Recurrent 16p11.2 microdeletions in autism
Ravinesh A. Kumar1, Samer KaraMohamed1, Jyotsna Sudi1, Donald F. Conrad1,
Camille Brune5, Judith A. Badner4, T. Conrad Gilliam1, Norma J. Nowak6, Edwin H. Cook Jr5,
William B. Dobyns1,2,3and Susan L. Christian1,?
1Department of Human Genetics,2Department of Neurology,3Department of Pediatrics and4Department of
Psychiatry, University of Chicago, Chicago, IL 60637, USA,5Department of Psychiatry, University of Illinois at
Chicago, Chicago, IL 60612, USA and6Department of Cancer Genetics, Roswell Park Cancer Institute, Buffalo,
NY 14236, USA
Received November 21, 2007; Revised and Accepted December 19, 2007
Autism is a childhood neurodevelopmental disorder with a strong genetic component, yet the identification
of autism susceptibility loci remains elusive. We investigated 180 autism probands and 372 control subjects
by array comparative genomic hybridization (aCGH) using a 19K whole-genome tiling path bacterial artificial
chromosome microarray to identify submicroscopic chromosomal rearrangements specific to autism. We
discovered a recurrent 16p11.2 microdeletion in two probands with autism and none in controls. The deletion
spans ?500-kb and is flanked by ?147-kb segmental duplications (SDs) that are >99% identical, a common
characteristic of genomic disorders. We assessed the frequency of this new autism genomic disorder by
screening an additional 532 probands and 465 controls by quantitative PCR and identified two more patients
but no controls with the microdeletion, indicating a combined frequency of 0.6% (4/712 autism versus 0/837
controls; Fisher exact test P 5 0.044). We confirmed all 16p11.2 deletions using fluorescence in situ hybrid-
ization, microsatellite analyses and aCGH, and mapped the approximate deletion breakpoints to the edges of
the flanking SDs using a custom-designed high-density oligonucleotide microarray. Bioinformatic analysis
localized 12 of the 25 genes within the microdeletion to nodes in one interaction network. We performed phe-
notype analyses and found no striking features that distinguish patients with the 16p11.2 microdeletion as a
distinct autism subtype. Our work reports the first frequency, breakpoint, bioinformatic and phenotypic ana-
lyses of a de novo 16p11.2 microdeletion that represents one of the most common recurrent genomic dis-
orders associated with autism to date.
Autism (OMIM 209850) is a childhood neurodevelopmental
disorder apparent by 3 years of age and characterized by quali-
tative impairments in reciprocal social interaction, deficits in
verbal communication, restricted interests and repetitive beha-
viors. Autism comprises the severe end of the autism spectrum
disorders (ASD), which also include Asperger syndrome, per-
vasive developmental disorder not otherwise specified and
Rett syndrome. Prevalence rates for autism and ASD are 0.2
and 0.6%, respectively. Firm evidence of a genetic basis for
autism has been demonstrated by twin studies that show
60–91% concordance rates in monozygotic twins (1,2),
making autism the most heritable of all complex neuropsy-
chiatric disorders. The search for autism susceptibility loci
has involved several approaches including genome-wide
linkage analysis, and association and mutation studies of can-
didate genes (3,4). However, the molecular basis of autism
remains largely unknown.
A complementary approach to identify genetic suscepti-
bility variants for autism involves searching for chromosomal
abnormalities in patients. Microscopically visible chromoso-
mal rearrangements such as deletions and duplications have
been identified in 1–3% of autism subjects (5). The most
?To whom correspondence should be addressed at: Department of Human Genetics, University of Chicago, 920 East 58th Street, CLSC 319, Chicago,
IL 60637, USA. Tel: þ1 7738342971; Fax: þ1 7738348470; Email: firstname.lastname@example.org
# 2007 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License
(http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduc-
tion in any medium, provided the original work is properly cited.
Human Molecular Genetics, 2008, Vol. 17, No. 4
Advance Access published on December 21, 2007
by guest on June 1, 2013
frequent are maternal 15q11-13 duplications; other known
recurrent rearrangements include duplications of 17p12, and
deletions of 7q11.23, 17p11.2 and 22q11.2 (3). Recent high
resolution microarray-based studies in autism have identified
a number of novel submicroscopic copy number variants
(CNVs) including both deletions and duplications (6,7),
although all are rare and few are recurrent. We hypothesize
that more remain to be discovered.
We undertook a genome-wide investigation of autism-
specific CNVs in 180 autism probands and 372 control subjects
using a 19K whole-genome tiling path bacterial artificial
chromosome (BAC) microarray and identified a 16p11.2
microdeletion in two patients but no controls. We screened
an additional 532 autism probands and 465 controls by quanti-
tative PCR (qPCR), and identified two additional probands
deleted for 16p11.2. Here we provide phenotype data for
patients with microdeletions at this locus, present detailed
molecular characterization of the 16p11.2 deletion breakpoints,
and perform computational network studies of the ?25 genes
within the microdeletion. Our results demonstrate that micro-
deletions of 16p11.2 are among the most frequent recurrent
genomic disorders associated with autism to date.
We investigated 180 unrelated autism subjects and 372 control
subjects by array comparative genomic hybridization (aCGH)
using a whole-genome tiling path microarray comprised of
?19 000 BACs. In this initial study, we identified a recurrent
?500-kb microdeletion in two unrelated patients with autism
(HI0646 and HI0624) that was not detected in 372 control sub-
jects, indicating a frequency of 1.1% (2/180 autism versus 0/
372 controls; Fisher exact test P ¼ 0.11) (Fig. 1A and B).
The microdeletion has not been reported as a CNV in
healthy controls [Database of Genomic Variants (Build 36);
http://projects.tcag.ca/variation/]. We confirmed both deletions
by fluorescence in situ hybridization (FISH). Microsatellite
analyses and quantitative real-time PCR (qPCR) confirmed
paternity and showed that the deletions were de novo and
maternally derived (Fig. 1C–F). Chromosome analysis was
reported as normal for HI0624 and not listed for HI0646 on
the Autism Genetics Resource Exchange (AGRE) website
(http://www.agre.org/). A summary of our molecular studies
is presented in Table 1. We did not identify the 16p11.2 micro-
deletion in an affected sibling of HI0646, or in one affected
and one unaffected sibling of HI0624.
