SUPPLEMENTARY MATERIALS
Common genetic variants associate with risk of autism
I.SUPPLEMENTARY METHODS
- Discovery samples
- Recruitment and diagnosis
Subjects used for the genome-wide gene discovery included one cohort of case-parent triad families and one cohort of cases with autism and unrelated controls in Chinese population. Subjects were recruited mainly from the southeast and central south regions of China; and autism probands were diagnosed independently by two experienced psychiatrists using DSM-IV-TR (American Psychiatric Association, 2000) criteria for autism. The diagnostic procedure also included the assessment using a series of tools-neurological examination, mental status examination, the Childhood Autism Rating Scale (CARS)1 and Autism Behavior Checklist (ABC). Subjects with Asperger’s syndrome, Rett syndrome, and pervasive developmental disorder not otherwise specified (PPD-NOS), fragile-X syndrome were excluded from this study. Total 290 autism case-parent triad families (one affected offspring and two healthy parents) and 170 sporadic autism children were initially recruited, and self-reported ethnic group were recorded (Table 1S-1). Total 460 probands included 391 males and 69 females, and mean age of 5.08 years and mean age of onset 2.0 years (Table 1S-2).
Table 1S-1. Sample distribution by self-reported ethnicity
Ethnicity group / Total / Trios / SporadicHan Chinese / 441 / 279 / 162
Ethnic Minority Chinese / 19 / 11 / 8
Total / 460 / 290 / 170
Table 1S-2. Distribution of age of onset and age at exam for autism cases (months)
Type / Clinical assessment / Median / Mean / SD / Min / MaxSporadic / Age of onset / 24 / 24.40 / 9.00 / 8 / 53
Age of Assessment / 60 / 61.63 / 15.35 / 24 / 138
Trios / Age of onset / 24 / 23.71 / 8.38 / 8 / 48
Age of assessment / 58 / 60.59 / 22.29 / 22 / 204
Overall / Age of onset / 24 / 23.96 / 8.57 / 8 / 48
Age of assessment / 59 / 60.97 / 20.03 / 22 / 205
Note: The minimum age of onset was at 8 months, which was later according to parent's recall, but not assessed till later when subjects were recruited.
To make use of 170 sporadic cases of autism, 1,110 unrelated controls subjects (mean age 34.3±14.4 years, 660 males and 440 females) were also added to construct a case-control cohort. All control subjects were also from the southeast (813 from Anhui, 187 from Shandong provinces) and the central south region (110 from Hunan province) of China. The controls were generally matched to the cases on geographic locations. Control subjects had no history of autism or spectrum disorders, any other psychiatric diseases, familial history of psychiatric, neurological or autoimmune diseases. In summary, the study subjects including two independent cohorts of 290 case-parent triads and a case-control cohort of 1,280 subjects were genotyped for genome-wide gene discovery.
1.2 SNP Genotyping and quality controls
DNA extraction. Whole genomic DNA was extracted from the whole blood in all samples of autism cases and parents as well as unrelated normal controls for genome-wide genotyping of single nucleotide polymorphisms (SNP). The DNA was extracted using standard proteinase K digestion and phenol-chloroform method at the genotyping experiments were performed at the State Key Laboratory of Medical Genetics (SKLMG) in Changsha, China. Similarly for the additional unrelated controls the DNA extractionwas performed by Genotyping Core Facility at the State Key Laboratory Incubation Base for Dermatology in Hefei, China. The sample collections for unrelated controls have been described elsewhere (Zhang, XJ et al, Nature Genetics, 2009).
Karotype analysis. Previous reports have found that chromosome abnormality contributes to risk for autism. Autism cases with chromosome abnormality may be less likely caused by common genetic variants. We made some efforts to identify those individuals with potential chromosome abnormality for exclusions from our genome-wide association study in two Chinese samples. We performed standard G-banded chromosome analysis at 400-550 band level of autism subjects from the Chinese samples, which have lymphocytic cell sample available (400 in 460). Seven individuals were identified with chromosome abnormality.
Genome-wide SNP genotyping. All samples were genotyped using the Illumina HumanHap chips. All autism subjects and their parents if exist as well as normal 110 controls were genotyped usingthe HumanHap CNV370K; 1,000 unrelated normal control subjects were genotyped using Illumina HumanHap 610 Quad BeadChip. All SNPs in the HumanHap CNV370 are covered in the HumanHap610 Quad BeadChip.
Quality controls. Individuals with missing SNP call rate >2% were excluded; SNPs were zeroed out if Mendelian errors >5% and individuals were removed if Mendelian errors >3% for individuals from the analysis. Sample duplications and cryptic relatedness were checked by the Identity By State (IBS) analysis of genotypes data in 22 chromosomes. One of each related pair (IBD-sharing coefficients>0.10) was excluded. No sex error was founded based on heterozygote of X chromosome. As a result, a case-control cohort of 1,264 individuals (169 sporadic autisms cases and 1,095 controls) and 276 case-parent triad remained for downstream analysis.
To make better use of the two cohorts of Chinese samples, genome-wide imputation using IPMUTE2 was performed to resolve the genotyping coverage different from other three cohorts of European ancestry (see description below), in which one cohort genotyped using HumanHap 550K, two others using HumanHap 1M. The reference dataset used for the imputation was the combined SNP datasets of Chinese (CHB) and Japanese (JPT) sample from the phase II of HapMap3. The genomic interval (-int) was set to 5M, the effect size (-Ne) was set to 20000. Only SNPs with probability greater than 90% were kept in the dataset, genome-wide quality control (see below) was further applied to both cohorts after imputation . As a result, 1.07 million of SNPs, of which 702,234 SNPs were autosomal common to both cohorts and also on the HumanHap 1M chip.
1.3 Sample genetic heterogeneity reduction
Analysis of identity by state (IBS) of the whole genome genotype data and multidimensional scaling analysis were performed to examine genetic heterogeneity in each cohort. One hundred forty-four individuals (33 cases and 111 controls) were identified as outliers for removal, 1,120 (136 cases and 984 controls) remained in the case-control cohort for analysis (Figure 1S-1). Similar analysis was also performed for the triad cohort, and only one family out of 276 triads was declared as outliers in the case-parent triad sample (Figure 1S-2).
Figure 1S-1. Pair-wise plots of the first four MDS components in the case-control sample after outlier removed (red=case and blue=control).
