Genome-wide linkage disequilibrium mapping of QTL using data mining and evolutionary algorithm techniques

Dr P. Visscher 

(with Dr. S. Knott)

!!! PhD Studentship Available !!!: BBSRC CASE Award for 2001-2004

Background: With the availability of (SNP based) dense marker maps it will become possible to search for loci underlying complex disease traits and quantitative traits (i.e., search for quantitative trait loci or QTL) using whole genome scans. Such scans will provide masses of data. For example, it will be possible to have phenotypic records on 1000 or more individuals combined with, say, 50,000 SNP genotypes on each of these individuals. Various statistical tests have been proposed to test for the effects of candidate genes (or polymorphisms within candidate genes) on complex or quantitative traits. However, none of the methods consider (i) multiple markers simultaneously in quantifying the amount of observed linkage disequilibrium, and (ii) the statistical challenge of dealing with many candidate regions (or markers) which could display epistatic interactions combined with a finite set of individuals with observations (phenotypes). From a statistical point of view this is an optimisation (or model selection) problem. We suggest that a student tackles both these problems with appropriate statistical techniques.

Project: Using analytical methods and Monte Carlo simulations and, and possibly analysis of industry partner provided data, the student will (1) develop and test simple and robust methods to quantify multi-point linkage disequilibrium from population-wide samples, including the exploration of data mining algorithms, and (2) develop and test evolutionary algorithms (such as the genetic algorithm) to detect multiple interacting QTL in genome wide association studies. Parameters that will vary in the simulation studies are the number of QTL in the genome, their effects and interactions, the population evolutionary history, marker density, and the number of phenotyped individuals. This projects is suited to candidates with a background in the biosciences or maths/stats, with a proven aptitude for numerical skills.

Detection of quantitative trait loci in animal and human populations

Most traits of importance in human medicine and livestock genetics (for example, common diseases in man and milk yield in dairy cattle) are quantitative, i.e. influenced by environmental factors and multiple genes, and genomics projects worldwide are moving from the detection of major genes to tackling these complex quantitative traits. The genome revolution is opening exciting opportunities to study and utilise quantitative traits in livestock and man, and quantitative genomics will be essential and central in future studies. A number of projects are available to develop and apply methods to find loci underlying complex traits:

(i) Detection of quantitative trait loci for psychiatric disorders (in collaboration with the Department of Psychiatry of the University of Edinburgh)

(ii) Detection of quantitative trait loci in pig populations (in collaboration with the Roslin Institute)

(iii) Combining linkage and association analyses to detect quantitative trait loci (in collaboration with the Roslin Institute)

These projects are suitable for students with a background in quantitative genetics or statistics, or a background in animal science or biology with a demonstrated aptitude for numerical skills.

References

Lebreton, C.M, Visscher, P.M., Haley C.S., Semikhodskii, A., and Quarrie S.A. (1998). Nonparametric Bootstrap Method for Testing Close Linkage vs. Pleiotropy of Coincident Quantitative Trait Loci. Genetics 150: 931-943.

Visscher, P.M., Haley, C.S., Heath S.C., Muir W.J., and Blackwood, D.H.R. (1999) Detecting QTLs for uni and bipolar disorder using a variance component method. Psychiatric Genetics 9: 75-84.

Walling, G.A., Archibald, A.L., Cattermole, J.A., Downing, A.C., Finlayson, H.A., Nicholson, D., Visscher, P.M., and Haley, C.S. (1998). Mapping of quantitative trait loci on porcine chromosome 4. Animal Genetics 29: 415-424.

Quantitative genetic analysis of age at onset

The aim of this project is to develop and apply novel statistical methods to genetically analyse Age-at-Onset (AAO) data in animal and human populations. The purpose of such analyses is to quantify the amount of genetic variation that exists in populations for AAO, and to locate genome regions which contain genes affecting the AAO of a disease or other trait. The developed methods will be applied to data on age at puberty (pigs), ascites (chickens), and age-of-onset for psychiatric disorders in human populations. The hypothesis to be tested is that the new methods are more powerful in dissecting genetic variation underlying AAO traits.

Age at onset of a disease is a complex trait for genetic analyses because (i) it is influenced by both the environment and by multiple genes and (ii) it is measured on affected individuals only. There have been several quantitative genetic analyses proposed. The most common one is to ignore age data, and to concentrate on affected versus unaffected individuals. A second method is to adjust the affected/unaffected scores for age A third method is to sample the AAO conditional on the actual age and inferred genotype. All these methods have drawbacks, because they do not use the information efficiently by ignoring the nature of the censored records.

We propose to develop and apply survival analysis methodology to model AAO directly, and to estimate polygenic and QTL variance components for this trait, using pedigree information and genetic marker data. Survival analysis models the trait of interest (here, AAO) properly as a function of time, and takes into account that some records at uncensored (affected individuals) and some are censored (unaffected individuals). The effects of factors such as sex, contemporary group and genotype are estimated in terms of the risk of becoming affected.

The project will be supervised jointly by the University of Edinburgh and the Roslin Institute. The project would be suitable for a student with a background in quantitative genetics or statistics, or a background in animal science or biology with a demonstrated aptitude for numerical skills.

References

Daw et al., 1999, Am J Hum Genet 64: 839-851.

Ducrocq & Casella, 1996, Genet Sel Evol 28:505-529.

Hanson et al., 1998, Am J Hum Genet 63: 1130-1138.

Back to postgrad opportunities
 
 

 

Site design by
Comments to icapb.webperson@ed.ac.uk.
Last modified: November 6, 2000 by David Baird
(You are visitor 542 since November 13, 2000)