HapTree-X: An Integrative Bayesian Framework for Haplotype Reconstruction from Transcriptome and Genome Sequencing Data

摘要

By running standard genotype calling tools, it is possible to accurately identify the number of "wild type" and "mutant" alleles for each single-nucleotide polymorphism (SNP) site. However, in the case of two heterozygous SNP sites, genotype calling tools cannot determine whether "mutant" alleles from different SNP loci are on the same chromosome or on different homologous chromosomes (i.e. compound heterozygote). In many cases, the latter can cause loss of function while the former is healthy; therefore, it is necessary to identify the phase (or diplotype) - the copies of a chromosome on which the mutant alleles occur - in addition to the genotype. Identifying phase information for an individual is important in biomedical studies due to disease association of complex haplotype effects such as compound heterozygosity, as well as matching donor and host in organ transplantation. As more sequencing data becomes available, we seek to design efficient algorithms to obtain accurate and comprehensive phase information directly from transcriptomic, as well as the commonly-used genomic, NGS read data. Transcriptome sequencing data differs from genomic read data in that genes often have differential haplotypic expression [3] (expression bias between the maternal and paternal chromosomes of a particular gene). We are able to leverage this asymmetry to increase the number of SNPs of an individual that can be phased.