Figure 1S-2. Pair-wise plots of the first four MDS components in the case-parent triad sample
- Three European ancestry cohorts and quality control
Three family cohorts of European ancestry were used for meta-analysis to gain adequate power to detect genetic association in our study. They were samples from autism genetic research exchange (AGRE) from Autism Speaks genotyped using the Illumina HumanHap 550K2;the Simons Foundation for Autism Research Institute (SAFRI) Base ( genotyped using HumanHap 1M3 and the Autism Genome Project(AGP) typed using the HumanHap 1M chip. AGRE and AGP were from previous published GWAS2,4. The study design, subject recruitment and sample collection have been described elsewhere2,4,5 and also see description below.
Although these three additional cohorts have been through quality controls of themselves, we further performed the same genome-wide QC process as described for the Chinese cohorts. We sequentially removed SNPs with rare allele (minor allele frequency, MAF, <1%), higher genotype missing rate (>5%) and deviated from Hardy-Weinberg proportion (p<0.0001).Individuals with SNP missing rate >2% were excluded. Mendelian error was also checked, individuals with Mendelian error>3% and SNPs with 5% were removed.
2.1 Autism Genetic Resource Exchange (AGRE)
The AGRE dataset initially comprised 4,327 individuals, of which 4,300 have phenotype data2. Table 2S-1 presents the description of affected sample by diagnosis using ADOS and ADI-R scores. By consensus criteria of both ADOS and ADI-R, 1,056 autism patients were diagnosed , and they were from 230 singleton and 471 multiplex families. For those families were multiplex, we randomly selected one affected with autism and normal parents to construct case-parent triad family for TDT analysis. After the further genome-wide quality control as described above for two Chinese cohorts, a total of 1,836 individuals from 675 nuclear families remained for analysis.
Table 2S-1. Diagnosis of all cases with autism in the AGRE sample
ADI-R criteria / ADOS criteriaAutism / Spectrum / Not Spectrum/Autism / Total
ASD / 1 / 0 / 0 / 1
Autism / 1,056 / 175 / 57 / 1,288
Broad spectrum / 36 / 34 / 31 / 101
NQA / 27 / 15 / 6 / 48
Not met / 0 / 2 / 22 / 24
PDD-NOS / 0 / 1 / 0 / 1
Total / 1,120 / 227 / 116 / 1,463
Note: ASD= autism spectrium disorder; NQA=Not quite autism.
Through IBS and MDS analysis, we found the sample with marked genetic heterogeneity, and 659 individuals (35.89%) were identified as genetic outliers for removals. The remaining sample of 1,161 individuals appear to be homogenous and was used for genetic association analysis. This removal may be stringent, because if one individuals within a triad family was removed, the whole family was not qualified for statistical analysis.
2.2 Autism Genome Project (AGP) dataset
The AGP dataset comprised 4,076 individuals (2,507 males and 1,569 females) from 1,366 families, of which 1,284 families were with at least one affected and two parents 4. About 88.71% of samples (3,582 subjects) were of European ancestry. The sample was genotyped using the Illumina Human 1M-single Infinium BeadChip array.
Further genome-wide quality control was carried out before analysisin addition to the quality control during the genotyping4. We removed 8 individuals with sex errors based on prediction of sex chromosomes, all were females. The SNP genotyping rate was 99.62% among the remaining 4,068 individuals (2,507 males and 1,561 females ), and 30 of them removed for high missing genotyping ratefor individual subject (> 0.03 ), 16,480 markers to be excluded based on HWE test ( p<= 0.001 ). After frequency and genotyping pruning, there are 801,089 SNPs and 4,038 individuals including 1,319 affected and 2,719 parents (Table 2S-2), which comprised 1,243 qualified trios families, remaining for further data analysis.
Table 2S-2. Sample distribution by number of individuals per family in AGP sample
Original sample / After QCFamily size / Freq / % / Freq / %
1 / 9 / 0.22 / 10 / 0.25
2 / 56 / 1.37 / 128 / 3.17
3 / 3,927 / 96.34 / 3,816 / 94.5
4 / 72 / 1.77 / 72 / 1.78
6 / 12 / 0.29 / 12 / 0.3
Total / 4,076 / 99.99 / 4,038 / 100
Note: QC, including genome-wide QC and Mendelian errors checking; Freq=frequency.
After genome-wide QC of the original dataset, 3,816 individuals, containing 1,243 case-parent triads, were qualified for further analysis. IBS and MDS analyses identified 799 individuals as genetic outliers. Removals of 799 outliers (20.93%) from 3,816 family resulted in a sample of 3,017 and 822 qualified triad family for further analysis. The diagnosis by subphenotype for the sample after outliers removal is presented in Table 3S-1. It is noteworthy that 325 probands (39.54%) with IQ greater than 80, and 164 proband (19.95%) with ID below 70 (Table 2S-3).
Table 2S-3. Diagnosis of proband by sub-phenotype in AGP sample after outliers removed (total number of proband=822)
Frequency / %Autism spectrum / 822 / 100
Strict autism spectrum / 478 / 58.15
Non-verbal / 274 / 33.33
Verbal / 529 / 64.36
Low_IQ<70 / 164 / 19.95
High_IQ>80 / 325 / 39.54
2.3 Simons Foundation for Autism Research Initiative (SFARI) dataset
The SFARI dataset was from the Simons Simplex Collection (SSC) and comprised 4,449 individuals (1,142 cases with autism, 2610 males and 1839 females) in 1,138 families, of which 1,132 contained one affected, 5 had two affected individuals with autism and 1 family with only one unaffected individual (Table 2S-4). Most of them were diagnosed using ADOS with a number of variables collected for probands and parents. Further genome-wide quality control was performed before any analysis.
Table 2S-4. Dirtibution of number of affecteds by family size in SFARI sample (additional 1 family with only one unnafected invidual)
Family with one affected / Family with two affectedFamily size / Subjects / % / Subjects / %
1 / 1 / 0.02
3 / 687 / 15.53
4 / 3180 / 71.86 / 8 / 34.78
5 / 435 / 9.83 / 15 / 65.22
6 / 108 / 2.44
7 / 14 / 0.32
Total / 4425 / 100 / 23 / 100
Similarly, the IBS and MDS analysis of 4,449 individuals identified 1,152 individuals (25.89%) as genetic outliers. Remaining sample comprised of 3,297 individuals (835 cases and 2,462 parents), of which 689 trios were qualified for analysis and 615 probands (89.26%) had IQ greater than 80.
