Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 May 12.
Published in final edited form as: Hum Genet. 2022 Oct 17;142(2):217–230. doi: 10.1007/s00439-022-02496-z

Common genetic risk factors in ASD and ADHD co-occurring families

Anbo Zhou 1, Xiaolong Cao 1,5, Vaidhyanathan Mahaganapathy 1, Marco Azaro 1, Christine Gwin 1, Sherri Wilson 1, Steven Buyske 2, Christopher W Bartlett 3, Judy F Flax 1, Linda M Brzustowicz 1,4, Jinchuan Xing 1,4
PMCID: PMC10177627  NIHMSID: NIHMS1885860  PMID: 36251081

Abstract

Autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) are two major neurodevelopmental disorders that frequently co-occur. However, the genetic mechanism of the co-occurrence remains unclear. The New Jersey Language and Autism Genetics Study (NJLAGS) collected more than 100 families with at least one member affected by ASD. NJLAGS families show a high prevalence of ADHD and provide a good opportunity to study shared genetic risk factors for ASD and ADHD. The linkage study of the NJLAGS families revealed regions on chromosomes 12 and 17 that are significantly associated with ADHD. Using whole-genome sequencing data on 272 samples from 73 NJLAGS families, we identified potential risk genes for ASD and ADHD. Within the linkage regions, we identified 36 genes that are associated with ADHD using a pedigree-based gene prioritization approach. KDM6B (Lysine Demethylase 6B) is the highest-ranking gene, which is a known risk gene for neurodevelopmental disorders, including ASD and ADHD. At the whole-genome level, we identified 207 candidate genes from the analysis of both small variants and structure variants, including both known and novel genes. Using enrichment and protein–protein interaction network analyses, we identified gene ontology terms and pathways enriched for ASD and ADHD candidate genes, such as cilia function and cation channel activity. Candidate genes and pathways identified in our study improve the understanding of the genetic etiology of ASD and ADHD and will lead to new diagnostic or therapeutic interventions for ASD and ADHD in the future.

Introduction

Autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) are two neurodevelopmental disorders (NDDs) of high prevalence and severity. ASD is characterized by deficits in social interaction and social communication, and by restricted, repetitive, and stereotyped patterns of behavior, interests, and activities (American Psychiatric Association 2013). ADHD is characterized by inattention and hyperactive/ impulsive symptoms (American Psychiatric Association 2013). The two disorders have shown a high comorbid frequency where 20–50% of children with ADHD meet the criteria for ASD and 30–80% of ASD children meet the criteria for ADHD (Rommelse et al. 2010). Both conditions can have a severe negative impact on patients’ quality of life, and only worsen when co-occurring.

On the behavioral level, and especially regarding diagnosis and clinical care, it is often difficult to distinguish some of the behavioral characteristics that define ASD to those that define ADHD. Individual with ASD often suffer from symptoms consistent with ADHD. Lack of focus, impulsivity, and emotional outbursts may be present in both disorders yet the approach to intervention, both behaviorally and pharmacologically, can be substantially different. Clarifying the genetic relationship between these two disorders could inform future treatment decisions. If ASD and ADHD represent pleiotropic manifestations of the same disorder, a unified treatment approach might be effective to target the symptoms of both. If they stem from different etiologies, separate treatments for each might prove more effective. Therefore, a deep understanding of the genetic etiology of familial co-occurrence of ASD and ADHD is important to enhance treatment planning.

Extensive research has been conducted on both disorders and demonstrates overlapping genetic factors between the two conditions (Rommelse et al. 2010; Doherty and Owen 2014; Ghirardi et al. 2018; Johnson et al. 2015; Polderman et al. 2014; Ronald and Hoekstra 2011; Stergiakouli et al. 2017; Liu et al. 2020). For example, ADHD candidate causal genes DRD3 and MAOA are cautiously positively associated with ASD (reviewed in (Rommelse et al. 2010)). In genome-wide association studies (GWAS), 16 single-nucleotide variants (SNVs) related to ADHD were found to be possibly involved in ASD and 25 SNVs related to ASD are possibly involved in ADHD (reviewed in (Rommelse et al. 2010)). In copy number variation (CNV) studies, CNVs segregated with ADHD were also found to be enriched for ASD candidate genes. Correspondingly, ASD patients’ family members also carry ADHD diagnosed CNVs (Elia et al. 2010; Martin et al. 2014). While the studies have suggested shared genetic risk factors as one possible reason for the co-occurrence of the two disorders, they either used microarray or whole-exome sequencing data. In addition, those studies are based on the case–control study design and do not leverage the power of family studies. Therefore, the evidence of shared genetic risk factors between ASD and ADHD from these studies is mostly weak and inconclusive.

To further explore the genetic etiology of ASD and ADHD co-occurrence, we aimed to use whole-genome sequencing (WGS) analysis to identify SNVs and structural variants (SVs) responsible for the ASD or ADHD phenotypes segregated in the New Jersey Language and Autism Genetics Study (NJLAGS) project (Bartlett et al. 2014). NJLAGS is a project studying the genetic etiology of ASD and related NDDs and has conducted WGS on 272 samples from 73 families, each with at least one ASD patient. All family members have been characterized by extensive behavioral assessments of language and social functioning, restrictive and repetitive behaviors, and other co-occurring behaviors (Bartlett et al. 2014). Despite the original study’s focus on ASD, these families exhibit elevated rates of ADHD, with more than 40% of the ASD patients also diagnosed with ADHD. The co-occurrence of ASD and ADHD in the NJLAGS families provides an opportunity to identify genetic factors underlying the co-occurring ASD and ADHD.

Patients and methods

Sample collection and phenotype assessment

The sample in this study includes 79 families from a previous NJLAGS study (Bartlett et al. 2014) (Wave 1) plus an additional 32 families collected after the 2014 publication (Wave 2). The original aim of the NJLAGS study was to identify genetic variation related to DSM-IV autistic disorder, language impairment, as well as related disorders. Families were required to have at least one person with autistic disorder and one additional person with a language-learning impairment (Bartlett et al. 2014). During Wave 1, autism probands were required to have a strict diagnosis of autistic disorder based on the Autism Diagnostic Interview (ADI-R), Autism Diagnostic Observation Schedule (ADOS), and the Diagnostics and Statistical Manual-IV (DSM-IV) (American Psychiatric Association 2000). A second proband had to meet the criteria for Specific Language Impairment (SLI), a disorder where language development is delayed or deviant and cannot be explained by any other neurodevelopmental diagnosis. For Wave 2, NJLAGS sought to understand if the original strict autistic disorder diagnostic criteria were necessary, and therefore, the autism proband requirements were intentionally relaxed to become more in line with the less restrictive, newer DSM-5 criteria for Autism Spectrum Disorder (American Psychiatric Association 2013). All subjects gave informed consent or assent conforming to the guidelines for treatment of human subjects from the Institutional Review Board at Rutgers, The State University of New Jersey (IRB number: 13–112Mc).

For ASD probands, ADHD characteristics were identified during the intake interview by either a study child psychiatrist or a developmental pediatrician. In each case, the diagnosis was corroborated in detailed NJLAGS history questionnaires that asked the parents if their child had ever received a diagnosis of attention-deficit disorder (ADD) and/or ADHD: 1) Medical History Questionnaire, 2) Family History Questionnaire, and 3) Language Correlates Questionnaire. For all other family members, parents completed the same detailed Family History Questionnaire and Language Correlates Questionnaire regarding whether they or their children have received a diagnosis of several different neurodevelopmental or neuropsychiatric disorders, including ADD/ADHD. In all questionnaires ADD and ADHD were treated as one category and we used ADHD as a collective term for the affected individual.