We assessed the frequency of the 16p11.2 microdeletion in
autism by screening an additional 532 probands and 465 con-
trols using qPCR. We identified two additional probands
(HI2467 and HI2997) and no controls deleted for 16p11.2,
giving a frequency of 0.4% in this cohort (2/532 autism
versus 0/465 controls; Fisher exact test P ¼ 0.50). The fre-
quency of the 16p11.2 microdeletion combining both cohorts
is 0.6% (4/712 autism versus 0/837 controls; Fisher exact
test P ¼ 0.044). We confirmed the new microdeletions by
FISH, microsatellite analysis and aCGH (Table 1), also con-
firming paternity (data not shown). Chromosome analysis
was reported as normal for HI2977 and not listed for
HI2467. We demonstrated that the deletion in HI2977 was
de novo and paternally derived; his affected sibling did not
carry the microdeletion. The last family proved to be more
complex; the deletion in HI2467 was paternally inherited
and apparently de novo as the deletion was not found in
either parent (or in her unaffected sister). However, her
affected brother also had the microdeletion, demonstrating
post-zygotic mosaicism for the 16p11.2 deletion and complete
association of the microdeletion with autism in this family.
The pedigree structures of all four families are shown in
Figure 2. We also identified the reciprocal duplication
product of the 16p11.2 microdeletion in one subject with
autism (HI0128), her unaffected mother, and two controls
The genomic structure and gene content of the 16p11.2
microdeletion were interrogated using the UCSC Browser
(http://genome.ucsc.edu; hg 18). Segmental duplications
(SDs) of ?147 kb and .99% identity were found to flank
each end of the microdeletion (Fig. 3A), a common character-
istic of genomic disorders. We tested the hypothesis that the
between the SDs by using a custom-designed chromosome
16p11.2 oligonucleotide array. Although the array data pro-
duced low signal-to-noise ratio due to the highly repetitive
genomic structure of our target region, we could approximate
the breakpoints in all five patients (Fig. 3B). We estimated the
telomeric and centromeric breakpoint boundaries at ?29.6
and ?30.2 Mb, respectively, which are in close proximity to
the flanking SDs, thereby supporting our hypothesis of
To better map the deletion breakpoints, we performed repli-
cate hybridization experiments in subject HI2467 and com-
bined these results using segmentation analysis (Fig. 3C).
Although complicated by the very high homology between
the SDs, these results demonstrate a sharp breakpoint at the
centromeric or internal end of the telomeric SD (arrow 1).
However, the centromeric SD has two apparent breakpoints
located near the internal (arrow 2) and external (arrow 3)
ends of the repeat with an intermediate log2ratio between
them. We interpret the intermediate log2 ratio to reflect
reduction of sequence within the SD from the normal four
copies to three copies, as many of the probes from the SD
regions have very high homology and cannot be reliably dis-
tinguished. This analysis unambiguously maps the breakpoints
of the 16p11.2 microdeletion to the edges of flanking SDs,
consistent with our previously estimated telomeric and centro-
meric breakpoint boundaries.
The microdeletion contains ?25 genes (Fig. 3A), with
another four located within the SDs. To study the potential con-
tribution of individual or sets of genes in the microdeletion to
autism, we investigated whether these genes interact bio-
logically using the Ingenuity Pathway Analysis (IPA) tool,
which uses known functional relationships and interactions
between gene products reported in the literature. Twelve of
the 25 genes were mapped to a single genetic network
(Fig. 4) that included genes involved with cell-to-cell signaling
and interaction. The pathways for 3 of the 25 genes (DOC2A,
MAPK3 and ALDOA) included post-synaptic density genes
that have been hypothesized to underlie autism (8).
We analyzed the behavioral profile on all patients with
16p11.2 microdeletions based on the phenotype data currently
Human Molecular Genetics, 2008, Vol. 17, No. 4 629
by guest on June 1, 2013
available in the AGRE database (Table 2). We found no
striking features that distinguish subjects with the 16p11.2
microdeletion as a distinct subtype of autism, although a
majority had a trend toward behavioral difficulties involving
aggression and overactivity. While all patients in the AGRE
database have had physical examinations, data regarding
facial appearance were entered for only one sibling pair.
The boy with the 16p11.2 microdeletion (HI0646) had
minor facial dysmorphism consisting of downslanting palpeb-
ral fissures, prominent ears and broad nasal root. However,
his brother with autism who did not have the microdeletion
also had minor facial dysmorphism, which suggests that the
minor facial dysmorphism may be familial rather than
related to the microdeletion.
We also reviewed available phenotype data for subjects
with duplications of 16p11.2. One girl (HI0128) with autism
and duplication 16p11.2 had severe compulsions and rituals
consistent with obsessive-compulsive disorder. However, she
also had a 1.0-Mb deletion of 7q31.2 and 890-kb duplication
of 16q22.1, making interpretation of her phenotype difficult;
Figure 1. Identification of a new recurrent de novo 16p11.2 microdeletion in autism. (A and B) BAC aCGH demonstrates the presence of a deletion at 16p11.2 in
two patients (black arrows highlight the vertical line of dots that indicate the presence of a deletion). The aCGH results show the log2ratio of the reference versus
patient DNA on the vertical axis. Each individual BAC is represented as a single blue dot. Known CNVs are represented as green dots. The horizontal axis shows
the position of each BAC along the chromosome. (C and D) Microsatellite analysis demonstrates that the deletions are de novo and maternally inherited. (E and
F) FISH results confirm the microdeletion at 16p11.2. BAC RP11-1107E4 (green arrow) and BAC RP11-114A15 (red arrow) show two copies of the green probe
and a single copy of the red probe indicating a deletion of RP11-114A15. A similar result was obtained using RP11-74E23 (red) and RP11-455F5 (green) (data
630 Human Molecular Genetics, 2008, Vol. 17, No. 4
by guest on June 1, 2013
only the 16q22.1 duplication was de novo. Her mother
(HI0126), who carries both the 7q31.2 deletion and 16p11.2
duplication, has depressive and anxiety symptoms (onset
13–19 years for both) without formal diagnosis of depression
or anxiety disorder, as well as undiagnosed learning disabil-
ities and other unspecified behavioral problems. The two
control subjects (04C34861 and 04C36902) with duplication
16p11.2 noted several minor behavioral abnormalities such
as compulsions, generalized anxiety, specific phobias and
panic attacks, but no mental health diagnoses. While both
had compulsions—defined as an inability to resist certain
activities in the presence of obsessions—the frequency was
nominally or not significant in comparison to other controls.