- Power calculation
To assist in selection of the cutoff threshold for individual SNPs, we performed power analysis based on the actual sample size in both Chinese cohorts, assuming 1% population risk and under additive model 6,7. As it showed both Chinese cohorts only provide reasonable power to detect genetic association for moderate effect size even at alpha of 0.1.
Table 3S-1. Power analysis of a case-control cohort of 136 cases and 984 controls
MAFRR / 0.01- / 0.06- / 0.11- / 0.16- / 0.21- / 0.26- / 0.31- / 0.36- / 0.41- / 0.46-
Alpha=0.05
1.200 / 0.060 / 0.108 / 0.151 / 0.188 / 0.219 / 0.244 / 0.264 / 0.278 / 0.286 / 0.290
1.400 / 0.088 / 0.271 / 0.419 / 0.529 / 0.608 / 0.662 / 0.699 / 0.721 / 0.733 / 0.735
1.600 / 0.132 / 0.496 / 0.714 / 0.828 / 0.888 / 0.920 / 0.938 / 0.947 / 0.950 / 0.950
1.800 / 0.190 / 0.709 / 0.899 / 0.960 / 0.981 / 0.989 / 0.993 / 0.994 / 0.995 / 0.994
2.000 / 0.259 / 0.860 / 0.973 / 0.994 / 0.998 / 0.999 / 0.999 / 1.000 / 1.000 / 1.000
Alpha=0.10
1.200 / 0.115 / 0.182 / 0.239 / 0.286 / 0.324 / 0.354 / 0.377 / 0.393 / 0.402 / 0.406
1.400 / 0.155 / 0.385 / 0.544 / 0.651 / 0.722 / 0.768 / 0.798 / 0.816 / 0.826 / 0.828
1.600 / 0.215 / 0.620 / 0.810 / 0.897 / 0.937 / 0.958 / 0.968 / 0.973 / 0.975 / 0.975
1.800 / 0.288 / 0.807 / 0.944 / 0.980 / 0.992 / 0.996 / 0.997 / 0.998 / 0.998 / 0.998
2.000 / 0.371 / 0.918 / 0.988 / 0.997 / 0.999 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000
Note: Highlighted are where power below 55%.
Table 3S-2. Power analysis of 275 trios under additive model and 1% population risk
MAFRR / 0.01- / 0.06- / 0.11- / 0.16- / 0.21- / 0.26- / 0.31- / 0.36- / 0.41- / 0.46-
Alpha=0.05
1.2 / 0.062 / 0.116 / 0.166 / 0.208 / 0.244 / 0.273 / 0.296 / 0.312 / 0.323 / 0.327
1.4 / 0.093 / 0.297 / 0.461 / 0.580 / 0.662 / 0.719 / 0.756 / 0.779 / 0.792 / 0.795
1.6 / 0.140 / 0.536 / 0.759 / 0.868 / 0.921 / 0.948 / 0.962 / 0.969 / 0.972 / 0.972
1.8 / 0.202 / 0.747 / 0.924 / 0.974 / 0.989 / 0.995 / 0.997 / 0.998 / 0.998 / 0.998
2.0 / 0.273 / 0.885 / 0.983 / 0.997 / 0.999 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000
Alpha=0.10
1.2 / 0.117 / 0.193 / 0.258 / 0.311 / 0.354 / 0.388 / 0.413 / 0.431 / 0.443 / 0.448
1.4 / 0.161 / 0.415 / 0.586 / 0.697 / 0.769 / 0.814 / 0.843 / 0.861 / 0.870 / 0.873
1.6 / 0.226 / 0.657 / 0.845 / 0.924 / 0.958 / 0.974 / 0.982 / 0.985 / 0.987 / 0.987
1.8 / 0.303 / 0.836 / 0.960 / 0.988 / 0.996 / 0.998 / 0.999 / 0.999 / 0.999 / 0.999
2.0 / 0.388 / 0.935 / 0.992 / 0.999 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000
Alpha=0.15
1.2 / 0.170 / 0.259 / 0.331 / 0.389 / 0.434 / 0.469 / 0.495 / 0.514 / 0.525 / 0.531
1.4 / 0.222 / 0.497 / 0.664 / 0.765 / 0.826 / 0.864 / 0.888 / 0.901 / 0.909 / 0.911
1.6 / 0.295 / 0.729 / 0.889 / 0.949 / 0.973 / 0.984 / 0.989 / 0.991 / 0.992 / 0.993
1.8 / 0.380 / 0.882 / 0.975 / 0.993 / 0.998 / 0.999 / 0.999 / 1.000 / 1.000 / 1.000
2.0 / 0.469 / 0.957 / 0.996 / 0.999 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000 / 1.000
Note: Highlighted are where power below 55%.
- Genetic association analysis
We first performed TDT analysis of the case-parent triad cohorts and logistic regression analysis for the case-control cohort. Genetic model was also assessed for the case-control data. Because each of two GWAS cohorts in Chinese alone may not able to provide enough power according to our power analysis, we set alpha=0.10for both the case-parent triad and the case-control cohort. Individual SNPs passed the threshold and combined analysis showed nominal significance of association (p<0.01) were included for meta-analysis with SNPs that showed same direction of over transmitted allele in three additional case-parent triad cohorts of European ancestry. Since there were marked heterogeneous in autism IQ between AGP and SFARI cohorts, we only included triad samples that probands had IQ greater than 80 for the analysis.
We then obtained combined evidence for genome-wide significant (p<10x10-8, which was <0.05/500,000 SNPs tested, using the Bonferroni correction for multiple testing) association through meta-analysis of five cohorts. The combined p values were calculated using Stouffer’s Z-score method, a modified form of Fisher’s combination method for overall meta-analysis.
where k is the number of independent cohorts, andis the inverse of the standard normal cumulative distribution based on the p value for an individual SNP in the th cohort.