Genotyping and sequencing

DNA extraction was performed by RUCDR (Piscataway, NJ), either from blood DNA (WB) or minimally passaged (3–6 times) lymphoblastoid cell lines (LCL). Wave 1 Affymetrix Axiom array genotyping experiments have been described previously (Bartlett et al. 2014). For Wave 2, single-nucleotide polymorphism (SNP) data were generated with the Illumina Infinium PsychArray-24 v1 array (Illumina, San Diego, CA), which includes 593,260 SNPs. SNP with a population minor allele frequency (MAF) > 1% were selected. Quality control on SNP genotypes was conducted by array batch and by array type, as described previously (Bartlett et al. 2014). The quality control criteria include individual/SNP genotype completion, relationship checking, Mendelian errors checking, and ancestry inference. The linkage analysis only included samples that clustered with the CEU (Utah residents with Northern and Western European ancestry from the CEPH collection) samples from the HapMap reference data, as determined by EIGENSTRAT using the recommend parameters in the documentation (Patterson et al. 2006).

Initially, a subset of 10,899 SNPs in common across the Wave 1 and 2 arrays was selected for the linkage analysis, which minimize marker-to-marker linkage disequilibrium, while retaining a high MAF (> 30%) to provide suitable genomic coverage of recombination events in the pedigrees. Overlap in some genomic regions was too low to retain acceptable information content as measured by MERLIN (Abecasis et al. 2002). In those regions, array-specific SNPs were included. This procedure did not pose an issue with missing data within families, since every family was genotyped using only a single array type.

The whole-genome sequencing was performed in four batches by three vendors (Table S1). All samples were sequenced using Illumina paired-end format with a spec of 30 × coverage of the human genome. Same as the linkage analysis, the WGS analysis only included samples that clustered with CEU samples from the HapMap reference data. Some samples were excluded in downstream analyses because of quality issues. A few were dropped because the subjects withdrew from the study. For samples that were sequenced in more than one batch, the best quality run was used for the analysis. The raw sequencing reads, variants, and genotypes for all samples are available in the National Institute of Mental Health Data Archive (NDA) under experiments C1932 and C2933.

Linkage analysis

Linkage analyses were conducted with KELVIN (v2.3.3). KELVIN implements the posterior probability of linkage (PPL) metric to estimate the probability that a genetic location is linked with a tested trait (Vieland et al. 2011). Primary linkage analysis of the phenotypes was conducted on each wave separately, and the linkage evidence was sequentially updated across the waves using Bayes’ rule to provide a single metric for linkage evidence. A secondary linkage analysis was conducted using all families jointly in a single pooled analysis of each trait. By comparing the sequentially updated result to the pooled result, the role of heterogeneity in the dataset can be inferred qualitatively. Because stratifying on an irrelevant trait will, on average, produce the same result as a pooled analysis (Govil and Vieland 2008), if the sequentially updated PPLs were appreciably higher than pooled results, heterogeneity demarcated by wave is likely to present in the data.

Based on the previous simulations of the null distribution in the NJLAGS sample when correcting for three phenotypes (Bartlett et al. 2014), a PPL of 0.32 or greater is consistent with a genome-wide error rate of p < 0.001, a PPL of 0.26 corresponds to p < 0.01, and a PPL of 0.11 corresponds to p < 0.05. These threshold values are similar to previous studies of the false-positive rate of the PPL after correction for testing multiple phenotypes (Bartlett et al. 2002; Logue et al. 2003).

Small variant (SNV/indel) and SV calling

Alignment of paired-end FASTQ files was performed using the BWA-MEM algorithm (v0.7.12) (Li and Durbin 2009) to the Human Genome Reference Consortium Build 37 (hg19) using default parameters. The output was converted to BAM format using SAMtools view (v0.1.19) (Li et al. 2009). BAM files from read alignment were then processed using the GATK v3.5.0 variant calling pipeline following the best practice recommendation for alignment processing and variant calling (DePristo et al. 2011; Wang and Xing 2013). Starting from sorted and indexed individual BAM files, a series of GATK alignment processing procedures were conducted, including PCR duplicate removal and base quality score recalibration. SNVs and small insertion/deletions (indels) were called per individual using HaplotypeCaller before being jointly called by GenotypeGVCF. All samples from different sequencing batches were jointly called along with the 1000 Genomes project European ancestry samples (CEU, GBR, FIN) as controls to reduce the batch effect for downstream analysis (Sudmant et al. 2015). After variant call, we employed variant quality score recalibration using VariantRecalibrator and ApplyRecalibration as outlined in the GATK protocol.

SV calling has been historically difficult for tools using a single algorithm (Ho et al. 2019). Therefore, an ensemble algorithm, MetaSV (Mohiyuddin et al. 2015), along with its components (Breakseq2 (Abyzov et al. 2015), break-dancer (Chen et al. 2009), CNVnator (Abyzov et al. 2011), and Pindel (Ye et al. 2009)), was used for SV discovery. Local realignment by SPAdes (Bankevich et al. 2012) and AGE (Abyzov and Gerstein 2011) was carried out to further improve breakpoint resolution. MetaSV then combined all the produced evidence into a single call set. The outputs of different chromosomes from the same individual were then merged back to one file using VCFtools (Danecek et al. 2011). Basic statistics for the SV calls were calculated by SURVIVOR (Sedlazeck et al. 2017) for quality control. SVs were then filtered to select SVs that had consensus calls from at least two SV callers in MetaSV.

Variant annotation and selection

SNVs and indels were annotated by VAT in the VAAST package (v2.0.2) (Hu et al. 2013) and condensed into CDR files to represent one family per file by VST in the VAAST package. The variants were filtered to include those that have an MAF < 5% in the ExAC dataset excluding psychiatric cohorts (ExAC non-psych v0.3.1 https://gnomad.broadinstitute.org/ ). For control samples, 635 GTEx WGS samples were obtained from the GTEx project (dbGaP accession number phs000424.v7.p2) (GTEx Consortium 2020) and condensed into a single group. SV calls were annotated by AnnotSV (v3.0.9) (Geoffroy et al. 2018).

Gene prioritization

For SNVs and indels, the gene prioritization tool pVAAST (v0.02) was used to prioritize candidate genes (Hu et al. 2014). A pVAAST score was calculated for each gene from its variants’ linkage pattern, association evidence, MAF, and functional prediction. pVAAST analyses were performed for the two ADHD linkage regions and the whole genome, under both dominant and recessive modes of inheritance. The linkage region analyses were performed at 106 permutations per gene and the whole-genome analyses were performed at 105 permutations per gene to determine the significance. For running pVAAST, the pedigrees were processed and trimmed to include individuals from one pair of ancestral parents per pedigree for the dominant mode, or a two-generation subset of the pedigree for the recessive mode. The individuals were selected to maximize the number of sequenced and affected samples in each family.

Candidate genes expressed in at least one of the brain expression databases were selected for downstream analysis. For linkage regions, candidate genes that had a Bonferroni-corrected p-value < 0.05 were considered significant and selected. For whole-genome analyses, candidate genes that have a p-value < = 10−5 and a LOD score > = 1 in the pVAAST analysis were selected for downstream analysis.

For SVs, AnnotSV annotation provides a ranking score for potential pathogenicity (Geoffroy et al. 2018). SVs that are predicted to be likely pathogenic (class 4) or pathogenic (class 5) were selected. In each individual, the severity score of an SV in classes 4 and 5 is calculated as: severity score = (class − 3) * number of SV copies, where the number of copies is set at 1 for heterozygous SVs, and 2 for homozygous SVs. A custom script was used to convert results from an SV-based report into a gene-based report. For each gene, the annotated severity scores of all SVs overlapping a given gene were aggregated for affected individuals and unaffected individuals separately. The genes were ranked based on the aggregated SV severity scores in affected individuals. Genes were selected for downstream analysis using the following criteria: 1. containing SV(s) in more than one affected individual; 2. expressed in at least one of the brain expression databases; 3. SV(s) overlapping the exonic region; and 4. the aggregated score in affected individuals is higher than that in unaffected individuals.

