Figure 2

A 3-SNP block is more complicated. A. Cartoon of a 3-SNP block from CEU_TRIO member NA12878 on chromosome 9 overlapping exon 8 of gene CBWD6. The format is similar to Figure 1. B. Sequencing reads, as displayed on the Integrative Genomics Viewer [9] dramatically illustrate the fact that the “adjacent” SNPs in a block are not contiguous in the genome. Note the blanks in several read based haplotypes. Our parsimonious clustering procedure combines read-based haplotypes CG_ and _GA into CGA. Other possible completions of the CG_ partial haplotype (like CGT or CGC) are mathematically possible but excluded because they are not seen in the existing reads. It is clear that at least three haplotypes are required to support this data: CGG, CGA and TTA.