Introduction 1 3 2 4 5 7 8 9 There are two approaches to identifying genetic contributors to disease. The first is a genome wide search in which linkage or association analysis is used to identify regions of the genome that may contain autism susceptibility genes. The second is the candidate gene approach, which investigates a specific gene or genes for involvement in autism risk. In the candidate gene approach, genes are chosen for study based on either what is known about the gene’s function, its location (for example in a recognized linkage peak), or a combination of both. Several candidates are hypothesized to be involved in autism; however, no single candidate gene has consistently emerged as involved in autism risk. 10 11 13 14 15 GABRB3 GABRA5 GABRG3 16 17 18 22 23 24 Signaling in the GABAergic system is mediated by receptors for the neurotransmitter GABA. There are 19 known GABA receptor subunits arranged in clusters throughout the genome. Functional pentamers formed by various combinations of these subunits result in receptors of varying properties and sensitivities. The amounts and functional capabilities of individual receptor subunits that form a specific pentamer can affect the amount and quality of signaling in different parts of the brain. p GABRA4 23 GABRB1 25 26 27 Materials and methods Samples N 28 29 30 31 32 33 34 35 23 Classification of history of seizure activity in autism patients was based on question 92 from the ADI-R, which queries for both current and lifetime presence of convulsions, seizures, and epilepsy. Caregiver responses to question 92 are coded to indicate no seizure activity, seizure activity with no definitive diagnosis of epilepsy, and seizures with a definite diagnosis of epilepsy. Using lifetime ratings, two groups of families were defined: those in which no seizure activity was reported, and those in which seizure activity was present in at least one autism patient. In addition, question 92 allows for coding of febrile seizures. Families with only febrile seizures were classified as negative for seizure activity and not included in the seizure subset analysis. Both families with positive and negative history of seizure activity were included in our overall dataset. This resulted in a dataset of 41 Caucasian families with a positive history for seizures. Molecular analyses and genotyping 23 GABRA1 GABRA6 GABRB2 GABRG2 GABRP GABRA2 GABRA4 GABRB1 GABRG1 GABRB3 GABRA5 GABRG3 GABRR1 GABRR2 GABRA4 GABRB1 N N GABRA4 GABRB1 1 http://genome.ucsc.edu http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=snp GABRB1 GABRA4 Fig. 1 GABRA4 GABRB1 Horizontal arrows Vertical arrows Pter cen  36 Statistical analysis 37 38 39 40 23 41 42 43 GABRA4 GABRB1 n= 44 GABRA4 Results GABRA4 p GABRB1 p 23 GABRA4 p GABRP p GABRA4 GABRB1 N N 1 p p GABRA4 p p p GABRA4 p p p p GABRB1 Table 1 GABRA4 GABRB1 GABRA4 GABRB1 SNP Caucasian African-American SNP Caucasian African-American a a a a PDT Geno PDT PDT Geno PDT RS7678338 0.9350 0.9923 0.2230 0.4190 RS1866989 0.3071 0.3860 0.5553 0.7892 RS6447517 0.8826 0.7034 0.505 0.3930 RS2351299 0.4529 0.0822 0.5775 0.8614 RS17599102 0.8055 0.8913 0.7518 0.8805 RS10016388 0.1585 0.2660 0.2367 0.4259 RS7660336 0.0833 0.0368(G/G) 0.5164 0.7410 RS1372496 0.2362 0.3740 0.2482 0.3715 RS1512136 0.9052 0.9869 0.2888 0.3575 RS3114084 0.0934 0.1942 0.2059 0.3281 RS17599165 0.0015(T) 0.0009(T/T) 0.6547 0.4304 RS4482737 0.1495 0.2504 1.0000 1.0000 HCV1592545 0.7798 0.9427 0.2230 0.5313 HCV11353524 0.1959 0.3117 1.0000 1.0000 RS7685553 1.0000 0.8419 0.4913 0.6044 RS3775534 0.1893 0.1831 0.4913 0.4913 RS1912960 0.0073(C) 0.0046(C/C) 0.4111 0.5110 HCV2119841 0.3352 0.0838 0.2278 0.4111 RS2055943 0.9671 0.9434 0.4927 0.7607 RS6287 0.4045 0.3571 0.6171 0.6984 RS2280073 0.1404 0.0955 0.0287(G) 0.1100 RS6289 0.9349 0.5554 0.8658 0.9220 RS16859788 0.3173 0.3173 0.0253(A) 0.0412(A/A) RS6290 0.4285 0.1973 0.3173 0.4594 RS17599416 0.0040(A) 0.0043(A/A) 0.8084 0.8084 4P0413 1.0000 1.0000 0.7389 0.7389 RS3792208 1.0000 0.4980 0.1797 0.1797 RS10028945 0.3272 0.4584 1.0000 0.8179 RS10517174 0.9484 0.0894 0.7150 0.8903 RS3832300 0.4094 0.6352 0.6547 0.6547 RS7694035 0.4337 0.6266 0.3657 0.3657      RS3792211 0.9057 0.8382 0.6547 0.2895      RS2229940 0.7873 0.9375 0.5485 0.7866      RS13151759 0.7529 0.9326 0.2367 0.5759      RS13151769 0.4740 0.7436 0.1824 0.4768      Bold indicates significant values a r 2 2 r 2 r 2 2 p p p 3 Table 2 Minor allele frequencies and linkage disequilibrium in Caucasian and African-American datasets  Bold indicates significant values a MAF Table 3 Haplotype frequency and association in Caucasian population RS7660336 RS17599165 RS1912960 RS17599416 Haplotype frequency P G T C A 0.503 0.0763 C T C A 0.277 0.9214 C T G A 0.124 0.6094 C A G G 0.082 0.0006 Bold indicates significant values GABRA4 GABRB1 GABRA4 GABRB1 p p p p p p r 2 p GABRA4 GABRB1 p p p p p p 4 Table 4 GABRA4 GABRB1 Input SNPs Significant interactions Gene SNP number SNP  SNPs P GABRA4 1 RS7660336 One-way 3 0.024 2 RS17599165 Two-way 3, 5 0.004 3 RS1912960  4, 5 0.014 4 RS17599416 Three-way 1, 3, 5 0.012 GABRB1 5 RS2351299  2, 3, 5 0.012 6 RS4482737  3, 4, 5 0.038 7 RS3832300  1, 4, 5 0.047 Discussion GABRA4 GABRA4 GABRB1 N GABRA4 GABRA4 GABRA4 GABRA4 GABRB1 GABRA4 GABRB1 GABRB1 GABRA4 GABRA4 r 2 2 GABRA4 GABRB1 GABRA4 4 GABRA4 GABRB1 GABRB1 GABRA4 Some differences in LD do exist between the two ethnic groups as well; however, the majority of the differences are small. The largest differences in LD are in pairwise values with rs16859788, which appear to mostly be due to the fact that the SNP is practically monoallelic in the Caucasian population. The Caucasian dataset suggests that there is a significant association of SNP haplotypes with autism risk, while the African-American set does not. However, this difference may be due to the small size of the African-American dataset. Therefore, while it appears that minor allele frequency differences can account for the lack of association of rs16859788 in the Caucasian dataset, additional studies are needed to determine whether the other differences in results between the two ethnic groups are due to sample size differences, differential LD with the causative variation in the two populations, or are population-specific risk alleles. GABRA4 GABRB1 GABRB1 GABRA4 GABRB1