Gene annotation, pathway, and enrichment analyses

For gene function predictions, the pLI (probability of being loss-of-function intolerant) score was extracted from gnomAD (v3.0) for each gene (Karczewski et al. 2020). Brain developmental gene expression data were obtained from the GTEx project (v8) (GTEx Consortium 2020), the BrainSpan Atlas of the Developing Human Brain project (RNA-Seq GENCODE v10 summarized to genes, http://www.brainspan.org/static/download.html ) (Miller et al. 2014), and the Human Developmental Biology Resource (https://www.ebi.ac.uk/gxa/experiments/E-MTAB-4840/Downloads ) (Lindsay et al. 2016), as previously described (Cao et al. 2021). The genes’ association with specific cells or diseases are predicted by the HumanBase (https://hb.flatironinstitute.org) algorithm that learns from large genomics data collections (Greene et al. 2015). The gene knock-out mouse behavior was downloaded from the International Mouse Phenotyping Consortium (IMPC) (ftp://ftp.ebi.ac.uk/pub/databases/impc/latest/ ) (Dickinson et al. 2016). Candidate genes for several NDDs were collected from previous studies and the gene sets and genes are summarized in Table S2.

The protein–protein interaction (PPI) networks were generated based on gene interactions from three databases, including STRING (v11.5) (Szklarczyk et al. 2019), GIANT (v2.0) (Wong et al. 2018) and ConsensusPathDB (v35) (Kamburov et al. 2011). The cut-off for the “combined score” for STRING was set to 700. For GIANT, interactions with score > 0.6 in tissues from brain and neuron system were used, including brain, astrocyte_of_the_cerebellum, astrocyte_of_the_cerebral_cortex, basal_ganglion, brainstem, central_nervous_system, central_nervous_system_pericyte, cerebellar_cortex, cerebellum, cerebral_cortex, cerebral_hemisphere, corpus_striatum, forebrain, frontal_cortex, glial_cell, global, globus_pallidus, gray_matter, gray_matter_of_telencephalon, gyrus, hindbrain, hippocampal_formation, hypothalamus, lobe_of_cerebral_hemisphere, neocortex, nervous_system, neural_cell, neuron, nucleus_ of_brain, pallium, parietal_lobe, peripheral_nervous_system, pituitary_gland, pons, spinal_cord, substantia_nigra, telencephalic_nucleus, and telencephalon. The Python NetworkX package (https://networkx.org) was used to visualize the PPI network and the positions of genes in the network were saved for final plotting with R (v4.1.0) packages ggplot2 and scatterpie. Details of interaction data download and processing procedures were described previously (Cao et al. 2021; Sun et al. 2022).

Gene set enrichment analyses were performed using over-representation analysis provided by ConsensusPathDB (Kamburov et al. 2011). All pathway, Gene Ontology (GO), and protein complex-based enrichment analyses were enabled, with the minimum two genes from the input and p-value cut-off set to 1. Enrichment p values for all genes were determined by ConsensusPathDB using a hypergeometric test and enrichment p-value for each gene list was calculated with Fisher’s exact test as described previously (Cao et al. 2021). An enriched term (i.e., GO, pathway, or protein complex) was selected for further analysis if: 1) the total number of genes belong to the term is ≤ 500; 2) the term includes more than three genes from all input genes (ADHD, ASD/ADHD, and NDD genes); and 3) enrichment p-value < 0.01 for ADHD and/or ASD/ADHD gene lists.

Results

Overall study design

For samples in the NJLAGS cohort, we conducted the SNP genotyping microarray analysis and the WGS analysis in parallel (Fig. 1). Genotypes from the genotyping microarray were used to perform linkage analysis and the WGS data were then used to prioritize genes based on SNVs/indels and SVs in both linkage regions and the whole genome. All families that undergone WGS were a part of the linkage study, except for family FAM52, where DNA sample is only available for the proband and therefore was not included in the linkage study (Table S1). To account for the genetic heterogeneity of ASD and ADHD, candidate gene sets were subjected to enrichment and pathway analyses.

Fig.1.

Fig.1

Open in a new tab

Project overview and overall analysis flow

Among the 73 families involved in WGS, 47 (64.3%) have individuals who are also affected by ADHD. For the 493 total individuals within the 73 families, 125 are affected by either ASD or ADHD (Table 1). Among the 98 individuals affected by ASD, 41 (41.8%) are also affected by ADHD (Table 1). Males are more likely to be affected by both ASD and ADHD compared to females (male/female ratio: 2.73, T test p = 0.03).

Table 1.

Summary of WGS families and their phenotypes

All Male Female Male/female ratio Families
ASD 98 77 21 3.67 73
ADHD 68 46 22 2.09 47
ASD and ADHD 41 30 11 2.73 36
ASD or ADHD 125 93 32 2.91 73
All samples 493 281 212 1.33 73
Open in a new tab

Linkage analysis highlights linkage peaks for ADHD

A total of 524 individuals from 111 families were genotyped for the linkage analysis. The fully constructed pedigrees contain 707 individuals, after adding missing individuals (e.g, grandparents) to link sibships of cousins. Families have an average of 5.7 individuals (including ungenotyped individuals) and 4.2 genotyped individuals.

We performed multipoint linkage analysis of the data for three phenotypes: “ADHD”, where individuals diagnosed with ADHD are considered affected; “ADHD or ASD” (referred to as “ASD/ADHD” in the following text), where both ADHD and ASD affected individuals are considered affected; and “ADHD and ASD”, where only individuals diagnosed with both ADHD and ASD are considered affected. The results are summarized in Fig. 2 with peaks summarized in Table S3. The ADHD phenotype was linked to 17p13.1–2, meeting the conventional standard for declaring linkage (PPL = 0.38, p < 0.001) as determined by a simulation study to estimate the empirical null distribution (Fig. 2). Because the pooled PPL was larger than the sequentially updated PPL, we infer that the locus is largely homogenous and these results do not offer evidence that data from either of the two waves are inconsistent with linkage. As such, the ADHD locus did not depend on the strictness of the ASD criteria used in family ascertainment. The locus on 12q12–15 met the criteria for suggestive linkage to ADHD (PPL = 0.27, p < 0.01, Fig. 2). Similar to the locus on 17p, the pooled PPL for 12q was larger than the sequentially updated PPL, indicating that the linkage with ADHD did not differ across the two waves with different ASD recruitment criteria. Several additional loci met the criteria of nominal linkage. The ADHD phenotype is nominally linked to 3p13 (pooled PPL = 0.11, sequential PPL = 0.23) with evidence for heterogeneity across ASD criteria and 19q12–13.1 (pooled PPL = 0.20, sequential PPL = 0.14) with no evidence for heterogeneity. The phenotype ADHD/ASD had two nominal linkage peaks, both suggesting heterogeneity, on 19p13.3 (pooled PPL = 0.11, sequential PPL = 0.18) and 20q13.13–13.33 (pooled PPL = 0.02, sequential PPL = 0.19). The “ADHD and ASD” phenotype did not show evidence for linkage, likely due to the small sample size of the phenotype (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Linkage analysis for ASD and ADHD. Sequential and pooled analysis results are shown in blue and red lines, respectively. The red dotted line indicates the PPL = 0.26, which corresponds to p = 0.01

Whole‑genome sequencing and variant discovery

We selected 272 individuals from the recruited families to undergo WGS to ~ 30 × coverage. The sequencing reads that passed quality filters were mapped to the human genome reference hg19 and genomic variants are called jointly using GATK for small variants (SNVs and indels) and MetaSV for SVs.