The present study reports the recurrent nature and first fre-
quency assessment of a novel and typically de novo 16p11.2
microdeletion that is significantly associated with autism: 4/
712 autism versus 0/837 controls; Fisher exact test P ¼
0.044. We are aware of no other studies of autism that have
Figure 2. Pedigree structure of families with 16p11.2 microdeletions. Squares indicate males and circles indicate females. Individuals affected with autism are
presented as black symbols while white symbols are unaffected. The family ID is presented at the top of each pedigree. Individual IDs are presented beneath each
symbol. Individuals with the 16p11.2 microdeletion are indicated.
Table 1. Summary of molecular studies in patients with 16p11.2 rearrangements
Individual PhenotypeRearrangementInheritanceBAC aCGH FISH MicrosatelliteqPCROligo aCGH
Maternal (de novo)
Maternal (de novo)
Paternal (de novo)
na, parental DNA not available.
2, not done.
Human Molecular Genetics, 2008, Vol. 17, No. 4631
by guest on June 1, 2013
reported a single microdeletion or duplication variant with
P , 0.10. The frequency of the 16p11.2 microdeletion
reported in this study is likely an underestimate of the actual
occurrence in autism as we tested primarily multiplex families,
which would be expected to have a higher frequency of inher-
ited and lower frequency of de novo abnormalities compared
with simplex families (6). Our study suggests that the
16p11.2 microdeletion is one of the most common recurrent
genomic disorders associated with autism.
While our study is the first to report the recurrent nature of
16p11.2 microdeletions in autism, it is not the first report of
this deletion. The 16p11.2 microdeletion was recently ident-
ified in a female with Asperger syndrome from a simplex
family, without further details regarding the phenotype (6).
A similar deletion was reported in monozygotic twins present-
ing with seizure disorder, mild mental retardation and aortic
valve anomalies (9). Both twins and a third brother without
the microdeletion had additional congenital anomalies, point-
ing to another genetic disorder segregating in the family that
may account for the aortic valve anomalies. Neither twin
had formal testing for autism, although one had more severe
expressive and receptive language deficits than expected
from his overall level of functioning, as is characteristic of
individuals with ASD (Supplementary Materials, Data S1 for
Finally, one patient with a large ?8.7-Mb deletion of
16p11.2–p12.2 that overlaps the region deleted in our patients
was recently reported (10). This deletion encompasses ?104
genes, so that a more severe phenotype would be expected.
Accordingly, she was reported to have developmental abnorm-
alities most consistent with moderate mental retardation, facial
dysmorphism including Robin sequence, multiple hand
anomalies and unexplained syncope; however, no mention of
autism was included. The additional abnormalities described
in this girl are not likely due to deletion of genes in our
16p11.2 microdeletion, as only one proband in our study
showed minor and non-specific craniofacial dysmorphism.
However, follow-up studies to obtain additional phenotype
Figure 3. Breakpoint characterization of a recurrent 16p11.2 microdeletion in autism.(A) Schematic of the 16p11.2 microdeletion. The location of the micro-
deletion is indicated on a diagram of Chromosome 16 followed by a horizontal black line representing an enlargement of the region. Large blue arrows represent
the?147-kb SDs. The genes located within this region, not drawn to scale, are listed above the black line. BACs labeled for FISH are represented by green
(overlap SDs) and red (within microdeletion) lines. Known CNVs are represented by blue lines. (B) The approximate breakpoints (dashed lines) are reported
for all five patients using a high density custom chromosome 16 oligonucleotide microarray. The short normal flanking regions outside the dashed lines contain
probes that are located mostly above the single horizontal red line, whereas the region between the dashed lines contain a large number of probes below the
horizontal line (albeit noisy data), indicating copy number loss that is consistent with the FISH analyses. The array results show the log2ratio of the reference
versus patient DNA on the vertical axis. Each individual probe is represented as a single black dot. The horizontal axis shows the position of each probe along the
chromosome. Vertical dashed lines represent the estimated telomeric and centromeric breakpoint boundaries. (C) The log2ratios from two replicate experiments
on HI2467 were averaged to create a lower noise data set. The red line shows a 1000-probe moving average of these data, while the 4 underlying black lines
depict the mean probe intensity of (from left to right): the telomeric segmental duplication (SD), the unique deleted region, the centromeric SD, and copy-normal
sequence. A sharp breakpoint at the centromeric or internal end of the telomeric SD (arrow 1) is indicated, as well as two apparent centromeric breakpoints
located near the internal (arrow 2) and external (arrow 3) ends of the repeat.
632 Human Molecular Genetics, 2008, Vol. 17, No. 4
by guest on June 1, 2013
data on these and other patients with the microdeletion are
Our data raise several important questions regarding the
16p11.2 microdeletion. First, could the deletion be non-
pathogenic, with the observed difference in frequency
between the autism and control cohorts simply a chance
finding? We think this is unlikely considering all of the data
available. The microdeletion was found in none of 837
control patients in our study, in none of the control subjects
listed in the Database of Genomic Variants (although the plat-
forms varied), and in none of 360 patients with mental retar-
dation (see below). It results in loss of at least 25 genes, a
sufficiently large number that a phenotype would be expected,
particularly considering that the list includes several good can-
didate genes (also see below). Finally, our P-value was signifi-
cant (P ¼ 0.044) when considering our entire cohort.
Next, is the 16p11.2 microdeletion specific for ASD, or can
it cause other developmental disorders as well? The deletion
was found in none of 360 patients with mental retardation
studied using a whole-genome tiling BAC array (11,12).
Several other large CNV studies of mental retardation have
been done with no deletions of 16p11.2 found, although the
platforms used included few or no probes in this region
(13–15). While further studies are needed, these studies
already suggest that the 16p11.2 microdeletion is not asso-
ciated with mental retardation, or alternatively causes mental
retardation less often than autism.