1
II. SUPPLEMENTARY TABLES
Table S1. Top signals in the combined analysis of two Chinese cohorts (Observed SNPs)
Case-parent triad cohort / Case-control cohortChr / BP / SNP / Gene / A1 / A2 / OR / P / OR / P / Pcomb
1 / 114749804 / rs6537825 / TRIM33 / A / G / 1.596 / 2.33E x 10-4 / 1.38 / 1.00 x 10-3 / 8.27 x 10-7
1 / 114875303 / rs10858046 / TRIM33 / T / C / 1.495 / 1.32 x 10-3 / 1.53 / 7.19 x 10-5 / 3.80 x 10-7
1 / 114875399 / rs10858047 / TRIM33 / C / T / 1.448 / 3.37 x 10-3 / 1.34 / 4.55 x 10-3 / 4.51 x 10-5
1 / 114900615 / rs1877455 / BCAS2 / A / G / 0.6543 / 6.25 x 10-4 / 0.57 / 7.40 x 10-5 / 1.78 x 10-7
1 / 114922984 / rs4839385 / BCAS2 / G / A / 1.409 / 4.74 x 10-3 / 1.85 / 2.02 x 10-5 / 5.41 x 10-7
1 / 115029016 / rs2268697 / AMPD1 / G / A / 1.469 / 1.51 x 10-3 / 1.76 / 1.00 x 10-4 / 5.90 x 10-7
1 / 115029419 / rs2268699 / AMPD1 / A / C / 1.412 / 4.59 x 10-3 / 1.68 / 3.38 x 10-4 / 5.67 x 10-6
1 / 115041339 / rs926938 / AMPD1//NRAS / A / G / 1.477 / 1.39 x 10-3 / 1.73 / 8.99 x 10-5 / 4.92 x 10-7
1 / 115065208 / rs10489525 / CSDE1 / A / G / 0.7248 / 1.05 x 10-2 / 0.60 / 4.54 x 10-4 / 1.80 x 10-5
3 / 21295122 / rs9859926 / VENTXP7 / G / T / 0.6859 / 2.52 x 10-3 / 0.70 / 1.19 x 10-3 / 9.51 x 10-6
3 / 64771010 / rs423536 / ADAMTS9 / T / C / 1.424 / 2.69 x 10-2 / 1.47 / 1.92 x 10-4 / 2.65 x 10-5
5 / 86948822 / rs11749237 / CCNH / A / G / 1.579 / 1.90 x 10-4 / 1.33 / 2.44 x 10-2 / 2.33 x 10-5
5 / 86949700 / rs13185988 / CCNH / C / T / 1.589 / 1.54 x 10-4 / 1.33 / 2.44 x 10-2 / 1.97 x 10-5
5 / 86980173 / rs3909481 / CCNH / A / G / 1.589 / 1.54 x 10-4 / 1.34 / 2.25 x 10-2 / 1.78 x 10-5
5 / 86987264 / rs13178161 / CCNH / T / C / 1.589 / 1.54 x 10-4 / 1.34 / 2.25 x 10-2 / 1.78 x 10-5
5 / 178449089 / rs1562234 / ZNF354C / G / A / 1.389 / 7.41 x 10-3 / 1.50 / 1.18 x 10-3 / 2.83 x 10-5
7 / 7648964 / rs3757526 / RPA3 / T / G / 2.444 / 9.60 x 10-4 / 1.94 / 9.06 x 10-3 / 2.91 x 10-5
7 / 7707439 / rs7800009 / RPA3 / T / G / 2.368 / 1.15 x 10-3 / 1.93 / 9.42 x 10-3 / 3.57 x 10-5
7 / 77065177 / rs6955503 / PTPN12 / T / C / 0.5213 / 1.68 x 10-4 / 0.66 / 4.28 x 10-2 / 4.25 x 10-5
8 / 57013973 / rs1126327 / LYN / A / G / 2.714 / 1.86 x 10-2 / 2.88 / 2.20 x 10-4 / 1.89 x 10-5
11 / 85350031 / rs2077815 / PICALM / G / A / 1.317 / 3.35 x 10-2 / 1.60 / 1.52 x 10-4 / 2.89 x 10-5
11 / 85379778 / rs7938033 / PICALM / A / C / 1.358 / 1.63 x 10-2 / 1.58 / 1.66 x 10-4 / 1.29 x 10-5
11 / 85393384 / rs527162 / PICALM / C / T / 1.291 / 5.08 x 10-2 / 1.60 / 5.33 x 10-5 / 2.26 x 10-5
11 / 85393680 / rs680119 / PICALM / T / C / 1.324 / 3.21 x 10-2 / 1.58 / 2.95 x 10-4 / 4.59 x 10-5
11 / 85458970 / rs669556 / PICALM / C / T / 1.262 / 7.19 x 10-2 / 1.56 / 5.33 x 10-5 / 3.63 x 10-5
13 / 109268342 / rs336230 / IRS2 / A / G / 1.434 / 5.61 x 10-2 / 1.47 / 1.15 x 10-3 / 2.06 x 10-5
14 / 52883170 / rs1255329 / DDHD1 / T / C / 1.41 / 3.92 x 10-2 / 1.76 / 1.63 x 10-4 / 3.73 x 10-5
14 / 52886261 / rs1255320 / DDHD1 / C / T / 1.41 / 3.92 x 10-2 / 1.76 / 1.63 x 10-4 / 3.73 x 10-5
15 / 34933282 / rs2167194 / MEIS2 / G / A / 1.339 / 9.69 x 10-2 / 1.60 / 2.22 x 10-5 / 3.00 x 10-5
18 / 45400123 / rs11664342 / LIPG / G / A / 1.537 / 1.02 x 10-3 / 1.44 / 1.36 x 10-2 / 4.73 x 10-5
18 / 46007489 / rs894771 / CCDC11 / C / T / 2.833 / 2.18 x 10-2 / 3.59 / 4.32 x 10-5 / 6.37 x 10-6
22 / 16085622 / rs5748965 / CECR1 / T / C / 1.384 / 2.12 x 10-2 / 1.66 / 3.87 x 10-4 / 3.48 x 10-5
22 / 23119632 / rs6004146 / KIAA0376 / T / C / 0.359 / 5.95 x 10-4 / 0.29 / 1.30 x 10-2 / 2.86 x 10-5
Table S2. Functional SNPs in LD (r2>0.5)with the genome-wide associated loci or SNPsinvolved with three haplotypes associated with autism in Chinese population
SNP / Position / Allele / LDsnp / TFBS / Splicing / miRNA / nsSNP / Reg. / Consved / Gene / Allele / Asian / CHBrs3789615 / 114742849 / C/T / rs11102800|rs6537825|rs3827735 / -- / -- / -- / -- / 0 / 0.988 / TRIM33 / T / 0.456 / 0.488