The number of small variants and SVs was consistent across samples. For small variants, around 4 million SNVs were identified for most of the samples, and the number of indels ranged from 700,000 to 1.2 million (Table S1). For SVs, an average of 743,280 SVs (including unknown type) were discovered for each individual, ranging from 325,443 to 1,291,816 across samples. We selected SVs that were identified by at least two methods for the downstream analysis. An average 1,802 SVs passed the filter in each individual (see Methods for detail), ranging from 1,332 to 2,620 (Table S1). Most SVs are within the 50–1,000 bps range and the largest category is deletion (Figure S1).

Candidate genes associated with ADHD linkage regions

To identify the risk genes for ADHD, we first examined SNVs and indels within the chromosome 12 and 17 significant ADHD linkage regions using the pVAAST workflow. pVAAST prioritizes genes by integrating segregation information, population allele frequency, and functional impact of variants in a gene (Hu et al. 2014). After Bonferroni correction of the p values (1,299 genes for chromosome 12 and 563 for chromosome 17 linkage regions) and selecting genes express in brain, we identified 34 genes that met the significance criteria (Bonferroni-corrected p < 0.05) for the dominant mode and 3 genes for the recessive mode (Table 2). Proline glutamate and leucine rich protein 1 (PELP1) is present in both lists. Among the 36 unique genes, 18 are located in the chromosome 12 linkage region and 18 in the chromosome 17 linkage region, respectively. Four genes (KDM6B, COL2A1, PTPRB, PER1) were reported in one or more NDD disease gene databases (Table 2). In addition to SNVs and indels, we also identified SVs using the WGS data. Within the linkage regions, an average of 83 SVs were identified per individual, with a minimum of 32 and a maximum of 157 across the cohort. None of the genes passed our SV gene filtering criteria (see whole-genome SV analysis below for filtering detail).

Table 2.

Candidate genes in linkage regions

Gene Chr Inheritance P value pLI Database
KDM6B chr17 Dominant 5.63E-04 1 SFARI, iPSYCH
MYH13 chr17 Dominant 5.63E-04 0
USP6 chr17 Dominant 5.63E-04 0
MYBBP1A chr17 Dominant 5.63E-04 0
KMT2D chr12 Dominant 1.30E-03 1
MYO1A chr12 Dominant 1.30E-03 0
COL2A1 chr12 Dominant 1.30E-03 1 iPSYCH
PTPRB chr12 Dominant 1.30E-03 0.94 SFARI
ITGA7 chr12 Dominant 1.30E-03 0
STAT2 chr12 Dominant 1.30E-03 0.04
NEURL4 chr17 Dominant 5.63E-04 1
TIMELESS chr12 Dominant 1.30E-03 0
CCDC65 chr12 Dominant 1.30E-03 0
PFAS chr17 Dominant 5.63E-04 0
NLRP1 chr17 Dominant 5.63E-04 0
EIF4B chr12 Dominant 1.30E-03 1
NELL2 chr12 Dominant 1.30E-03 1
WNT10B chr12 Dominant 1.30E-03 0
ALG10B chr12 Dominant 2.60E-03 0
PELP1 chr17 Dominant 1.69E-03 1
CHRNE chr17 Dominant 1.69E-03 0
GRASP chr12 Dominant 3.90E-03 0.93
ARHGEF15 chr17 Dominant 2.82E-03 0
TROAP chr12 Dominant 6.50E-03 0
TRPV3 chr17 Dominant 4.50E-03 0
P2RX5 chr17 Dominant 4.50E-03 NA
DNAJC22 chr12 Dominant 1.04E-02 0
ACADVL chr17 Dominant 1.13E-02 0
ZZEF1 chr17 Dominant 1.13E-02 1
TEKT1 chr17 Dominant 1.13E-02 0
ANKRD33 chr12 Dominant 2.60E-02 1
PER1 chr17 Dominant 1.13E-02 0.88 SFARI, ADH-Dgene
LYZ chr12 Dominant 2.60E-02 0
USP43 chr17 Dominant 1.69E-02 0.01
PELP1 chr17 Recessive 5.63E-04 1
SHPK chr17 Recessive 1.13E-03 0
ESPL1 chr12 Recessive 3.90E-03 1
Open in a new tab

The highest-ranking gene is Lysine Demethylase 6B (KDM6B) on chromosome 17. KDM6B has two missense variants segregating in two families. One mutation (17–7751888-C-G, p.Thr761Ser) is predicted to be possibly damaging by PolyPhen2. The other (17–7749972-G-T, p.Val209Leu) is predicted to be benign by PolyPhen2. Three additional missense variants are also present in patients but do not follow strict segregation patterns (Table S4). KDM6B is expressed in the brain and it is intolerant to mutations (pLI = 1). The KDM6B protein demethylates trimethylated lysine-27 on histone H3. Pathogenic alterations in histone lysine methylation and demethylation genes have been associated with multiple NDDs (Stolerman et al. 2019), including ADHD (Insa Pineda and Gómez González 2021; Kafle et al. 2021). KDM6B is also annotated in NDD databases such as SFARI and iPSYCH, and it is identified as one of the 102 risk genes in a recent large ASD study (Satterstrom et al. 2020).

The highest-ranking gene in the chromosome 12 linkage region is lysine methyltransferase 2D (KMT2D), which is also involved in lysine methylation, similar to KDM6B. Two missense variants segregate in two families (12-49428694-T-C, p.Asp3419Gly; 12-49436073-C-T, p.Asp1970Asn), both are predicted to be damaging by both SIFT and PolyPhen2. Nine additional missense variants are present in patients but do not follow strict segregation patterns (Table S4). KMT2D is expressed in the brain and intolerant to mutations (pLI = 1).

Another candidate gene from the chromosome 17 linkage region, Neuralized E3 Ubiquitin Protein Ligase 4 (NEURL4), has a missense variant (17-7224433-C-T, p.Gly1120Arg) segregating in one family and four additional missense variants in other affected individuals. NEURL4 is expressed in the brain and intolerant to mutations (pLI ~ 1). NEURL4 protein is a scaffold protein, which maintains normal centriolar homeostasis and prevents formation of ectopic microtubule organizing centers (Li et al. 2012). Although no publication directly links NEURL4 with NDDs, evidence has shown that microtubule related genes are related to neuronal migration and dendritic functioning (Chang et al. 2018). Knockout of NEURL4 in mice causes neurological phenotypes such as decreased prepulse inhibition (Dickinson et al. 2016).

In the chromosome 12 linkage region, we also identified the circadian rhythm controller gene Timeless Circadian Regulator (TIMELESS) with a missense mutation (12–56827209-C-A, p.Ala129Ser). The mutation is segregated into two families and is predicted to be deleterious by SIFT, suggesting its potential role in ADHD etiology. Sleep disorders are common in both ASD and ADHD patients (Ming and Walters 2009). A previous study found mutations in circadian-relevant genes, including TIMELESS, are more frequent in ASD affected individuals than controls (Yang et al. 2016).

Novel and known risk genes associated with ADHD and ASD/ADHD in the whole‑genome analysis

Because ASD and ADHD are highly heterogeneous and multiple variants/genes could contribute to their genetic etiology, next we performed pVAAST analysis on SNVs/indels for the whole genome. After selecting genes with evidence of inheritance (LOD score > = 1) and express in the brain (see Methods for detail), we identified 16, 25, 70, and 29 candidate genes for ADHD-dominant model, ADHD-recessive model, ASD/ADHD-dominant model, and ASD/ ADHD-recessive model, respectively (Table S5).