Finally, does the 16p11.2 microdeletion cause ASD by itself
or only serve as a major risk factor in combination with other
susceptibility variants? Only future studies will answer this
question, as only a few genes (or non-genetic contributing
factors) causing or contributing to autism have so far been
identified. Perhaps the microdeletion alone is sufficient to
cause mild ASD such as Asperger syndrome (6), while
additional susceptibility variants are needed to cause severe
autism. Our novel observation of de novo CNVs in patients
with autism but not in several of their affected sibs suggests
that other ASD risk factors may also be segregating in these
families. If so, then the 16p11.2 microdeletion may represent
a susceptibility variant contributing to the disorder in combi-
nation with other genetic variants shared or unshared by the
Figure 4. Functional network analysis of genes on 16p11.2. Genes are represented as nodes and the biological relationship between two nodes is represented as
an edge (line). All edges are supported by at least one reference from the literature, from a textbook or from canonical information stored in the Ingenuity Path-
ways Knowledge Base. Nodes are displayed using various shapes that represent the functional class of the gene product. Shaded nodes represent genes residing
within the microdeletion.
Human Molecular Genetics, 2008, Vol. 17, No. 4633
by guest on June 1, 2013
Considering all of these observations together, we hypoth-
esize that 16p11.2 microdeletions are a risk factor for ASD
generally, and may be causal for mild ASD in some families.
Further, our study raises interesting questions that have not
been discussed in previous reports of CNV in autism, such
as the role of genetic heterogeneity within families. Had the
current study been performed primarily in simplex families,
the discordant genotypes between affected siblings would
not have been detected. The present work emphasizes the
importance of studying multiplex as well as simplex families
for the identification of both inherited and spontaneous
For a few deletion syndromes such as 22q11.2, both the del-
etion and reciprocal duplication have been associated with
autism (16–18) (Christian et al., submitted). It is not yet
clear whether this will be true for the 16p11.2 locus as well.
The reciprocal duplication product of the 16p11.2 microdele-
tion has been reported in a mother and daughter with mild
mental retardation (19) and in two more unrelated patients
with autistic behavior (20). However, both studies used
lower resolution techniques such as karyotyping and FISH
so that the duplicated region may be larger than the
?500-kb segment highlighted in the current study. Our work
and that of others suggest that analysis of 16p11.2 duplications
should be explored further in autism as well as obsessive-
Several of the genes that reside within the 16p11.2 microde-
letion represent promising candidates for autism based on
known expression and functional data. MAPK3 is expressed
in human fetal and adult brains; Mapk32/2mice display
abnormal avoidance behavior, hyperactivity, reduced long-
term potentiation and immune system abnormalities (21).
MAZ is expressed in several tissues with the highest
expression in brain found in motor and midfrontal cortex.
MAZ directly regulates genes involved in GABA signaling
(22), neuronal differentiation (23) and the serotonin pathway
Table 2. Phenotype data for 16p11.2-deleted probands and their affected siblings
Autism Diagnostic Interview-Review-Revised (ADI-R)
Restricted, repetitive behavior domain
Age of 1st words (mos)
Age of 1st phrases (mos)
Any language regression
Any regression of other skills
History of seizures
Austism Diagnostic Observation Schedule (ADOS)
Physical measures and abormalitiesb
Birth weight (g)
Head circumference (cm)
Ear length (cm)
Functioning/Psychiatric Symptoms/Medical History
Age (years), IQ testing
Other medications (current)
Risper, Fluox, Gaba
Sex (M, male, F, female); ADI-R classification (Aut, Autism); Phrases (NA, skill not acquired); Language regression (loss of W, words, I, intent); Skill
regression (loss of skills; C, communication, S, social, A, academic, M, motor); Anxiety includes social anxiety or fears; Aggression norms (V,
visuospatial), ADOS classfication (Aut, Autism; ASO, Autism Spectrum Disorder). Percentiles for birth weight, height, and weight based on median age
(e.g. 6.5years for a 6-year-old) CDC growth charts unless percentile provided inAGREdatabase. Verbal IQ is from thePeabody Picture Vocabulary Test
(PPVT) standard score; non-verbal IQ is from the Raven’s Progressive Matrices estimated non-verbal IQ. For psychiatric symptoms, yes, presence of
symptoms. no, absence of symptoms, DX, diagnosis of disorder. Medications (methylphenidate.
aRefers to affected sibling without a deletion of 16p 11.2.
bAbnormal physical features (including congenital defects) and dysmorphic features were absent in H10646 and sibling.
634 Human Molecular Genetics, 2008, Vol. 17, No. 4
by guest on June 1, 2013
(24). DOC2A is expressed predominantly in the brain, loca-
lizes to synaptic vesicles, and is hypothesized to regulate
synaptic activity though Ca-dependent mechanisms (25), con-
sistent with the proposed role of Ca2þsignaling in autism (26).
Doc2a2/2mice display alterations in synaptic transmission
and long-term potentiation as well as learning and behavioral
deficits that include abnormal passive avoidance behavior
(27). SEZ6L2 is expressed at high levels in human brain;
Sez6l22/2mice display behavior and neurological abnormal-
ities that include impaired motor coordination and abnormal
excitatory postsynaptic currents (28). HIRIP3 is a widely
expressed gene that interacts directly with HIRA, a major can-
didate for the DiGeorge syndrome and related developmental
disorders (29). IL6, which is not located within the deletion,
was found as a node in the identified network. IL6 is known
to play a role in brain development, learning and memory,
response to disease injury in the nervous system and regulation
of the balance between neurogenesis and gliogenesis (30,31),
and has recently been identified as a key mediator of the
effects of maternal immune activation on fetal brain develop-
In summary, our work provides the first evidence of a recur-
rent 16p11.2 microdeletion in autism and ASD, and strength-
ens evidence for the involvement of 16p11.2 in other
cognitive, language and social impairments. We hypothesize
that perturbations in the network of genes identified in this
study, especially those within the microdeletion region,
underlie the autism phenotype. Our discovery of affected sib-
lings without the microdeletion supports a role for genetic het-
erogeneity in these families, and leads us to hypothesize that
the same phenomenon may be found for other autism loci
and potentially for other common complex disorders.
