The emerging paradigm for extensive genetic studies relies on a haplotype map to guide study design and analysis.
The research aims to tackle the following computational issues that are essential for such studies: 1. Development of methodology for selection and utilization of tag SNPs to be typed in an association study. Specifically, the statistical power to detect association is suggested as a figure of merit for candidate tag SNP sets, and a goal for optimizing such a set. Furthermore Bayesian framework is developed to generalize and synthesize current approaches to the identity, prioritization, and evaluation of the tagged variants. 2. Development of refined models for haplotype ancestry. A model of two HMM levels is proposed. The """"""""bottom"""""""" model describes the descent of modern samples from ancestral chromosome segments that underlie common haplotypes; the top model portrays genealogies leading to those ancestral segments. This block-free framework has potential to help identify the causative allele in association studies. 3. Participation in gene discovery studies of Type II Diabetes and Inflamatory Bowel Disease, ongoing in our lab. These data-driven projects will focus method development efforts to the practical biological applications.
Singh, Angad Pal; Zafer, Samreen; Pe'er, Itsik (2013) MetaSeq: privacy preserving meta-analysis of sequencing-based association studies. Pac Symp Biocomput :356-67 |
Lowe, Jennifer K; Maller, Julian B; Pe'er, Itsik et al. (2009) Genome-wide association studies in an isolated founder population from the Pacific Island of Kosrae. PLoS Genet 5:e1000365 |
Barzuza, Tamar; Beckmann, Jacques S; Shamir, Ron et al. (2008) Computational problems in perfect phylogeny haplotyping: typing without calling the allele. IEEE/ACM Trans Comput Biol Bioinform 5:101-9 |