Figure 1

Method. (A) Investigated sites with exon/intron boundaries. We searched dbSNP build 127 (green box) for SNPs on splice dinucleotides (+1, +2, -1, -2) at both edges of whole introns of the human genes defined by H-InvDB (orange box). We named these SNPs "sdSNPs", and examined whether their alleles satisfied the GT-AG rule (red box), which requires that the 5' and 3' ends of introns should be GT and AG sequences, respectively. Blue boxes and dotted lines are exonic and intronic regions. (B) Allele estimation method for SNPs at splice dinucleotides. To estimate the SNP (SNP2, red arrow) allele (red question mark) of each transcript sequence that had been spliced out, we used different SNPs (SNP1, blue arrow) located in the nearest exon of the same transcript. Assuming linkage disequilibrium between a SNP at a splice dinucleotide (e.g., SNP2, whose alleles are G and C) and that in an exon (SNP1 whose alleles are A and T), we estimated alleles of SNP2 in transcript 1 (brown) and transcript 2 (pink) using a BAC clone sequence (gray bar). In this case, transcript 1 was estimated as the same haplotype as the BAC sequence because transcript 1 has A at the SNP1 as in BAC sequence. See text (Methods) for detail. (C) Example in which a SNP gives another exon boundary This example shows that the SNP (bold in box) whose alleles are A and G make two possible exon/intron boundaries consistent with the GT-AG rule (Boundary-1 Allele-G and Boundary-2 Allele-A). When the G allele is given, the exon/intron boundary is the next base of the SNP. When A alleles are given, the boundary is shifted one base downstream. The lower-case and capital letters indicate intronic and exonic regions, respectively. The red "ag" indicates a match to 3' splice site sequences.