MATERIALS AND METHODS
The Institutional Review Board at the University of Chicago
approved this study. The original 180 AGRE samples were
simplex in 19 and multiplex in 161 families. For the additional
230 AGRE samples, 11 were simplex and 219 multiplex
families, while the 302 National Institutes of Mental Health
(NIMH) Genetics Initiative autism samples included 57
simplex and 245 multiplex families. Therefore, the total
cohort of 712 autism probands consisted of 87 simplex and
625 multiplex families. DNA from 837 control subjects was
selected from the NIMH Genetics Initiative Control sample
set, and were used for both the aCGH (n ¼ 372) and qPCR
(n ¼ 465) experiments. All control subjects were characterized
for Axis I disorders. Although individuals were characterized
for Axis I disorders, individuals meeting criteria for diagnosis
of an Axis I disorder were not excluded from the sample.
Autistic traits were not assessed in this sample.
Detailed data of all subjects with autism, their parents and
some unaffected siblings are available from the AGRE
website, and were reviewed, compiled and analyzed by the
authors (C.B. and E.H.C.). These data included detailed
medical histories and physical exams, and results from the
Autism Diagnostic Interview—Revised (33,34), the Autism
Diagnostic Observation Schedule (35) and other cognitive
testing. Data represented as missing in Table 2 was unavail-
able in the AGRE database.
BAC CGH array
The minimal tiling RPCI BAC array contains ?19 000 BAC
clones that were chosen by virtue of their STS content,
paired BAC end-sequence and association with heritable dis-
orders and cancer. The backbone of the array consists of
4603 BAC clones that were directly mapped to specific,
single chromosomal positions by FISH. Each clone was
printed in duplicate on amino-silanated glass slides type Aþ
(Schott Nexterion typeAþ) using a MicroGrid ll TAS
arrayer (Apogent Discoveries). The BAC DNA printing sol-
utions were prepared from sequence connected RPCI-11
BACs by ligation-mediated PCR as described previously
(36). The BAC DNA spots are ?80 mm in diameter with a
150 mm center-to-centerspacing
?38 000 elements. The printed slides were dried overnight
and UV-crosslinked (350 mJ) in a Stratalinker 2400 (Strata-
gene) immediately before hybridization. A complete list of
the BAC clones printed on the 19K array can be found on
the RPCI microarray facility website. Reference and test
sample genomic DNA (1 mg each) was fluorescently labeled
using the BioArray CGH Labeling System (Enzo Life
Sciences). For this study, each control DNA was used as an
independent test sample against a reference DNA pool of 20
sex-mismatched individuals, and each AGRE patient was
used as the test sample against the same sex-mismatched refer-
ence DNA pools. The DNA was denatured in the presence of
random primers at 998C for 10 min in a thermocycler, quickly
cooled to 48C and transferred onto ice. The samples were
labeled with a dNTP-cyanine 3 mix (or dNTP-cyanine 5)
and Klenow followed by incubation overnight at 378C. Unin-
corporated nucleotides were removed using a QIAquick PCR
purification column (Qiagen) and the labeled probe was
eluted with 2 ? 25 ul washes. Prior to hybridization, the test
and reference probes were resuspended in 110 ml SlideHyb
Buffer #3 (Ambion) containing 5 ml of 20 mg/ml Cot-1 DNA
and 5 ml of 100 mg/ml Yeast tRNA (Invitrogen), heated to
95˚C for 5 min and placed onto ice. Hybridization to the 19K
BAC microarrays was performed for 16 h at 558C using a
GeneTAC hybridization station (Genomic Solutions, Inc.) as
described previously (37). After hybridization, the slides
were automatically washed in the GeneTAC station with redu-
cing concentrations of SSC and SDS. The slides were scanned
using a GenePix 4200AL Scanner (Molecular Devices) to gen-
erate high-resolution (5 mm) images for both Cy3 (test) and
Cy5 (control) channels. Image analysis was performed using
the ImaGene (version 6.1.0) software from BioDiscovery,
Inc. The log2 test/control ratios were normalized using a
sub-grid Loess correction. Mapping data for each BAC was
acquired from the March 2006 version of the human genome
and added to the resulting log2test/control values. BACs over-
lapping SDs or published CNVs were flagged (38–41). To
select abnormal loci for further analysis, the aCGH data
were analyzed using two methods. (i) All clones outside of
creating an array of
Human Molecular Genetics, 2008, Vol. 17, No. 4 635
by guest on June 1, 2013
five SD were identified using software developed by Shuang
Liu. The autosomes and each sex chromosome were analyzed
separately. All pseudoautosomal BACs were included with the
autosome analyses. (ii) Manual curation was also performed to
identify loci with multiple abnormal clones that were detected
as a single clone using a threshold of five SD. All loci with two
contiguous abnormal BACs were then screened against the
control data. Only CNVs not present in the controls were
Oligonucleotide CGH arrays and breakpoint analyses
We characterized the breakpoints using a NimbleGen Systems
(Madison, WI) custom oligonucleotide Chr. 16 array (HG18;
NCBI build36) covering
designed under the following specifications: isothermal
probes (50–74 bp) with target Tm of 76; synthesis cycle
limit of 148 cycles; filtered at average 15-mer frequency of
50; interval spacing at 2 bp (mean) and 3 bp (median). We
also employed a fine tiling commercial design for chr16
from human genome (hg18; NCBI Build 36) designed as
above except that interval spacing was 158 bp (mean) and
165 bp (median). All hybridizations were performed as pre-
viously described (42). Our estimates of the telomeric break-
point boundaries were consistent with the data generated
from a lower density Chromosome 16 microarray (Nimble-
Gen); however, the centromeric boundary could not be accu-
rately recovered due to a paucity of probes in this region on
the lower resolution array (data not shown). A Gaussian
mixture model approach was used to identify change points
between two adjacent segments of different copy number.
