EM Algorithm For Haplotype Inference Incorporating Ungenotyped Individuals In Pedigree Studies

In modern biology genetics , organisms' haplotype information has an essential role , not only for its contribution in improving the power and nicety degree in genetics linkage and association studies, it's also used to detect genetics variants and complex-disease genes . With the rapid development of the modern science and technology , we have gained the genotypes data of organisms , but the haplotypes are unobservabled directly . There are many scholars commit theirselves to this work for several years , and have produced a series of methods , in which statistical method is the most primary one . However , these methods they all need an assumption , which is basic and important , that all of the genotype data sets are completely , and the pedigree data sets are not discussed in detail . But in real-world , the data we have gained are always enormous as large pedigrees , and within them there are many uncompletely or partly missing genotype individuals ,therefore those methods are all out of power .To contend with the limitation and weakness of the existed algorithms , we propose a new EM method to inference haplotypes more effectively and conveniently in large pedigrees which have ungenotyped individuals . At first , we make use of conventional EM algorithm to deal with the large pedigrees , which have completely genotype data , and gained a much detailed algorithm to inference haplotype , to deal with nuclear family data and pedigree data respectively . On this base , we pay more attention to discuss the pedigrees with ungenotyped individuals , using the kindred within these individuals , consider organisms' genetics relation in every nuclear family . For the genotyped organism in nuclear family , we apply the Judge operator to eliminate impossible haplotypes , and apply the Induce operator to reinforce some important one's information . In the present article , we describe an novel EM algorithm to estimate the haplotype frequencies and get some characters of the method . Then we make use of the estimation of the haplotype frequencies 9 to inference haplotypes of each individual .