From SVs, we identified 209 and 227 genes that contain exonic SVs in at least one affected individual for ADHD and ASD/ADHD, respectively (Table S6). To further prioritize genes affected by SVs, we calculated the combined severity scores of SVs within each gene and determined their brain expression levels. We then selected genes whose combined severity scores were larger in affected individuals than in unaffected ones and were expressed in the brain. A total of 43 and 44 genes passed our SV gene filtering criteria for ADHD and ASD/ADHD, respectively (Table S6).

Next, we created two high-confidence candidate gene sets for ADHD and ASD/ADHD. For ADHD, we combined significant linkage region candidate genes, high-confidence candidate genes from the whole-genome pVAAST analyses, and the SV analysis. For ASD/ADHD, we combined candidate genes from whole-genome pVAAST and SV analyses. A total of 207 unique genes were included in these two confidence gene sets, including 113 genes for ADHD and 138 for ASD/ADHD (Table 3). We annotated these genes using public resources, including gene association databases, literature search databases, tolerance to loss-of-function mutation scores, mouse knock-out experiments, and expression databases (Table S7, see Methods for detail). A total of 31 genes (15%) were described by previous studies or existing NDD gene databases (Table S2), including 22 genes from the SFARI database. Three genes (ANKRD11, KDM6B, COL2A1) overlap risk genes in the iPSYCH study of ASD and ADHD (Satterstrom et al. 2019).

Table 3.

Candidate gene summary

SNV/Indel Genes Families* SV Genes Families* Total (unique) Genes Families*
ADHD
 Linkage 36 19 0 0 36 19
 WGS 40 47 43 42 83 47
 Total (Unique) 70 47 43 42 113 47
ASD/ADHD
 WGS 94 73 44 73 138 73
 Total (Unique) 140 73 61 73 207 73
Open in a new tab
*

For SNV/indel, families with segregating variants in a given gene were counted. For SV, families with at least one affected individual with an exonic SV were counted

In addition to genes overlapping known NDD genes, many other candidate genes showed functional relevance to neurodevelopmental processes or have been reported to be associated with NDDs. For example, we identified Protein disulfide-isomerase A2 (PDIA2) in the whole-genome pVAAST analysis. With the ADHD-dominant model, PDIA2 had a missense mutation (16–334899-G-A, p.Ala188Thr) segregating in two families and was predicted as damaging by SIFT. It had additional 27 variants segregating in 13 families. PDIA2 encodes a member of disulfide-isomerase family of endoplasmic reticulum (ER) proteins that catalyze protein folding and thiol–disulfide interchange reactions and is highly expressed in the cerebellar cortex. HumanBase predicted a strong association of PDIA2 to ASD (rank 1) and a study investigating the role of ER stress in autism identified PDIA2 upregulation in autistic subjects (Dong et al. 2018).

In whole-genome SV analysis, we discovered a novel candidate gene Calcium Voltage-Gated Channel Auxiliary Subunit Alpha2delta 4 (CACNA2D4). CACNA2D4 is one of the genes that encode the alpha-2/delta subunit of the voltage-dependent calcium channel, among three other genes (CACNA2D1, CACNA2D2, CACNA2D3) (Vergult et al. 2015). While a dozen genes encoding the alpha1 and beta subunits of the calcium channel are previously identified in the SFARI database (Table S2), CACNA2D4 was absent. In our study, CACNA2D4 was found to contain SVs in affected individuals in five families in the ASD/ADHD analysis (one SV overlaps intron 16 to intron 26 and other SVs within intron 26). The SV overlapping intron 16 to intron 26 is a 36 kb deletion (chr12:1948911–1985036) that is rare in general populations (< 0.16% in European populations in gnomAD, https://gnomad.broadinstitute.org/variant/DEL_12_124891?dataset=gnomad_sv_r2_1 ). Two previous ASD studies identified CAC-NA2D4 deletions in ASD patients and support our finding (Prasad et al. 2012; Smith et al. 2012).

Candidate genes participate in pathways related to neurological disorders

To explore the relationships among the candidate genes and to identify the potential shared mechanisms between ASD, ADHD, and other NDDs, we performed enrichment analysis for the 207 candidate genes with 1,629 known NDD genes (Table S2) and determined the significance of enrichment for each gene list (Table S8). ADHD and ASD/ADHD showed strong enrichment in GO terms and pathways related to cilium and microtubule, suggesting that many genes in this group are involved in brain development (Table S8). For example, pathway “Ciliopathy” (WP4803) is significantly enriched in both ADHD and ASD/ADHD gene sets, overlapping five and seven genes in the two sets, respectively. “cilium or flagellum-dependent cell motility” (GO:0001539) is one of the top enriched GO terms in ADHD gene list, and overlap four genes in the candidate genes (BBS4, CCDC65, TEKT1, TEKT5).

Other processes that are known to relate to NDDs are also present. For example, “cation channel activity” (GO:0005261) is enriched in both ADHD and ASD/ADHD gene sets, overlapping seven and eight genes in the two sets, respectively. Besides shared enriched terms and pathways, each gene set also contains phenotype-specific enrichments. For example, striatum development (GO:0021756) and cerebral cortex development (GO:0021987) are significantly enriched in the ASD/ADHD gene set, overlapping two and four candidate genes, respectively. On the other hand, “sensory perception of light stimulus” is only significantly enriched in ADHD gene set (GO:0050953), overlapping seven genes in the set.

Protein–protein interaction network of candidate genes

To further examine the relationship among the candidate genes, we constructed PPI networks using our candidate genes and known NDD genes (Fig. 3, Table S9). Of the 113 ADHD genes, 23 were connected with known NDD genes into a single PPI network with interactions supported by at least 2 databases (Fig. 3A, 45 genes in total), and 50 of the remaining genes were connected with known NDD genes with interactions supported by at least one database (Fig. 3B, 150 genes). Similarly, for the 138 ASD/ADHD genes, 44 were connected after adding NDD genes with interactions supported by ≥ 2 databases (Fig. 3C, 71 genes in total), and 31 of the remaining genes were connected after adding NDD genes with interactions in ≥ 1 database (Fig. 3D, 83 genes). Several enriched terms, such as histone methyltransferase complex and cation channel activity, showed extensive connection with known genes for ASD, ADHD, and other NDDs, supporting the shared genetic etiology of ADHD and ASD.

Fig. 3.

Fig. 3

Open in a new tab

Protein–protein interaction (PPI) networks of candidate genes. Candidate genes were ADHD risk genes in (A, B) and ASD/ ADHD risk genes in (C, D). A, C: PPI networks of the candidate genes. Other known NDD genes were added as intermediate nodes if they interact with more than one candidate gene supported by PPI evidence from ≥ 2 databases. B, D: PPI network of candidate genes that are not in (A) and (C), respectively. Known genes were added as intermediate nodes if they interact with more than one candidate gene with PPI evidence from any PPI database. PPI networks were defined by three databases, ConsensusPathDB, STRING, and GIANT_v2. Genes were colored by gene lists. ASD/ADHD genes were used to color genes in (A, B) and ADHD genes were used to color genes in (C, D). To simplify the network, PPIs between non-candidate genes were removed. Genes belonging to two GO terms, methylation and cation channel activity, were highlighted in A, C, respectively

Discussion

ASD and ADHD co-occur at a high frequency and studies suggested shared genetic risk factors between the two disorders (Rommelse et al. 2010; Doherty and Owen 2014; Ghirardi et al. 2018; Johnson et al. 2015; Polderman et al. 2014; Ronald and Hoekstra 2011; Stergiakouli et al. 2017). Given the worsened negative impact on patients’ quality of life when the two conditions co-occur, it is crucial to understand the underlying genetic risk factors for each disorder and the co-occurrence. In the NJLAGS cohort, ADHD occurs in 41.8% of ASD patients, consistent with previous studies (reviewed in (Rong et al. 2021)). The family design of the NJLAGS study provides a unique opportunity to identify inherent and de novo genetic factors underlying both ASD and ADHD.