Briefly, a large window was manually designed that spanned
an equal number of probes in each segment. The model was
then constructed as a mixture of two probe sets, with the
mixture proportions determined by the (unknown) location
of the breakpoint. Data from the probes spanning the interval
from the left edge of the design window to the breakpoint
parameterize one of the mixture distributions; the other set
of probes, the other distribution. The change point was then
identified by evaluating the likelihood of the data across a
grid of points (one point per interval between probes),
re-estimating the other parameters in the model at each
point. The maximum likelihood estimates for the three
change points apparent in Figure 1H are: 29 557 173–
29 559 112(changepoint1);
30 253 847–30 253 852 (change 3). All coordinates are rela-
tive to NCBI 36. Strikingly, these estimated change points cor-
respond closely to the edges of known SDs: a 147-kb
duplication on the centromeric flank of the deletion covers
the sequence from 30 107 356–30 254 364. The other half of
this pair is located at 29 368 017–29 514 353; however,
there is a third, 22-kb duplication from 29 534 625–
29 557 497, whose centromeric edge corresponds exactly
with change point 1 identified here. It is possible that a
CNV within this 22-kb SD is creating additional changes
between HI2467 and the reference, thus complicating the
analysis, or that the 22-kb SD is directly involved in NAHR
with the centromeric 147-kb SD. Moving averages of the Nim-
blegen data were constructed with the ‘caTools’ package
29 318 000–30 304 400 Mb
30 107 181 (change2);
written in the ‘R’ programming language (http://cran.rpro-
Quantitative real-time PCR
qPCR was performed using two TaqManwprimer/probe
amplicon sets designed against the 16p11.2 microdeletion
region (Assays-by-Design; Applied Biosystems) as well as a
reference primer/probe set (TaqManwRNaseP gene, Applied
Biosystems). Sequences for the 16p11.2-region primers and
50-CCTCTCTCTTCCCCACAAAGG-30; SEZ6L2-R, 50-TG
GACAGCCTGGTTCTCTCT-30; SEZ6L2-probe, 50- CCTC
ACC-30; C16orf54-probe, 50-CTTCCCAGGCTCC-30. Reac-
tions were run in triplicate in 6 ul total volume containing
40 ng genomic DNA, 2X TaqMan Universal Master Mix
(ABI part number 4304437) and 20X primer/probe mix.
Each assay included a no-template control and three controls
subjects each validated by FISH for 16p11.2 copy number
(i.e. one, two and three copies, respectively). We repeated
qPCR for every subject showing copy number alterations at
16p11.2 by re-running each sample an additional six times
for the target and reference probes. Each experiment was per-
formed using a 384-well optical PCR plate and the 7900 HT
Real-Time PCR System (Applied Biosystems) under the fol-
lowing cycling conditions: initial step at 508C for 2 min, dena-
turation at 958C for 10 min, followed by 40 cycles at 958C for
15 s. Data analysis was performed using the SDS v2.1 soft-
ware0;s Relative Quantitation feature. Copy number of the
16p11.2 target region was defined as 2T
the difference threshold cycle number for the test and refer-
ence loci. For each subjects showing copy number alterations
at the target loci, we confirmed our finding by re-running each
sample six times for the target and reference probes.
2DDC, where DCTis
Fluorescence in situ hybridization
Lymphoblastoid cell lines for each proband and selected
family members were acquired from the Rutgers University
Cell and DNA Repository and cultured using standard tech-
niques. RPCI-11 BACs that defined the boundaries of the
selected CNVs were acquired from several sources including
the Wellcome Trust Sanger Institute and the Roswell Park
Cancer Institute. BAC probes were labeled by nick translation
using either Spectrum Green or Spectrum Orange fluorescent
dyes (Abbott Labs). FISH was performed using standard tech-
niques. Slides were analyzed with a Zeiss Axioplan 2 fluor-
escent microscope with a cooled CCD camera and Applied
Imaging CytoVision v3.7 software.
Microsatellites were selected from the UCSC Genome
Browser microsatellite or simple repeat tracks and primers
were designed using the MIT Primer3 program. For a single
reaction, a master mix of 1 ml 10? PCR buffer with 15 mM
MgCl (2), 1 ml 10 mM dNTP, 0.1 ml Ampli Taq Gold
enzyme, 0.8 ml 10 mM primers (forward and reverse) and
636Human Molecular Genetics, 2008, Vol. 17, No. 4
by guest on June 1, 2013
7.1 ml sterile H2O was prepared. One microliter DNA (10 ng/
ml) was added to each reaction. The PCR reaction was run in
ABI 9700 thermocyclers using the following conditions: hot
start at 948C for 10 min, 948C for 30 s, 558C for 30 s, 728C
for 30 s for 35 cycles followed by a final extension step at
728C for 10 min. Samples were analyzed on an ABI 3730
XL DNA sequencing analyzer and processed using GeneMap-
per 3.7 software (Applied Biosystems).
Gene accession numbers for our ‘focus’ genes (i.e. the 25
genes residing within the 16p11.2 microdeletion) were
imported into the IPA software (Systems, Mountain View,
CA, USA; https://analysis.ingenuity.com/pa), a web-delivered
application that enables discovery, visualization and explora-
tion of molecular interaction networks. Genetic networks
were ranked by a score that takes into account the number
of focus genes and size of networks. The probability that a col-
lection of genes equal to or greater than the number in a
network could be achieved by chance alone was calculated.
A score of three indicates that there is a 1/1,000 chance that
the focus gene(s) are in a network due to random chance.
Scores of three or higher have a 99.9% confidence level of
not being generated by random chance alone and was used
as the cut-off for identifying gene networks.
To assess the significance of the difference of the frequency of
the 16p11.2 microdeletion between autistic subjects and con-
trols, the Fisher Exact Test was used, as implemented at
Supplementary Material is available at HMG Online.
We gratefully acknowledge the resources provided by the
Autism Genetic Resource Exchange (AGRE) Consortium
and the participating AGRE patients and families. Additional
autism families and the control samples were acquired as
part of the NIMH Center for Collaborative Genetic Studies
on Mental Disorders.
Conflict of Interest statement. None declared.
This study was supported by the National Institutes of Health,
National Institute of Neurological Diseases and Stroke
(1R01NS51812 to S.L.C.) and National Institute of Mental
Health (1R01MH64547-01 to T.C.G), the National Alliance
for Autism Research to S.L.C., and an Autism Speaks Post-
Doctoral Fellowship Award to E.H.C. Jr. and C.B. AGRE is
supported, in part, by a grant to D.H.G. from the National
Institute of Mental Health (NIMH) (MH64547).
1. Bailey, A., Le Couteur, A., Gottesman, I., Bolton, P., Simonoff, E.,
Yuzda, E. and Rutter, M. (1995) Autism as a strongly genetic disorder:
evidence from a British twin study. Psychol. Med., 25, 63–77.
