************************************************************************* Department of Mathematical Sciences The Johns Hopkins University SEMINAR ************************************************************************* Russell Schwartz April 5, 2001 Informatics Research Group 304 Whitehead Hall Celera Genomics Preseminar: 3:00 p.m. Refreshments: 3:30 p.m. Seminar: 4:00 p.m. ************************************************************************* HAPLOTYPE SEPARATION OF SHOTGUN ASSEMBLY FRAGMENTS ************************************************************************* ABSTRACT This talk will examine computational aspects of the problem of deducing genetic haplotypes from genome sequence fragment data. Fragment data is generally produced from an organism's DNA by "shredding" it into small pieces, which are individually sequenced and later reassembled as part of a process called shotgun assembly. In a higher organism, the fragments will typically be made from two slightly different copies of the genome due to the fact that the chromosomes that contain genetic material come in pairs. We are interested in locating from a pool of fragments subsets each of whose members come from a single chromosome. We call this problem haplotype separation. This talk will define the problem, present some formalizations, and examine the computational complexity of the problem. It will then describe a practical algorithm developed for a specific formalization of the problem and discuss its performance in practice. *************************************************************************