In this study, we identified two linkage peaks (12q12–15, 17p13.1–2) for the ADHD phenotype from 111 families. Using WGS data from 272 selected samples, we discovered 36 significant genes within the linkage regions and explored their gene functions. Expanding the analysis to the whole genome revealed more risk genes for both ADHD and ASD/ADHD phenotypes. Notably, both top candidate genes in the two linkage regions (KDM6B, KMT2D) are involved in the histone lysine methylation process. Histone modification is a central process regulating gene expression during neurodevelopment and mutations in KDM6B have been linked to multiple NDDs (Stolerman et al. 2019), including ADHD (Insa Pineda and Gómez González 2021; Kafle et al. 2021). At the whole genome level, 31 of the high-confidence candidate genes were identified in previous studies or NDD gene databases. In addition, we observed multiple NDD-related pathways overrepresented in the high-confidence ADHD and ASD/ ADHD gene sets (Table S8). Taken together, the results demonstrate the power of combining linkage analysis, rare variant analysis, and SV analyses to uncover risk genes.

By studying ADHD-affected patients in families with ASD probands, we identified genetic risk factors underlying ASD and ADHD co-occurrence. Three candidate genes (ANKRD11, KDM6B, COL2A1) were reported in a previous study of ASD and ADHD comorbidity in Danish cohorts (Satterstrom et al. 2019). In the Danish cohorts, all three genes contain multiple protein-truncating variants or missense variants in ASD or ADHD patients, and ANKRD11 (Ankyrin Repeat Domain Containing 11) is one of the top candidate genes with a significant excess of constrained rare protein-truncating variants (crPTVs) (Satterstrom et al. 2019). In addition, several genes (KDM6B, PTPRB, PER1) that are significant within the ADHD linkage region are described in previous ASD studies (Abrahams et al. 2013; Schork et al. 2019; Zhang et al. 2012). We did not identify overlapping candidate genes with several recent across-disorder meta-analysis of ASD and ADHD (Baranova et al. 2022; Rao et al. 2022) and previous ADHD linkage studies (Rommelse et al. 2010; Corominas et al. 2020; Zhou et al. 2008; Taurines et al. 2012). The small number of overlapping candidate genes and the lack of overlapping linkage peaks among studies support the highly heterogenous genetic etiology of ASD and ADHD. By providing an extensive candidate gene set, we are not only reaffirming the linkage between ASD and ADHD, but also suggesting novel potential risk pathways for further investigation.

Several important GO terms and pathways recurred in both gene sets in our enrichment analysis, namely cilium/ microtubule, ion channel, histone methylation, and chromatin remodeling. Pathways we discovered that are important in ADHD etiology are also vital in ASD etiology. These pathways revealed interesting aspects of mechanisms in ASD and ADHD etiology as an interconnected network of hundreds of genes and pathways (Iakoucheva et al. 2019) and underlined the complexity of genetic casualties of NDDs. A recent article argued that several NDDs share common genetic risk factors and there might not be “Autism-specific” genes (Myers et al. 2020). Our results support this view. We also identified some important but underappreciated processes such as the circadian rhythm (GO:0007623), which has phenotypic effects in cognition, mood, and reward-related behaviors. Research documented that the prevalence of insomnia ranges from 50 to 80% in ASD patients, compared to 9–50% in age-matched typically developing children (Charrier et al. 2017). However, the underlying mechanism of association and genes involved remain to be determined. The genes related to circadian rhythm we identified in this study (MYBBP1A, PER1, TIMELESS) could shed light on the genetic etiology of sleep disorders in ASD and ADHD patients.

Our study has a few limitations. The limited sequencing capacity led to the sequencing of only a selected number of ADHD-affected individuals, as the NJLAGS dataset was collected primarily to identify ASD and language-impaired probands. However, recently, we sent out follow-up questionnaires to all NJLAGS families requesting an update on ADHD diagnoses. We specifically asked who had received a diagnosis from a medical professional (i.e., psychiatrist, developmental pediatrician, neurologist, and psychologist). Of the 34 families who responded to our questionnaire, 15 were families who had some sequencing data and whose data are included in this manuscript. From these 15 families, 28 individuals had received a formal diagnosis of ADD/ADHD by a medical professional. Importantly, only 13 of these family members had been initially identified as having a diagnosis of ADD/ADHD. In some cases these were younger family members who were too young at the time of data collection to be identified as affected while in other cases, we did not have this diagnostic information available. These updated findings will be included in future ADHD in genetic analyses. Moving forward, we anticipate extensive sequencing efforts to continue for the NJLAGS project resulting in the sequencing of all NJLAGS families. As more ADHD-affected individuals are identified and sequenced, this more complete pedigree information and stronger statistical power will allow us to further discover potential common risk factors underlying the ASD and ADHD etiology.

Another limitation of this study is we focused mainly on variants affecting protein coding sequences. Non-coding regulatory variants are known to play an important role in NDDs, although predicting their functional impact is still difficult (Novikova et al. 2021). With the rapid development of functional prediction algorithms for non-coding variants, we will identify additional candidate genes affected by regulatory variants using the cohort’s WGS data in the near future.

In conclusion, using linkage analysis and WGS-based variant analysis, we examined the genetic etiology of ASD and ADHD in the NJLAGS cohort. The family-based design of the cohort allows us to leverage the power of variant segregation while taking into account the phenotype heterogeneity of ASD and ADHD. By examining both small variants and large SVs, we identified genes and pathways, previously known or novel, that could contribute to ASD or ADHD etiology. Because of the difficulty of distinguishing some of the ASD and ADHD behavioral characteristics, understanding the genetic risk factors for these disorders can provide guidance to clinical intervention approaches, both behaviorally and pharmacologically. Future studies, especially functional assessment of these candidate genes, will elucidate their roles in ASD and ADHD and improve the diagnosis and treatment of these disorders.

Supplementary Material

Supplemental Tables
Supplemental Figure S1

Acknowledgements

We thank the patients who participated in and contributed to this study. We gratefully acknowledge access to the HPC facilities and support of the computational STEM and bioinformatics scientists from the Office of Advanced Research Computing at Rutgers University.

Funding

This study was supported by NIMH grants R01 MH-070366 and RC1 MH-088288 to LMB; New Jersey Governor’s Council for Medical Research and Treatment of Autism grant CAUT12APS006 to LMB and CAUT19APL028 to JX and LMB.

Footnotes

Declarations

Conflict of interest The authors declare that they have no competing interests.

Ethical approval The study is approved by the Institutional Review Board at Rutgers, The State University of New Jersey (IRB number: 13-112Mc).

Consent to participate Informed written consent or assent was obtained from all subjects involved in the study.

Consent for publication Not applicable.

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s00439-022-02496-z.

Data availability

The raw sequencing reads, variants, and genotypes for all samples are available in the National Institute of Mental Health Data Archive (NDA) under experiments C1932 and C2933.