2. Steffenburg, S., Gillberg, C., Hellgren, L., Andersson, L., Gillberg, I.C.,
Jakobsson, G. and Bohman, M. (1989) A twin study of autism in
Denmark, Finland, Iceland, Norway and Sweden. J. Child Psychol.
Psychiatry, 30, 405–416.
3. Veenstra-VanderWeele, J. and Cook, E.H. Jr. (2004) Molecular genetics
of autism spectrum disorder. Mol. Psychiatry, 9, 819–832.
4. Freitag, C.M. (2007) The genetics of autistic disorders and its clinical
relevance: a review of the literature. Mol. Psychiatry, 12, 2–22.
5. Vorstman, J.A., Staal, W.G., van Daalen, E., van Engeland, H.,
Hochstenbach, P.F. and Franke, L. (2006) Identification of novel autism
candidate regions through analysis of reported cytogenetic abnormalities
associated with autism. Mol. Psychiatry, 11, 18–28.
6. Sebat, J., Lakshmi, B., Malhotra, D., Troge, J., Lese-Martin, C., Walsh, T.,
Yamrom, B., Yoon, S., Krasnitz, A., Kendall, J. et al. (2007) Strong
association of de novo copy number mutations with autism. Science, 316,
7. Autism Genome Project ConsortiumSzatmari, P., Paterson, A.D.,
Zwaigenbaum, L., Roberts, W., Brian, J., Liu, X.Q., Vincent, J.B., Skaug,
J.L., Thompson, A.P. et al. (2007) Mapping autism risk loci using genetic
linkage and chromosomal rearrangements. Nat. Genet., 39, 319–328.
8. Geschwind, D.H. and Levitt, P. (2007) Autism spectrum disorders:
developmental disconnection syndromes. Curr. Opin. Neurobiol., 17,
9. Ghebranious, N., Giampietro, P.F., Wesbrook, F.P. and Rezkalla, S.H.
(2007) A novel microdeletion at 16p11.2 harbors candidate genes for
aortic valve development, seizure disorder, and mild mental retardation.
Am. J. Med. Genet. A, 143, 1462–1471.
10. Ballif, B.C., Hornor, S.A., Jenkins, E., Madan-Khetarpal, S., Surti, U.,
Jackson, K.E., Asamoah, A., Brock, P.L., Gowans, G.C., Conway, R.L.
et al. (2007) Discovery of a previously unrecognized microdeletion
syndrome of 16p11.2-p12.2. Nat. Genet., 39, 1071–1073.
11. de Vries, B.B., Pfundt, R., Leisink, M., Koolen, D.A., Vissers, L.E.,
Janssen, I.M., Reijmersdal, S., Nillesen, W.M., Huys, E.H., Leeuw, N.
et al. (2005) Diagnostic genome profiling in mental retardation.
Am. J. Hum. Genet., 77, 606–616.
12. Koolen, D.A., Vissers, L.E., Pfundt, R., de Leeuw, N., Knight, S.J.,
Regan, R., Kooy, R.F., Reyniers, E., Romano, C., Fichera, M. et al. (2006)
A new chromosome 17q21.31 microdeletion syndrome associated with a
common inversion polymorphism. Nat. Genet., 38, 999–1001.
13. Fan, Y.S., Jayakar, P., Zhu, H., Barbouth, D., Sacharow, S., Morales, A.,
Carver, V., Benke, P., Mundy, P. and Elsas, L.J. (2007) Detection of
pathogenic gene copy number variations in patients with mental
retardation by genomewide oligonucleotide array comparative genomic
hybridization. Hum. Mutat., 28, 1124–1132.
14. Friedman, J.M., Baross, A., Delaney, A.D., Ally, A., Arbour, L.,
Armstrong, L., Asano, J., Bailey, D.K., Barber, S., Birch, P. et al. (2006)
Oligonucleotide microarray analysis of genomic imbalance in children
with mental retardation. Am. J. Hum. Genet., 79, 500–513.
15. Wagenstaller, J., Spranger, S., Lorenz-Depiereux, B., Kazmierczak, B.,
Nathrath, M., Wahl, D., Heye, B., Glaser, D., Liebscher, V., Meitinger, T.
et al. (2007) Copy-number variations measured by
single-nucleotide-polymorphism oligonucleotide arrays in patients with
mental retardation. Am. J. Hum. Genet., 81, 768–779.
16. Vorstman, J.A., Morcus, M.E., Duijff, S.N., Klaassen, P.W., Heineman-de
Boer, J.A., Beemer, F.A., Swaab, H., Kahn, R.S. and van Engeland, H.
(2006) The 22q11.2 deletion in children: high rate of autistic disorders and
early onset of psychotic symptoms. J. Am. Acad. Child Adolesc.
Psychiatry, 45, 1104–1113.
17. Simon, T.J., Bish, J.P., Bearden, C.E., Ding, L., Ferrante, S., Nguyen, V.,
Gee, J.C., McDonald-McGinn, D.M., Zackai, E.H. and Emanuel, B.S.
(2005) A multilevel analysis of cognitive dysfunction and
psychopathology associated with chromosome 22q11.2 deletion syndrome
in children. Dev. Psychopathol., 17, 753–784.
18. Mukaddes, N.M. and Herguner, S. (2007) Autistic disorder and 22q11.2
duplication. World J. Biol. Psychiatry, 8, 127–130.
19. Engelen, J.J., de Die-Smulders, C.E., Dirckx, R., Verhoeven, W.M.,
Tuinier, S., Curfs, L.M. and Hamers, A.J. (2002) Duplication of
Human Molecular Genetics, 2008, Vol. 17, No. 4 637
by guest on June 1, 2013
chromosome region (16)(p11.2 –. p12.1) in a mother and daughter with Download full-text
mild mental retardation. Am. J. Med. Genet., 109, 149–153.
20. Finelli, P., Natacci, F., Bonati, M.T., Gottardi, G., Engelen, J.J., de
Die-Smulders, C.E., Sala, M., Giardino, D. and Larizza, L. (2004) FISH
characterisation of an identical (16)(p11.2p12.2) tandem duplication in
two unrelated patients with autistic behaviour. J. Med. Genet., 41, e90.
