Due to a late withdrawal the following BBSRC funded PhD studentship has become 
available commencing October 2000.  The funding covers fees plus an annual 
living allowance which is currently £7,380 pa.

To be eligible candidates should be UK residents with a good first degree in 
statistics, mathematics or computer science, or a biological science with a 
strong quantitative content.  Knowledge of the underlying biology or computing 
is not essential, but enthusiasm to learn about these areas is required. 
Residents of other EU countries are also eligible but in such cases the 
studentship pays tuition fees but does not provide an annual living  
allowance.

Prospective candidates should contact Dr Mike Denham as soon as possible.


Dr Mike Denham
Department of Applied Statistics
The University of Reading
PO Box 240
Reading RG6 6FN

Tel: 0118 9318914
Fax: 0118 9753169

Email: [log in to unmask]


ABOUT THE PROJECT:


Methodology for QTL analysis using very dense marker maps 

Supervisors: Dr Mike Denham & Dr John Whittaker 

The problem of mapping the genes controlling quantitative traits (QTL), and then
exploiting these loci in breeding programmes via marker assisted selection (MAS)
or introgression has received much recent attention.  However, existing
methodology for inbred line crosses is in general only suitable for relatively
sparse marker maps.  Current methodology is based around linear models, where
the phenotype is the response variable and markers are explanatory variables.
Some form of selection procedure, such as the Akaike Information Criterion (AIC)
is used to select a set of important markers which will be used to control for
background genetic variation whilst the genome is scanned for major genes.
However, as marker density increases so does the correlation between nearby
markers and this approach becomes increasingly unworkable.  Current exciting
developments in molecular genetics, such as the availability of high density SNP
maps, will thus require new statistical methodology to deal with the very large
number of highly correlated variables produced.  Fortunately, similar
statistical problems are encountered in other areas, notably chemometrics.  Here
regularisation methods such as ridge, principal components and partial least
squares (PLS) and new variable selection methods have been used to fit linear
models relating spectral absorbance data consisting of the order of 1000
explanatory variables to chemical composition using as few as 12 observations.
We will adapt these methods for use in QTL mapping and MAS.  Initially, we will
concentrate on QTL mapping in F2 and backcross populations with univariate
Gaussian responses, before moving on to consider MAS and introgression.  More
general responses and population structures will be considered if time permits.
The methods developed will be tested on simulated data throughout the course of
the project.

ABOUT THE DEPARTMENT:

The Department of Applied Statistics has a long record of excellence in teaching 
and research connected with the application of statistics to the life sciences.  
The Department's highly-regarded MSc in Biometry is supported by the MRC, the 
BBSRC, the EPSRC and the pharmaceutical industry.  MSc courses form part of PhD 
students programmes where appropriate.

The successful applicant will form part of the Department's Statistical Genetics 
group which currently consists of three teaching staff, one emeritus professor, 
four postdoctoral researchers, and three PhD students.



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%