A survey of current Bayesian gene mapping method

John Molitor,Paul Marjoram,David Conti,Duncan Thomas

doi:10.1186/1479-7364-1-5-371

John Molitor, Paul Marjoram + Show 2 more

Open Access

https://doi.org/10.1186/1479-7364-1-5-371

Copy DOI

Journal: Human Genomics	Publication Date: Aug 1, 2004
Citations: 8	License type: cc-by

Affiliation: University of Southern California

Abstract

Recently, there has been much interest in the use of Bayesian statistical methods for performing genetic analyses. Many of the computational difficulties previously associated with Bayesian analysis, such as multidimensional integration, can now be easily overcome using modern high-speed computers and Markov chain Monte Carlo (MCMC) methods. Much of this new technology has been used to perform gene mapping, especially through the use of multi-locus linkage disequilibrium techniques. This review attempts to summarise some of the currently available methods and the software available to implement these methods.

Highlights

Bayesian methods have become extremely popular in genetic analysis, in part because they allow for the incorporation of background information into the model
While Markov chain Monte Carlo (MCMC) methods are, by themselves, not Bayesian methods, they are most often utilised in a Bayesian context, as the random nature of parameters in a Bayesian model allow for MCMC methods to be utilised in a natural way
The authors have summarised some of the more popular currently available methods in this area. (For a more technical review of Bayesian haplotype association methods, see Thomas et al.25) One area that has not been discussed here is the issue of phase estimation

Summary

Introduction

Bayesian methods have become extremely popular in genetic analysis, in part because they allow for the incorporation of background information into the model. Many excellent introductions to MCMC methods exist.[2] Much of this powerful Bayesian-based computational machinery has been applied to the field of gene mapping. LD refers to a non-random association of alleles within haplotypes It is these associations that are used in gene mapping techniques.[9] Bayesian methods utilise LD through the use of likelihoods that exploit these allelic associations. The first approach is to examine the association of continuous sets of markers (ie haplotypes) with disease (see below). The location of a putative disease-causing mutation is used as a point of reference for haplotype risk estimation Another approach is to examine the association between alleles and disease status but to model dependency q HENRY STEWART PUBLICATIONS 1473 – 9542. A third, arguably more ambitious, approach is to approximate ancestral trees without modelling the entire coalescent (see below)

Haplotype methods

Clustering methods

Conclusion