References

  1. Abecasis GR, Cherny SS, Cookson WO, Cardon LR (2002) Merlin– rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30(1):97–101 [DOI] [PubMed] [Google Scholar]
  2. Abrahams BS, Arking DE, Campbell DB, Mefford HC, Morrow EM, Weiss LA et al. (2013) SFARI gene 2.0: a community-driven knowledgebase for the autism spectrum disorders (ASDs). Mol Autism 4:1–3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Abyzov A, Gerstein M (2011) AGE: defining breakpoints of genomic structural variants at single-nucleotide resolution, through optimal alignments with gap excision. Bioinformatics 27(5):595–603 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Abyzov A, Urban AE, Snyder M, Gerstein M (2011) CNVnator: an approach to discover, genotype, and characterize typical and atypical CNVs from family and population genome sequencing. Genome Res 21(6):974–984 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Abyzov A, Li S, Kim DR, Mohiyuddin M, Stütz AM, Parrish NF et al. (2015) Analysis of deletion breakpoints from 1092 humans reveals details of mutation mechanisms. Nat Commun 6(1):1–12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. American Psychiatric Association (2000) Diagnostic and statistical manual of mental disorders, 4th edition text revision (DSM-IV-TR). American Psychiatric Association, Washington DC [Google Scholar]
  7. American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, fifth edition (DSM-5). American Psychiatric Association, Washington, DC [Google Scholar]
  8. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS et al. (2012) SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol 19(5):455–477 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Baranova A, Wang J, Cao H, Chen JH, Chen J, Chen M et al. (2022) Shared genetics between autism spectrum disorder and attention-deficit/hyperactivity disorder and their association with extraversion. Psychiatry Res 314:114679. [DOI] [PubMed] [Google Scholar]
  10. Bartlett CW, Flax JF, Logue MW, Vieland VJ, Bassett AS, Tallal P et al. (2002) A major susceptibility locus for specific language impairment is located on 13q21. Am J Hum Genet 71(1):45–55 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bartlett CW, Hou L, Flax JF, Hare A, Cheong SY, Fermano Z et al. (2014) A genome scan for loci shared by autism spectrum disorder and language impairment. Am J Psychiatry 171(1):72–81 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cao X, Zhang Y, Abdulkadir M, Deng L, Fernandez TV, Garcia-Delgar B et al. (2021) Whole-exome sequencing identifies genes associated with Tourette’s disorder in multiplex families. Mol Psychiatry 26(11):6937–6951 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chang Q, Yang H, Wang M, Wei H, Hu F (2018) Role of microtubule-associated protein in autism spectrum disorder. Neurosci Bull 34(6):1119–1126 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Charrier A, Olliac B, Roubertoux P, Tordjman S (2017) Clock genes and altered sleep-wake rhythms: their role in the development of psychiatric disorders. Int J Mol Sci 18(5):938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chen K, Wallis J, McLellan M, Larson D, Kalicki J, Pohl C et al. (2009) BreakDancer: an algorithm for high-resolution mapping of genomic structural variation. Nat Methods 6(9):677–681 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Corominas J, Klein M, Zayats T, Rivero O, Ziegler GC, Pauper M et al. (2020) Identification of ADHD risk genes in extended pedigrees by combining linkage analysis and whole-exome sequencing. Mol Psychiatry 25(9):2047–2057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA et al. (2011) The variant call format and VCFtools. Bioinformatics 27(15):2156–2158 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C et al. (2011) A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet 43(5):491–498 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Dickinson ME, Flenniken AM, Ji X, Teboul L, Wong MD, White JK et al. (2016) High-throughput discovery of novel developmental phenotypes. Nature 537(7621):508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Doherty JL, Owen MJ (2014) Genomic insights into the overlap between psychiatric disorders: implications for research and clinical practice. Genome Med 6(4):1–13 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dong D, Zielke HR, Yeh D, Yang P (2018) Cellular stress and apoptosis contribute to the pathogenesis of autism spectrum disorder. Autism Res 11(7):1076–1090 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Elia J, Gai X, Xie HM, Perin JC, Geiger E, Glessner JT et al. (2010) Rare structural variants found in attention-deficit hyperactivity disorder are preferentially associated with neurodevelopmental genes. Mol Psychiatry 15(11):1122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Geoffroy V, Herenger Y, Kress A, Stoetzel C, Piton A, Dollfus H et al. (2018) AnnotSV: an integrated tool for structural variations annotation. Bioinformatics 34(20):3572–3574 [DOI] [PubMed] [Google Scholar]
  24. Ghirardi L, Brikell I, Kuja-Halkola R, Freitag CM, Franke B, Asherson P et al. (2018) The familial co-aggregation of ASD and ADHD: a register-based cohort study. Mol Psychiatry 23(2):257–262 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Govil M, Vieland V (2008) Practical considerations for dividing data into subsets prior to PPL analysis. Human Hered 66:223–237 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Greene CS, Krishnan A, Wong AK, Ricciotti E, Zelaya RA, Himmelstein DS et al. (2015) Understanding multicellular function and disease with human tissue-specific networks. Nat Genet 47(6):569–576 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Consortium GTEx (2020) The GTEx consortium atlas of genetic regulatory effects across human tissues. Science 369(6509): 1318–1330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Ho SS, Urban AE, Mills RE (2019) Structural variation in the sequencing era. Nat Rev Genet 21(3):171–189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hu H, Huff CD, Moore B, Flygare S, Reese MG, Yandell M (2013) VAAST 2.0: improved variant classification and disease-gene identification using a conservation-controlled amino acid substitution matrix. Genet Epidemiol 37(6):622–634 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hu H, Roach JC, Coon H, Guthery SL, Voelkerding KV, Margraf RL et al. (2014) A unified test of linkage analysis and rare-variant association for analysis of pedigree sequence data. Nat Biotechnol 32(7):663–669 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Iakoucheva LM, Muotri AR, Sebat J (2019) Getting to the cores of autism. Cell 178(6):1287–1298 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Insa Pineda I, Gómez González CL (2021) The KDM6B mutation: phenotype and clinical characteristics-report of a case. Rev Psiquiatr Salud Ment. 10.1016/j.rpsm.2020.12.005 [DOI] [PubMed] [Google Scholar]
  33. Johnson MH, Gliga T, Jones E, Charman T (2015) Annual research review: infant development, autism, and ADHD–early pathways to emerging disorders. J Child Psychol Psychiatry 56(3):228–247 [DOI] [PubMed] [Google Scholar]
  34. Kafle OP, Wang X, Cheng S, Ding M, Li P, Cheng B et al. (2021) Genetic correlation analysis and transcriptome-wide association study suggest the overlapped genetic mechanism between gout and attention-deficit hyperactivity disorder. Can J Psychiatry 66(12):1077–1084 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kamburov A, Pentchev K, Galicka H, Wierling C, Lehrach H, Herwig R (2011) ConsensusPathDB: toward a more complete picture of cell biology. Nucleic Acids Res 39:D712–D717 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Karczewski KJ, Francioli LC, Tiao G, Cummings BB, Alfoldi J, Wang Q et al. (2020) The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581(7809):434–443 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25(14):1754–1760 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N et al. (2009) The sequence alignment/map format and SAMtools. Bioinformatics 25(16):2078–2079 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Li J, Kim S, Kobayashi T, Liang FX, Korzeniewski N, Duensing S et al. (2012) Neurl4, a novel daughter centriole protein, prevents formation of ectopic microtubule organizing centres. EMBO Rep 13(6):547–553 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lindsay SJ, Xu YB, Lisgo SN, Harkin LF, Copp AJ, Gerrelli D et al. l (2016) HDBR expression: a unique resource for global and individual gene expression studies during early human brain development. Front Neuroanat. 10.3389/fnana.2016.00086 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Liu S, Rao S, Xu Y, Li J, Huang H, Zhang X et al. (2020) Identifying common genome-wide risk genes for major psychiatric traits. Hum Genet 139(2):185–198 [DOI] [PubMed] [Google Scholar]
