Syedur Rahman

Imperfect Phylogeny Haplotype Inference: Using Regions of Perfect Phylogeny Syedur Rahman

The haplotype inference (HI) problem is defined as the problem of inferring 2n haplotype pairs from n observed genotype vectors. The inference of haplotype information from genotype data (the latter of which is more readily available) is very useful in researching genes affecting health, disease and responses to drugs and environmental factors. The PPH or the Perfect Phylogeny Haplotype model assumes that inferred haplotypes from a sample can be derived using a single tree, i.e. a perfect phylogeny. However, there are biological events such as recombination that violate this model. Stochastic methods on the other hand can infer haplotypes despite recombination but they can be time consuming and their inferences often depend on the initial state randomly chosen during a run. The research described in this monograph aimed to analyse previous models and solutions to the haplotype inference problem and engineer algorithms that would infer haplotypes from genotypes in the presence of recombination using disjoint and overlapping regions of perfect phylogeny and scale better in terms of time complexity.