rs6537825 / 114749804 / A/G / rs6537825|rs11102800|rs10858047 / -- / -- / -- / Y / 0.32 / 1 / TRIM33 / G / 0.663 / 0.625
rs12083584 / 114774952 / C/T / rs11582563|rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / -- / Y / -- / -- / 0.26 / 1 / TRIM33 / T / 0.961 / 0.952
rs11585926 / 114775213 / C/T / rs11585926|rs11582563|rs7511633|rs6661053|rs11102807|rs11589568 / -- / -- / -- / -- / 0 / 0.469 / TRIM33 / T / 0.972 / 0.97
rs6670743 / 114813007 / C/T / rs11582563|rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / Y / -- / -- / -- / 0.19 / 1 / TRIM33 / T / 0.955 / 0.952
rs7520209 / 114821233 / C/T / rs11102800|rs1877455|rs3827735 / -- / -- / -- / -- / NA / 0.998 / TRIM33 / C / 0.579 / 0.554
rs12078204 / 114824708 / A/C / rs11582563|rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / -- / -- / -- / -- / NA / 0.386 / TRIM33 / A / 0.961 / 0.944
rs11102807 / 114863107 / A/G / rs11102807|rs11582563|rs7511633|rs6661053|rs11589568|rs11585926 / Y / -- / -- / -- / 0 / 0 / TRIM33 / A / 0.579 / 0.952
rs11102808 / 114863165 / A/G / rs11582563|rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / Y / -- / -- / -- / 0 / 0.002 / TRIM33 / A / 0.96 / 0.952
rs10858046 / 114875303 / C/T / rs10858047|rs6537825|rs1877455|rs926938 / -- / -- / -- / -- / 0 / 0.446 / TRIM33 / C / 0.644 / 0.571
rs10858047 / 114875399 / C/T / rs10858047|rs6537825|rs926938 / -- / -- / -- / -- / NA / 0.482 / TRIM33 / T / 0.683 / 0.619
rs2137365 / 114880208 / C/G / rs10858047|rs6537825|rs1877455|rs926938 / -- / -- / -- / -- / 0.5 / 0.996 / TRIM33 / G / 0.175 / 0.565
rs11587400 / 114881181 / C/T / rs11587400 / -- / -- / -- / -- / 0.45 / 1 / TRIM33 / C / 0.889 / 0.917
rs8128 / 114912206 / C/A / rs10858047|rs6537825|rs1877455|rs926938 / -- / -- / Y / -- / 0 / 0 / BCAS2 / A / 0.628 / 0.56
rs3789624 / 114917957 / A/G / rs10858047|rs6537825|rs926938 / -- / -- / -- / -- / 0 / 0.429 / BCAS2 / A / 0.624 / 0.633
rs1564824 / 114925270 / G/A / rs10858047|rs6537825|rs1877455|rs926938 / Y / -- / -- / -- / 0 / 0 / BCAS2 / A / 0.652 / 0.583
rs222493 / 114929610 / G/A / rs10858047|rs6537825|rs1877455|rs926938 / Y / -- / Y / -- / 0 / 0 / DENND2C / G / 0.652 / 0.571
rs222498 / 114947209 / C/T / rs10858047|rs6537825|rs1877455|rs926938 / -- / -- / -- / -- / 0 / 0.374 / DENND2C / C / 0.588 / 0.554
rs12136548 / 114969620 / C/T / rs11587400 / -- / -- / -- / Y / 0.09 / 0.425 / DENND2C / T / 0.344 / 0.911
rs7533963 / 114989320 / C/G / rs10858047|rs6537825|rs1877455|rs926938 / -- / -- / -- / -- / 0 / 0.477 / DENND2C / C / 0.635 / 0.602
rs11803022 / 115014095 / A/G / rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / Y / -- / -- / -- / 0.44 / 0 / DENND2C / A / -- / 0.97
rs6679869 / 115019250 / C/G / rs7511633|rs6661053|rs11102807|rs11589568|rs11585926 / Y / -- / -- / -- / 0 / 0.001 / AMPD1 / G / 0.978 / 0.97
rs11588144 / 115040878 / A/G / rs10858047|rs1877455|rs926938|rs10489525 / Y / -- / -- / -- / NA / 0.004 / AMPD1||NRAS / A / 0.614 / 0.566
rs926938 / 115041339 / G/A / rs926938|rs10858047|rs1877455|rs10489525 / Y / -- / -- / -- / 0 / 0.006 / AMPD1||NRAS / G / 0.624 / 0.542
rs11587596 / 115043380 / A/G / rs10858047|rs1877455|rs926938 / Y / -- / -- / -- / 0 / 0.028 / AMPD1||NRAS / G / 0.624 / 0.567
rs14804 / 115051366 / G/A / rs11102807 / -- / -- / Y / -- / 0 / 0.897 / NRAS / G / 0.956 / 0.969
rs8453 / 115061122 / T/G / rs8453 / -- / -- / -- / -- / 0.32 / 1 / CSDE1 / G / 0.876 / 0.91
rs10489525 / 115065208 / A/G / rs10489525|rs1877455|rs926938 / Y / -- / -- / -- / 0 / 0.255 / CSDE1 / G / 0.571 / 0.577
rs6692870 / 115065314 / C/T / rs10858047|rs926938|rs10489525 / Y / -- / -- / -- / 0 / 0 / CSDE1 / T / 0.618 / 0.577
rs6668128 / 115065512 / A/G / rs11102807 / Y / -- / -- / -- / 0 / 0 / CSDE1 / G / 0.972 / 0.97
rs7555948 / 115089224 / C/T / rs11102807 / -- / -- / -- / -- / 0 / 0.461 / CSDE1 / C / 0.978 / 0.966
