J, Schroer R, Yan J, et al. High frequency of neurexin 1beta
signal peptide structural variants in patients with autism. Neurosci
34. Gauthier J, Siddiqui TJ, Huashan P, et al. Truncating mutations in
NRXN2 and NRXN1 in autism spectrum disorders and schizo-
phrenia. Hum Genet. 2011;130(4):563–73.
35. Vaags AK, Lionel AC, Sato D, et al. Rare deletions at the neurexin
3 locus in autism spectrum disorder. Am J Hum Genet. 2012;90
36. Jamain S, Quach H, Betancur C, et al. Mutations of the X-linked
genes encoding neuroligins NLGN3 and NLGN4 are associated
with autism. Nat Genet. 2003;34(1):27–9.
37. Ylisaukko-oja T, Rehnström K, Auranen M, et al. Analysis of four
neuroligin genes as candidates for autism. Eur J Hum Genet.
38. Sato D, Lionel AC, Leblond CS, et al. SHANK1 Deletions in Males
with Autism Spectrum Disorder. Am J Hum Genet. 2012;90(5):879–
39. Berkel S, Marshall CR, Weiss B, et al. Mutations in the SHANK2
synaptic scaffolding gene in autism spectrum disorder and mental
retardation. Nat Genet. 2010;42(6):489–91.
40. Durand CM, Betancur C, Boeckers TM, et al. Mutations in the
gene encoding the synaptic scaffolding protein SHANK3 are as-
sociated with autism spectrum disorders. Nat Genet. 2007;39
41. Wang X, McCoy PA, Rodriguiz RM, et al. Synaptic dysfunction
and abnormal behaviors in mice lacking major isoforms of Shank3.
Hum Mol Genet. 2011;20(15):3093–108.
42. Chao HT, Zoghbi HY, Rosenmund C. MeCP2 controls excitatory
synaptic strength by regulating glutamatergic synapse number.
43. •• Guy J, Gan J, Selfridge J, et al. Reversal of neurological defects
in a mouse model of Rett syndrome. Science. 2007;315
(5815):1143–7. A proof-of-concept study that postnatal neurological
disorders may be rescuable.
44. Dölen G, Osterweil E, Rao BS, et al. Correction of fragile X
syndrome in mice. Neuron. 2007;56(6):955–62.
45. Foo JN, Liu JJ, Tan EK. Whole-genome and whole-exome se-
quencing in neurological diseases. Nat Rev Neurol. 2012 Jul 31.
46. Lyon GJ, Wang K. Identifying disease mutations in genomic
medicine settings: current challenges and how to accelerate prog-
ress. Genome Med. 2012;4(7):58.
47. Majewski J, Rosenblatt DS. Exome and whole-genome sequencing
for gene discovery: the future is now! Hum Mutat. 2012;33
48. Bucan M, Abrahams BS, Wang K, et al. Genome-wide anal-
yses of exonic copy number variants in a family-based study
point to novel autism susceptibility genes. PLoS Genet. 2009;5(6):
49. Brunetti-Pierri N, Berg JS, Scaglia F, et al. Recurrent reciprocal
1q21.1 deletions and duplications associated with microcephaly or
macrocephaly and developmental and behavioral abnormalities.
Nat Genet. 2008;40(12):1466–71.
50. Schaaf CP, Boone PM, Sampath S et al. Phenotypic spectrum and
genotype-phenotype correlations of NRXN1 exon deletions. Eur J
Hum Genet. 2012 May 23.
51. Willatt L, Cox J, Barber J, et al. 3q29 microdeletion syndrome:
clinical and molecular characterization of a new syndrome. Am J
Hum Genet. 2005;77(1):154–60.
52. Mervis CB, Dida J, Lam E, et al. Duplication of GTF2I results in
separation anxiety in mice and humans. Am J Hum Genet. 2012;90
53. Shinawi M, Liu P, Kang SH, et al. Recurrent reciprocal 16p11.2
rearrangements associated with global developmental delay, be-
havioral problems, dysmorphism, epilepsy, and abnormal head
size. J Med Genet. 2010;47(5):332–41.
334, Page 8 of 8Curr Psychiatry Rep (2013) 15:334