PhD STUDENTSHIPS, MRC BIOSTATISTICS UNIT, CAMBRIDGE

The BSU is an internationally recognised research unit specialising in
statistical modelling with application to medical, biological or public
health sciences. Details of the work carried out in the Unit appear on
the website www.mrc-bsu.cam.ac.uk. The Unit has at least 2 MRC PhD
studentships to start from October 2006. Awards cover Cambridge
University fees and a stipend for a period of 3 years. Awards are
subject to strict eligibility criteria. In addition, applicants must
have or expect to get a first or high 2.1 honours degree in mathematics,
statistics or a related discipline. A masters degree is highly desirable
but not essential.

Applications should be returned to Dr. Linda Sharples at the address
below or by e-mail by 28th February 2006. Interviews will take place in
the week beginning 13th March.

Examples of some of the projects available for PhD study are given
below. Other projects will also be considered.



Title: Analysis of genome wide association studies

Supervisor: Dr. Frank Dudbridge

With the completion of the human genome sequencing project, many large
scale projects are now underway to unravel the genetic basis of common
diseases. To analyse these studies effectively, new methods are required
to exploit the complexity of the underlying genetics while managing the
genomewide scale of the data. This project will explore

methods for finding associations from multiple hypothesis tests, for
integrating Bioinformatics knowledge with genetic epidemiology, and for
efficient analysis of multiple-cohort studies. The work will be
motivated by data from the European Bloodomics consortium, which is
identifying genetic risk factors in coronary artery disease.

Title: Benchmarking of longitudinal healthy active life expectancy methods

Supervisor: Dr. Fiona Matthews

Longer life expectancy is no longer the goal for most developed nations.
The aim is to live a healthy and longer life. Over recent years methods
have been developed for measuring healthy active life expectancy using
longitudinal data. To date these methods have been specific to the study
type that underpinned the research. There is now a consensus among HALE
researchers that these disparate methodologies need to be compared using
similar data sources with known properties. These discussions have been
led by Professor Carol Jagger who will co-supervise the project.
Professor Jagger co-editor of the book? Determining Health
Expectancies?, a 10 year compendium of the work of REVES (the
International Network on Health Expectancy and the Disability Process)
and responsible for the methodology section. The aim of this project
would be to collaborate with the worldwide group of HALE researchers to
compare each of the methodologies in turn. A detailed investigation of
each methodology together with the assumptions and limitations will be
undertaken. Further scope for methodological development of the current
techniqies will be evaluated. A toolkit for new researchers will be
developed and evaluated using data sources from the UK and the US.
Agreement is already in place with collaborators within the UK and
abroad to assist with data and programs for this project.

Title: Longitudinal studies of chronic diseases with applications to
psoriatic arthritis

Supervisor: Prof. Vern Farewell

This PhD project will make use of the unique data resource derived from
a 25 year study of over 600 patients with psoriatic arthritis. Building
on previous work, the aim will be to characterise patterns of disease
that may also inform treatment decisions. A particular focus will be on
the modelling of joint destruction at the individual joint level and its
relationship to inflammation patterns in the same and neighbouring
joints. Initial approaches will likely involve the use of random effects
models to reflect the complicated correlated outcome data of interest.
The appropriate definition and incorporation of explanatory variables at
both the joint and individual level will be required. There will also be
the possibility of incorporating genetic information into the model
building process.

Title: Studies of genetic association and integrative genomics

Supervisor: Dr. Carlo Berzuini

We are developing a general framework for studies of genetic association
based on unrelated cases and controls together with nuclear families
ascertained via an affected proband. The method we have in mind involves
a sequence of two analysis stages. In the first stage, all non-founder
individuals, which belong to a family without being among its

founders, are removed. The resulting - simplified - dataset is analyzed
via standard packages to generate a set of simulated (complete and
phase-solved) datasets with corresponding ``posterior weights". The
second stage of the analysis involves re-incorporating the removed
individuals, solving each family's chromosome transmission pattern
(allowing for genotyping error and recombination) and updating the
posterior weights to reflect the newly incorporated information. The
resulting (weighted) collection of imputed datasets is used for relative
risk estimation and or testing, for example via suitable Mantel-Haenszel
type tests, or via multiple weighted logistic regression. We aim at
enhancing and generalizing the above method, by incorporating new ideas
of haplotype clustering, as well as by incorporating knowledge from
bioinformatic databases. We shall apply the methods to our available DNA
database generated from an isolated population with high prevalence of
Multiple Sclerosis in Sardinia, Italy. This project will be jointly
supervised by Luisa Bernardinelli.

A second project concerns the development of statistical methods for
integrative genomics. We want to develop methods for a combined analysis
of data representing
different levels of organization of the genome, the proteome and the
metabolome. We want to apply these methods to data generated via
advanced experimental platforms within our European Project
collaboration. In particular, our available data will include
micro-array gene expression levels, genome-wide genotypes of individuals
affected by various clinical phenotypes (such as cardiovascular
disease), data from cell function assays, array CGH data, DNA
methylation data, small metabolite concentration profile data, and so
on. One idea will be to consider a list of "candidate" genes with
elevated functional annotation with respect to a specific clinical
phenotype, such as cardiovascular disease. Then we shall apply different
methods of multiple test correction (Bonferroni, false discovery rate,
(permutations) to find statistically significant associations of single
nucleotide polymorphisms (SNPs) along the genome with expression
variation in genes belonging to the above candidate list. More
specifically, we shall aim at identifying proximal (cis-) influences of
regulatory SNPs on the expression of particular genes involved in key
molecular pathways of the platelet cells. Another investigation will
involve application of Bayesian graphical modelling methods to combined
genotype and platelet functional assay data, the aim being to perform
simultaneous clustering of individuals into functional behaviour classes
and of genes or SNPs which are associated with these classes.

-- 
_________________________________________________________________________

Dr. Linda D. Sharples
MRC Biostatistics Unit and R&D Unit, Papworth NHS Foundation Trust
Cambridge, UK

Tel: 01223 330389 and 01480 364445
Fax: 01223 330388 and 01480 831450
[log in to unmask]
[log in to unmask]
__________________________________________________________________________