Speaker: Jin Ju
 
Day:  Wednesday, 11/10/2004
Room: ITE 336
Time:  3:30pm-4:30pm 

Title: Bayesian Methods for Haplotype Inference
----------------------------------------------------------------------------

Abstract: 
In this talk, application of Bayesian methods to haplotype inference 
will be shown.
Haplotype inference is the problem of inferring original haplotype pairs from given 
genotypes, one of which can be explained by many possible haplotype pairs.
Due to the hidden configurations (possible haplotype pairs per each genotype) and the exponentially 
growing size of these, traditional likelihood approach, e.g. MLE, is not efficient. 
To resolve this obstacle, an iterative optimization technique, Gibbs
Sampling, is applied to this problems. Moreover, two algorithms are be 
showing with two different prior models.

back to CSE-TS page