21. Mazzucchelli, C., Vantaggiato, C., Ciamei, A., Fasano, S., Pakhotin, P.,
Krezel, W., Welzl, H., Wolfer, D.P., Pages, G., Valverde, O. et al. (2002)
Knockout of ERK1 MAP kinase enhances synaptic plasticity in the
striatum and facilitates striatal-mediated learning and memory. Neuron,
22. Murakami, A., Matsuda, M., Nakasima, A. and Hirata, M. (2006)
Characterization of the human PRIP-1 gene structure and transcriptional
regulation. Gene, 382, 129–139.
23. Okamoto, S., Sherman, K., Bai, G. and Lipton, S.A. (2002) Effect of the
ubiquitous transcription factors, SP1 and MAZ, on NMDA receptor
subunit type 1 (NR1) expression during neuronal differentiation. Brain
Res. Mol. Brain Res., 107, 89–96.
24. Parks, C.L. and Shenk, T. (1996) The serotonin 1a receptor gene contains
a TATA-less promoter that responds to MAZ and Sp1. J. Biol. Chem.,
25. Groffen, A.J., Friedrich, R., Brian, E.C., Ashery, U. and Verhage, M.
(2006) DOC2A and DOC2B are sensors for neuronal activity with unique
calcium-dependent and kinetic properties. J. Neurochem., 97, 818–833.
26. Krey, J.F. and Dolmetsch, R.E. (2007) Molecular mechanisms of autism: a
possible role for Ca2þ signaling. Curr. Opin. Neurobiol., 17, 112–119.
27. Sakaguchi, G., Manabe, T., Kobayashi, K., Orita, S., Sasaki, T., Naito, A.,
Maeda, M., Igarashi, H., Katsuura, G., Nishioka, H. et al. (1999)
Doc2alpha is an activity-dependent modulator of excitatory synaptic
transmission. Eur. J. Neurosci., 11, 4262–4268.
28. Miyazaki, T., Hashimoto, K., Uda, A., Sakagami, H., Nakamura, Y.,
Saito, S.Y., Nishi, M., Kume, H., Tohgo, A., Kaneko, I. et al. (2006)
Disturbance of cerebellar synaptic maturation in mutant mice lacking
BSRPs, a novel brain-specific receptor-like protein family. FEBS Lett.,
29. Llevadot, R., Scambler, P., Estivill, X. and Pritchard, M. (1996) Genomic
organization of TUPLE1/HIRA: a gene implicated in DiGeorge
syndrome. Mamm. Genome, 7, 911–914.
30. Bauer, S., Kerr, B.J. and Patterson, P.H. (2007) The neuropoietic cytokine
family in development, plasticity, disease and injury. Nat. Rev. Neurosci.,
31. He, F., Ge, W., Martinowich, K., Becker-Catania, S., Coskun, V., Zhu,
W., Wu, H., Castro, D., Guillemot, F., Fan, G. et al. (2005) A positive
autoregulatory loop of Jak-STAT signaling controls the onset of
astrogliogenesis. Nat. Neurosci., 8, 616–625.
32. Smith, K.M., Bauer, L., Fischer, M., Barkley, R. and Navia, B.A. (2005)
Identification and characterization of human NR4A2 polymorphisms in
attention deficit hyperactivity disorder. Am. J. Med. Genet. B
Neuropsychiatr. Genet., 133, 57–63.
33. Lord, C., Rutter, M. and Le Couteur, A. (1994) Autism Diagnostic
Interview-Revised: a revised version of a diagnostic interview for
caregivers of individuals with possible pervasive developmental disorders.
J. Autism Dev. Disord., 24, 659–685.
34. Lord, C., Risi, S., Lambrecht, L., Cook, E.H. Jr., Leventhal, B.L.,
DiLavore, P.C., Pickles, A. and Rutter, M. (2000) The autism diagnostic
observation schedule-generic: a standard measure of social and
communication deficits associated with the spectrum of autism. J. Autism
Dev. Disord., 30, 205–223.
35. Lord, C., Rutter, M., Goode, S., Heemsbergen, J., Jordan, H., Mawhood,
L. and Schopler, E. (1989) Autism diagnostic observation schedule: a
standardized observation of communicative and social behavior.
J. Autism. Dev. Disord., 19, 185–212.
36. Nowak, N.J., Gaile, D., Conroy, J.M., McQuaid, D., Cowell, J., Carter, R.,
Goggins, M.G., Hruban, R.H. and Maitra, A. (2005) Genome-wide
aberrations in pancreatic adenocarcinoma. Cancer Genet. Cytogenet., 161,
37. Cowell, J.K., Wang, Y.D., Head, K., Conroy, J., McQuaid, D. and Nowak,
N.J. (2004) Identification and characterisation of constitutional
chromosome abnormalities using arrays of bacterial artificial
chromosomes. Br. J. Cancer, 90, 860–865.
38. Sharp, A.J., Locke, D.P., McGrath, S.D., Cheng, Z., Bailey, J.A., Vallente,
R.U., Pertz, L.M., Clark, R.A., Schwartz, S., Segraves, R. et al. (2005)
Segmental duplications and copy-number variation in the human genome.
Am. J. Hum. Genet., 77, 78–88.
39. Tuzun, E., Sharp, A.J., Bailey, J.A., Kaul, R., Morrison, V.A., Pertz, L.M.,
Haugen, E., Hayden, H., Albertson, D., Pinkel, D. et al. (2005) Fine-scale
structural variation of the human genome. Nat. Genet., 37, 727–732.
40. Iafrate, A.J., Feuk, L., Rivera, M.N., Listewnik, M.L., Donahoe, P.K., Qi,
Y., Scherer, S.W. and Lee, C. (2004) Detection of large-scale variation in
the human genome. Nat. Genet., 36, 949–951.
41. Sebat, J., Lakshmi, B., Troge, J., Alexander, J., Young, J., Lundin, P.,
Maner, S., Massa, H., Walker, M., Chi, M. et al. (2004) Large-scale copy
number polymorphism in the human genome. Science, 305, 525–528.
42. Selzer, R.R., Richmond, T.A., Pofahl, N.J., Green, R.D., Eis, P.S., Nair,
P., Brothman, A.R. and Stallings, R.L. (2005) Analysis of chromosome
breakpoints in neuroblastoma at sub-kilobase resolution using fine-tiling
oligonucleotide array CGH. Genes Chromosomes Cancer, 44, 305–319.
638Human Molecular Genetics, 2008, Vol. 17, No. 4
by guest on June 1, 2013