rs12140675 / 115104402 / C/T / rs11102807 / Y / -- / -- / -- / 0 / 0.489 / RP5-1000E10.4 / T / 0.971 / 0.97
Table S3. Functional SNPs in LD (r2>0.50) withgenome-wide associated loci or SNPs involved with three haplotypes associated with autism in European ancestry population
SNPs / Position / Allele / LDsnp / TFBS / Splicing / miRNA / nsSNP / Reg / Consverd / Nearby Gene / Allele / Asian / CEUrs3789615 / 114742849 / C/T / rs11102800|rs7511633|rs6661053 / -- / -- / -- / -- / 0 / 0.988 / TRIM33 / T / 0.456 / 0.438
rs6537825 / 114749804 / A/G / rs6537825 / -- / -- / -- / Y / 0.316 / 1 / TRIM33 / G / 0.663 / 0.920
rs12083584 / 114774952 / C/T / rs11102800|rs7511633|rs6661053|rs11102807|rs11585926 / -- / Y / -- / -- / 0.26 / 1 / TRIM33 / T / 0.961 / 0.680
rs11585926 / 114775213 / C/T / rs11585926|rs11582563|rs7511633|rs6661053|rs11589568 / -- / -- / -- / -- / 0 / 0.469 / TRIM33 / T / 0.972 / 0.732
rs6670743 / 114813007 / C/T / rs11102800|rs7511633|rs6661053|rs11102807|rs11585926 / Y / -- / -- / -- / 0.187 / 1 / TRIM33 / T / 0.955 / 0.674
rs7520209 / 114821233 / C/T / rs3827735 / -- / -- / -- / -- / NA / 0.998 / TRIM33 / C / 0.579 / 0.893
rs11102806 / 114856906 / A/G / rs11102807 / Y / -- / -- / -- / NA / 0 / TRIM33 / A / -- / 0.509
rs11102807 / 114863107 / A/G / rs11102807|rs7511633|rs6661053 / Y / -- / -- / -- / 0 / 0 / TRIM33 / A / 0.579 / 0.540
rs11102808 / 114863165 / A/G / rs7511633|rs6661053|rs11102807 / Y / -- / -- / -- / 0 / 0.002 / TRIM33 / A / 0.96 / 0.545
rs10858046 / 114875303 / C/T / rs11587400 / -- / -- / -- / -- / 0 / 0.446 / TRIM33 / C / 0.644 / 0.402
rs10858047 / 114875399 / C/T / rs10858047 / -- / -- / -- / -- / NA / 0.482 / TRIM33 / T / 0.683 / 0.839
rs2137365 / 114880208 / C/G / rs10858047 / -- / -- / -- / -- / 0.496 / 0.996 / TRIM33 / G / 0.175 / 0.835
rs11587400 / 114881181 / C/T / rs11587400 / -- / -- / -- / -- / 0.45 / 1 / TRIM33 / C / 0.889 / 0.652
rs6674801 / 114882040 / A/G / rs11587400 / -- / -- / -- / -- / 0.31 / 0 / BCAS2 / G / 0.831 / 0.460
rs3789624 / 114917957 / A/G / rs10858047 / -- / -- / -- / -- / 0 / 0.429 / BCAS2 / A / 0.624 / 0.871
rs1564824 / 114925270 / G/A / rs10858047 / Y / -- / -- / -- / 0 / 0 / BCAS2 / A / 0.652 / 0.853
rs222493 / 114929610 / G/A / rs10858047 / Y / -- / Y / -- / 0 / 0 / DENND2C / G / 0.652 / 0.853
rs12136548 / 114969620 / C/T / rs11587400 / -- / -- / -- / Y / 0.086 / 0.425 / DENND2C / T / 0.344 / 0.661
rs178445 / 114971389 / A/G / rs10858047 / -- / -- / -- / -- / 0.204 / 0.001 / DENND2C / A / 0.652 / 0.83
rs7533963 / 114989320 / C/G / rs11587400 / -- / -- / -- / -- / 0 / 0.477 / DENND2C / C / 0.635 / 0.517
rs6667218 / 115040505 / A/G / rs926938 / Y / -- / -- / -- / NA / 0 / AMPD1||NRAS / A / 1 / 0.643
rs11588144 / 115040878 / A/G / rs926938 / Y / -- / -- / -- / NA / 0.004 / AMPD1||NRAS / A / 0.614 / 0.531
rs926938 / 115041339 / G/A / rs926938 / Y / -- / -- / -- / 0 / 0.006 / AMPD1||NRAS / G / 0.624 / 0.500
rs11587596 / 115043380 / A/G / rs926938 / Y / -- / -- / -- / 0 / 0.028 / AMPD1||NRAS / G / 0.624 / 0.592
rs8453 / 115061122 / T/G / rs8453 / -- / -- / -- / -- / 0.324 / 1 / CSDE1 / G / 0.876 / 0.808
rs10489525 / 115065208 / A/G / rs10489525 / Y / -- / -- / -- / 0 / 0.255 / CSDE1 / G / 0.571 / 0.743
rs1286560 / 115177453 / T/C / rs926938 / Y / -- / -- / -- / NA / 0 / RP5-1000E10.4 / C / 0.461 / 0.533
rs360617 / 115205622 / G/A / rs926938 / -- / -- / -- / -- / 0 / 0.518 / SYCP1 / G / 0.532 / 0.475
rs360661 / 115256803 / G/A / rs926938 / -- / -- / -- / -- / 0 / 0.418 / SYCP1 / G / 0.526 / 0.381
rs360667 / 115282974 / C/G / rs926938 / -- / -- / -- / -- / 0 / 0.293 / SYCP1 / G / 0.556 / 0.405
Note: SNPs listed have functional evidence for either a TFBS, splicing, coding, has regulation potential (predicted score>0.25) or evidence for conservation (>0.25); LDsnp, SNPs or SNPs in three haplotypes associated with autism.
