Haplotype phase inference

Most of the information being collected on DNA variation among people does not identify which of the two parents transmitted which of the two copies of each gene. Even worse, the parent of origin is often scrambled for each single nucleotide polymorphism (SNP) that is identified, so that each gene may be represented by hundreds of pairs of SNP vectors or haplotypes. Collection of a population sample of this kind of genotype data, however, does contain information about these haplotypes, and inference of the haplotype pairs from this kind of data is referred to has haplotype phase inference. The problem has a rich geometric representation, and has spawned a wealth of algorithms that span graph theoretic to Bayesian approaches. Good solutions to this problem are strongly motivated by the efforts seeking to identify genes that underlie complex genetic disorders. The latest efforts in this area will be described and reviewed.