Haplotype-phasing in polyploids

Many agricultural crops like cotton, potato, alfalfa and many of the flowering plantsare polyploid. In order to manipulate these crops genetically so that they have superior traits like higher nutritive value, improved tolerance abilities against harsh climates etc., it is important to characterize the genetic markers associated with these traits. It is required to identify their genotypes and commonly found haplotypes. There are some reliable methods to perform genotyping and phasing of diploid species, but not many such methods are available for polyploids. When we tested different programs to perform genotyping of tetraploids, we learnt that the methods which are available are very specific to the ploidy and that the various available methods are not in agreement with one another for a given set of data. My PhD project aims to develop and implement a phasing method, which may be applied for species of any ploidy. Our phasing method, which is very close to completion is based on observation that haplotypes tend to form clusters and the similarity of the clusters varies along the genome according to a recombination rate. This idea is motivated by the model behind the method fastPHASE, a tool to phase diploids, detailed in Scheet & Stephans (2006). We use Hidden markov model (HMM) approach based model and then use Numerical optimization to estimate the parameters. In the end, we use backward sampling to perform the phasing, which is a standard method with HMM based approaches. In order to address the performance load which is typically imposed by HMM methods with large state spaces, we employ the particle filter method to sample and filter the states and this makes the size of our state space easily manageable.The  implementation with respect to coding (all the modules ) is complete, but it is yet to be tested rigorously on various kinds of data sets and then we intend to perform comparative analysis against current available methods of phasing.