Table S4. Permutation test for three haplotype associations at three loci in five cohorts
Permutation p valueHaplotype / Type / SKLMGtrios / SKLMGcc / AGRE / SFARI / AGP / P_comb
AMPD1-NRAS-CASDE1:rs926938|rs8453|rs10489525
AGG / Chisq sum / 0.006101 / 0.0002 / 0.21992 / 0.36231 / 0.17442 / 8.5E-06
AGG / minp / 0.020324 / 0.0002 / 0.32246 / 0.45749 / 0.17289 / 4.4E-05
TRIM33: rs6537825|rs11582563|rs11585926|rs11589568|rs7511633|rs6661053|rs11102800|rs3827735|rs11102807
AGTTGTCCA / Chisq sum / 0.00073 / 0.0202 / 0.54248 / 0.17430 / 0.16496 / 5.1E-05
AGTTGTCCA / minp / 0.00088 / 0.0202 / 0.15480 / 0.32585 / 0.25317 / 3.9E-05
TRIM33-BCAS2: rs10858047|rs11587400|rs1877455
TCT / Chisq sum / 0.006101 / 0.0002 / 0.14913 / 0.04337 / 0.70692 / 4.1E-06
TCT / minp / 0.020324 / 0.0002 / 0.17036 / 0.02170 / 0.76730 / 7.7E-06
Note: the permutation test was performed for whole marker for sum of Chi-square (Chisq sum) and min p value (minp), while the case-control data, permutation test was performed for specific haplotype.
P_comb, combined p values of permutation p values in five cohorts.
1
Table S5. The cis-association analysis of SNPs involved with three autism associated haplotypes with gene expressions in prefrontal cortex of postnatal human brains in Caucasian and African American sample, collected by CBDB/NIMH
Caucasian + AA(n=224) / Caucasian (n=109)Probes / Source / Type III SS / MS / F / P / Type III SS / MS / F / P
SNPs rs926938-rs8453-rs10489525 with AMPD1-NRAS-CSDE1
AMPD1 / rs926938 / 0.1249 / 0.0624 / 0.76 / 0.47 / 0.1657 / 0.0828 / 1.04 / 0.358
AMPD1 / rs8453 / 0.2302 / 0.1151 / 1.41 / 0.246 / 0.1617 / 0.0809 / 1.01 / 0.367
AMPD1 / rs10489525 / 0.0699 / 0.0349 / 0.42 / 0.657 / 0.1097 / 0.0548 / 0.68 / 0.5078
NRAS / rs926938 / 0.3995 / 0.1998 / 1.27 / 0.284 / 0.0268 / 0.0134 / 0.10 / 0.9067
NRAS / rs8453 / 0.9118 / 0.4559 / 2.92 / 0.056 / * / 1.0074 / 0.5037 / 3.94 / 0.0223 / **
NRAS / rs10489525 / 0.8223 / 0.4112 / 2.62 / 0.075 / 0.1042 / 0.0521 / 0.38 / 0.6835
CSDE1.1 / rs926938 / 0.4792 / 0.2396 / 2.98 / 0.053 / * / 0.1576 / 0.0788 / 0.90 / 0.4113
CSDE1.1 / rs8453 / 0.6942 / 0.3471 / 4.45 / 0.013 / ** / 0.7995 / 0.3997 / 4.89 / 0.0093 / **
CSDE1.1 / rs10489525 / 0.5378 / 0.2689 / 3.36 / 0.036 / ** / 0.432 / 0.216 / 2.53 / 0.0843 / *
CSDE1.2 / rs926938 / 0.0109 / 0.0054 / 0.02 / 0.985 / 0.2255 / 0.1128 / 0.21 / 0.8082
CSDE1.2 / rs8453 / 4.7676 / 2.3838 / 7.3 / 9E-04 / ** / 3.2951 / 1.6475 / 3.30 / 0.0407 / **
CSDE1.2 / rs10489525 / 0.3875 / 0.1937 / 0.55 / 0.58 / 0.2195 / 0.1098 / 0.21 / 0.8109
CSDE1.3 / rs926938 / 0.9489 / 0.4744 / 2.89 / 0.058 / * / 0.2464 / 0.1232 / 0.59 / 0.5565
CSDE1.3 / rs8453 / 1.0241 / 0.5121 / 3.13 / 0.046 / ** / 0.9551 / 0.4776 / 2.36 / 0.0993 / *
CSDE1.3 / rs10489525 / 1.3038 / 0.6519 / 4.05 / 0.019 / ** / 0.4478 / 0.2239 / 1.07 / 0.3463
SNPs in the second haplotype with TRIM33
TRIM33 / rs6537825 / 0.099 / 0.099 / 1.14 / 0.286 / 0.116 / 0.116 / 1.42 / 0.2354
TRIM33 / rs11582563 / 0.2329 / 0.1165 / 1.35 / 0.262 / 0.1199 / 0.060 / 0.73 / 0.4845
TRIM33 / rs11585926 / 0.2844 / 0.1422 / 1.65 / 0.195 / 0.1195 / 0.0598 / 0.73 / 0.4856
TRIM33 / rs11589568 / 0.2844 / 0.1422 / 1.65 / 0.195 / 0.1195 / 0.0598 / 0.73 / 0.4856
TRIM33 / rs7511633 / 0.1963 / 0.0982 / 1.13 / 0.324 / 0.4498 / 0.2249 / 2.85 / 0.0625 / *
TRIM33 / rs6661053 / 0.1906 / 0.0953 / 1.1 / 0.335 / 0.4498 / 0.2249 / 2.85 / 0.0625 / *
TRIM33 / rs11102800 / 0.4503 / 0.2251 / 2.64 / 0.074 / * / 0.4673 / 0.2337 / 2.97 / 0.056 / *
TRIM33 / rs3827735 / 0.1737 / 0.0868 / 1 / 0.369 / 0.0113 / 0.0056 / 0.07 / 0.9347
TRIM33 / rs1102807 / 0.4597 / 0.2299 / 2.68 / 0.071 / * / 0.4721 / 0.2361 / 3.00 / 0.0543 / *
SNP rs10858047-rs11587400-rs1877455with TRIM33-BCAS2
BCAS2.1 / rs1877455 / 0.0862 / 0.0431 / 0.39 / 0.681 / 0.0622 / 0.0311 / 0.27 / 0.7627
BCAS2.1 / rs10858047 / 0.0153 / 0.0077 / 0.04 / 0.9581 / 0.0925 / 0.0925 / 0.55 / 0.4612
BCAS2.1 / s11587400 / 0.1390 / 0.0695 / 0.62 / 0.5372 / 0.1012 / 0.0506 / 0.45 / 0.6390
BCAS2.2 / rs1877455 / 0.4360 / 0.2180 / 1.21 / 0.3016 / 0.1378 / 0.0689 / 0.40 / 0.6727
BCAS2.2 / rs10858047 / 0.0153 / 0.0077 / 0.04 / 0.9581 / 0.0925 / 0.0925 / 0.55 / 0.4612
BACS2.2 / Rs11587400 / 0.0688 / 0.0344 / 0.40 / 0.6708 / 0.1673 / 0.0837 / 1.04 / 0.3577
TRIM33 / rs1877455 / 0.1270 / 0.0635 / 0.75 / 0.4725 / 0.0243 / 0.0122 / 0.15 / 0.8588
TRIM33 / rs10858047 / 0.0300 / 0.0150 / 0.17 / 0.8403 / 0.0218 / 0.0218 / 0.27 / 0.6057
TRIM33 / rs11587400 / 0.0688 / 0.0344 / 0.40 / 0.6708 / 0.1673 / 0.0837 / 1.04 / 0.3577
Note: Highlighted, showing consistent cis-association in both combined sample and Caucasian only; **p<0.05, *P<0.1.