  42. Logue MW, Vieland VJ, Goedken RJ, Crowe RR (2003) Bayesian analysis of a previously published genome screen for panic disorder reveals new and compelling evidence for linkage to chromosome7. Am J Med Genet B Neuropsychiatr Genet 121b(1):95–99 [DOI] [PubMed] [Google Scholar]
  43. Martin J, Cooper M, Hamshere ML, Pocklington A, Scherer SW, Kent L et al. (2014) Biological overlap of attention-deficit/hyper-activity disorder and autism spectrum disorder: evidence from copy number variants. J Am Acad Child Adolesc Psychiatry 53(7):761–70e26 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Miller JA, Ding SL, Sunkin SM, Smith KA, Ng L, Szafer A et al. (2014) Transcriptional landscape of the prenatal human brain. Nature 508(7495):199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Ming X, Walters AS (2009) Autism spectrum disorders, attention deficit/hyperactivity disorder, and sleep disorders. Curr Opin Pulm Med 15(6):578. [DOI] [PubMed] [Google Scholar]
  46. Mohiyuddin M, Mu JC, Li J, Bani Asadi N, Gerstein MB, Abyzov A et al. (2015) MetaSV: an accurate and integrative structural-variant caller for next generation sequencing. Bioinformatics 31(16):2741–2744 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Myers SM, Challman TD, Bernier R, Bourgeron T, Chung WK, Constantino JN et al. (2020) Insufficient evidence for “autism-specific” genes. Am J Hum Genet 106(5):587–595 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Novikova G, Andrews SJ, Renton AE, Marcora E (2021) Beyond association: successes and challenges in linking non-coding genetic variation to functional consequences that modulate Alzheimer’s disease risk. Mol Neurodegener 16(1):27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Patterson N, Price AL, Reich D (2006) Population structure and eigenanalysis. PLoS Genet 2(12):e190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Polderman TJ, Hoekstra RA, Posthuma D, Larsson H (2014) The co-occurrence of autistic and ADHD dimensions in adults: an etiological study in 17,770 twins. Transl Psychiatry 4:e435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Prasad A, Merico D, Thiruvahindrapuram B, Wei J, Lionel AC, Sato D et al. (2012) A discovery resource of rare copy number variations in individuals with autism spectrum disorder. G3 2(12):1665–85 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Rao S, Baranova A, Yao Y, Wang J, Zhang F (2022) Genetic relationships between attention-deficit/hyperactivity disorder, autism spectrum disorder, and intelligence. Neuropsychobiology 2:1–13 [DOI] [PubMed] [Google Scholar]
  53. Rommelse NN, Franke B, Geurts HM, Hartman CA, Buitelaar JK (2010) Shared heritability of attention-deficit/hyperactivity disorder and autism spectrum disorder. Eur Child Adolesc Psychiatry 19(3):281–295 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Ronald A, Hoekstra RA (2011) Autism spectrum disorders and autistic traits: a decade of new twin studies. Am J Med Genet B 156b(3):255–274 [DOI] [PubMed] [Google Scholar]
  55. Rong Y, Yang C-J, Jin Y, Wang Y (2021) Prevalence of attention-deficit/hyperactivity disorder in individuals with autism spectrum disorder: a meta-analysis. Res Autism Spectr Disord 83:101759 [Google Scholar]
  56. Satterstrom FK, Walters RK, Singh T, Wigdor EM, Lescai F, Demontis D et al. (2019) Autism spectrum disorder and attention deficit hyperactivity disorder have a similar burden of rare protein-truncating variants. Nat Neurosci 22(12):1961–1965 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Satterstrom FK, Kosmicki JA, Wang J, Breen MS, De Rubeis S, An JY et al. (2020) Large-scale exome sequencing study implicates both developmental and functional changes in the neurobiology of autism. Cell 180(3):568–84.e23 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Schork AJ, Won H, Appadurai V, Nudel R, Gandal M, Delaneau O et al. (2019) A genome-wide association study of shared risk across psychiatric disorders implicates gene regulation during fetal neurodevelopment. Nat Neurosci 22(3):353–361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Sedlazeck FJ, Dhroso A, Bodian DL, Paschall J, Hermes F, Zook JM (2017) Tools for annotation and comparison of structural variation. F1000Res 6:1795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Smith M, Flodman PL, Gargus JJ, Simon MT, Verrell K, Haas R et al. (2012) Mitochondrial and ion channel gene alterations in autism. Biochim Biophys Acta 1817(10):1796–1802 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Stergiakouli E, Davey Smith G, Martin J, Skuse DH, Viechtbauer W, Ring SM et al. (2017) Shared genetic influences between dimensional ASD and ADHD symptoms during child and adolescent development. Mol Autism 8:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Stolerman ES, Francisco E, Stallworth JL, Jones JR, Monaghan KG, Keller-Ramey J et al. (2019) Genetic variants in the KDM6B gene are associated with neurodevelopmental delays and dysmorphic features. Am J Med Genet A 179(7):1276–1286 [DOI] [PubMed] [Google Scholar]
  63. Sudmant PH, Rausch T, Gardner EJ, Handsaker RE, Abyzov A, Huddleston J et al. (2015) An integrated map of structural variation in 2504 human genomes. Nature 526(7571):75–81 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Sun S, Miller M, Wang Y, Tyc KM, Cao X, Scott RT Jr et al. l (2022) Predicting embryonic aneuploidy rate in IVF patients using whole-exome sequencing. Hum Genet. 10.1007/s00439-022-2450-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J et al. (2019) STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 47(D1):D607–D613 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Taurines R, Schwenck C, Westerwald E, Sachse M, Siniatchkin M, Freitag C (2012) ADHD and autism: differential diagnosis or overlapping traits? A selective review. Atten Defic Hyperact Disord 4(3):115–139 [DOI] [PubMed] [Google Scholar]
  67. Vergult S, Dheedene A, Meurs A, Faes F, Isidor B, Janssens S et al. (2015) Genomic aberrations of the CACNA2D1 gene in three patients with epilepsy and intellectual disability. Eur J Hum Genet 23(5):628–632 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Vieland VJ, Huang Y, Seok SC, Burian J, Catalyurek U, O’Connell J et al. (2011) KELVIN: a software package for rigorous measurement of statistical evidence in human genetics. Hum Hered 72(4):276–288 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Wang S, Xing J (2013) A primer for disease gene prioritization using next-generation sequencing data. Genomics Inform 11(4):191–199 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Wong AK, Krishnan A, Troyanskaya OG (2018) GIANT 2.0: genome-scale integrated analysis of gene networks in tissues. Nucleic Acids Res 46(W1):W65–W70 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Yang Z, Matsumoto A, Nakayama K, Jimbo EF, Kojima K, Nagata K et al. (2016) Circadian-relevant genes are highly polymorphic in autism spectrum disorder patients. Brain Dev. 38(1):91. [DOI] [PubMed] [Google Scholar]
  72. Ye K, Schulz MH, Long Q, Apweiler R, Ning Z (2009) Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads. Bioinformatics 25(21):2865–2871 [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Zhang L, Chang S, Li Z, Zhang K, Du Y, Ott J et al. (2012) ADHDgene: a genetic database for attention deficit hyperactivity disorder. Nucleic acids Res 40:D1003–D1009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Zhou K, Dempfle A, Arcos-Burgos M, Bakker SC, Banaschewski T, Biederman J et al. (2008) Meta-analysis of genome-wide linkage scans of attention deficit hyperactivity disorder. Am J Med Genet B Neuropsychiatr Genet 147B(8):1392–1398 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Tables
NIHMS1885860-supplement-Supplemental_Tables.xlsx (5.2MB, xlsx)
Supplemental Figure S1
NIHMS1885860-supplement-Supplemental_Figure_S1.pdf (36.9KB, pdf)

Data Availability Statement

The raw sequencing reads, variants, and genotypes for all samples are available in the National Institute of Mental Health Data Archive (NDA) under experiments C1932 and C2933.

RESOURCES