1
Table S6. Descriptive statistics of mRNA expressions (log2 intensity ratio) in postmortem human prefrontal and temporal cortex between autism and normal controls of genes at the genome-wide associated loci (1p13.2)
Frontal cortex / Temporal cortexAutism (n=16) / Control (n=16) / Autism (n=10) / Control (n=11)
Probe / Mean / SD / Median / Mean / SD / Median / Mean / SD / Median / Mean / SD / Median
AMPD1 / ILMN_1759312 / 7.70 / 0.29 / 7.70 / 7.85 / 0.20 / 7.82 / 8.241 / 0.069 / 8.252 / 8.177 / 0.06 / 8.174
NRAS / ILMN_1775759 / 9.93 / 0.44 / 9.93 / 9.90 / 0.24 / 9.94 / 9.582 / 0.207 / 9.482 / 9.489 / 0.231 / 9.416
CSDE1.1 / ILMN_1685796 / 9.64 / 0.45 / 9.63 / 9.97 / 0.27 / 9.99 / 9.942 / 0.247 / 9.986 / 9.879 / 0.249 / 9.985
CSDE1.2 / ILMN_1683538 / 8.96 / 0.24 / 8.97 / 9.09 / 0.24 / 9.11 / 9.39 / 0.271 / 9.328 / 9.529 / 0.265 / 9.432
BCAS2 / ILMN_2148452 / 8.46 / 0.39 / 8.54 / 8.67 / 0.31 / 8.69 / 8.726 / 0.107 / 8.721 / 8.702 / 0.124 / 8.697
DENND2C / ILMN_2106573 / 8.39 / 0.28 / 8.43 / 8.38 / 0.34 / 8.32 / 8.674 / 0.098 / 8.686 / 8.602 / 0.12 / 8.614
TRIM33.1 / ILMN_2351930 / 8.87 / 0.37 / 8.97 / 8.99 / 0.25 / 8.94 / 9.192 / 0.181 / 9.226 / 9.184 / 0.084 / 9.165
TRIM33.2 / ILMN_1682316 / 12.1 / 0.47 / 12.25 / 12.46 / 0.24 / 12.45 / 12.456 / 0.208 / 12.407 / 12.54 / 0.303 / 12.541
TRIM33.3 / ILMN_1738672 / 7.67 / 0.24 / 7.75 / 7.87 / 0.24 / 7.79 / 8.243 / 0.097 / 8.213 / 8.256 / 0.085 / 8.253
PMI / 20.08 / 8.87 / 23.58 / 6.37 / 21.337 / 10.017 / 22.160 / 21.43 / 6.22 / 22.12
RIN / 6.78 / 1.07 / 7.52 / 0.66 / NA / NA
Age / 24.56 / 13.75 / 33.25 / 13.1 / 22.000 / 15.048 / 19.000 / 37.45 / 14.75 / 36.00
Table S7. Differential expression of 9 probes in human frontal and temporal cortex between 16 autisms and 16 normal controls based on post-hoc least-square mean estimates from MANOVA analysis (Roy’s largest root based F exact statistics).
Frontal cortex / Temporal cortexAutism(n=16) / Control(n=16) / Autism(n=10) / Control(n=11)
Est / Se. / Est / Se. / p / Est / Se. / Est / Se. / P
AMPD1 / 8.249 / 0.024 / 8.227 / 0.026 / 0.4875 / 8.193 / 0.021 / 8.171 / 0.024 / 0.4133
NRAS / 9.915 / 0.072 / 9.712 / 0.079 / 0.0463 / ** / 9.872 / 0.072 / 9.605 / 0.081 / 0.0099 / ***
CSDE1.1 / 9.692 / 0.060 / 9.851 / 0.065 / 0.0570 / * / 9.725 / 0.078 / 9.774 / 0.087 / 0.6307
CSDE1.2 / 9.136 / 0.042 / 9.132 / 0.046 / 0.9383 / 9.218 / 0.079 / 9.170 / 0.089 / 0.6471
BCAS2 / 8.670 / 0.037 / 8.798 / 0.040 / 0.0155 / ** / 8.741 / 0.045 / 8.796 / 0.051 / 0.3566
DENND2C / 8.651 / 0.042 / 8.594 / 0.046 / 0.3209 / 8.657 / 0.040 / 8.589 / 0.045 / 0.2038
TRIM33.1 / 9.064 / 0.037 / 8.958 / 0.041 / 0.0431 / ** / 9.031 / 0.030 / 8.937 / 0.034 / 0.0255 / **
TRIM33.2 / 12.120 / 0.044 / 12.154 / 0.048 / 0.5706 / 12.170 / 0.067 / 12.099 / 0.075 / 0.4203
TRIM33.3 / 8.243 / 0.019 / 8.297 / 0.020 / 0.0423 / ** / 8.232 / 0.027 / 8.241 / 0.031 / 0.8017
Overall F 9,17=2.78 / 0.0332 / ** / Overall, F9,11=1.87 / 0.162
Note: *= p<0.1; **=p<0.05; ***=p<0.01; age, sex, RIN and PMI were not significant.
1
III. SUPPLEMENTARY FIGURES
Figure S1. Chromosomal plot and Q-Q plot of p values from TDT analysis of case-parent triad cohort in Chinese
a) Chromosome plot: y axis is logarithm of p values, x axis is chromosome
b) Q_Q plot
Figure S2. Chromosomal plot and Q-Q plot of p values from case-control cohort in Chinese using logistic regression
a) Chromosome plot:y axis is logarithm of p values, x axis is chromosome