Biostatistics group at Imperial College

From Opasnet
Revision as of 07:06, 16 October 2007 by Jouni (talk | contribs) (copied from Word document for WP1.2)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Head of the group: Sylvia Richardson, Chair of Biostatistics.

The biostatistics group at Imperial College, headed by Sylvia Richardson, has a wide experience of applying statistical methods to environmental, biomedical and epidemiological data. Recently, the size, type and quality of such data have radically changed. Large data bases have been set up to monitor public health performance and to study geographical heterogeneity of disease risk at a small spatial scale. Moreover, there are many non-standard features of such data that need to be taken into account: missing or censored events, existence of intricate patterns of correlation and dependence, noisy, mis-measured and heterogeneous data with diverse sources of variability. All these features necessitate the development of new tools for integrating data efficiently from multiple and widely differing sources, new ways of modelling that rely on modularity, new approaches to the problem of multiple comparisons and false discovery and an increased interaction with the field of Bioinformatics.

The group mainly uses Bayesian methods that allow incorporating uncertainty at each level of the model. At the heart of these methods is the idea of hierarchical model building where a global picture of any complex data problem is constructed by: - using a number of local sub-models to capture different components of the problem, each with hidden variables, - organising these sub-models in a hierarchical fashion, - linking the local models via interpretable probabilistic relationships, with the ultimate aim of making probabilistic inference about all hidden variables. This strategy of model building is closely tied up with developments of efficient algorithms for estimation of the unknown quantities (e.g. exact probability propagation, maximisation in multidimensional spaces or stochastic (Monte Carlo) simulations).

One particular topic of interest has been measurement error models where Sylvia has an internationally recognized expertise. The methodology she proposed is described in a paper appeared on the American Journal of Epidemiology [1], and presents how the measurement error model can be modelled using conditional independence and how the uncertainty in the measures is taken into account allowing a distribution for the parameters. The approach has been applied to several case studies to assess risk of diseases based on different exposures (for instance the methodology was used to project the distribution of cancer risks from the Japanese atomic bomb survivors to the population of England and Wales [2] investigating flexible dose-response [3]). Other applications on risk assessment have been carried out by members of the Biostatistics group, for instance to investigate the effects of chlorination by products and other drinking water contaminants on the outcome of pregnancies or with the aim of creating a generic statistical framework for the optimal combination of complex spatial and temporal data from survey, census, real-time and telematics sources (part of the OPUS project: Optimising the use of partial information in urban and regional systems).

Sylvia has also led the group on ecological bias works [4], [5], [6], investigating the problems of using an ecological, i.e. group-level, approach from a Bayesian point of view and more recently focussing attention on how to synthesise grouped and individual level data in order to increase efficiency and reduce bias [7].

Moreover, Sylvia is an expert in statistical methodology, has written many papers on Bayesian methodology (see for example [8]-[12]), Markov Chain Monte Carlo methods (also editing a book on the topic [13]) and has applied these methodologies to different fields, (e.g. disease mapping, gene expression data).

For disease mapping one of the main projects the group is involved is The Small Area Health Statistics Unit (SAHSU), with the main aim of assessing the risk to the health of the population of exposure to environmental factors, with an emphasis on the use and interpretation of routine health statistics ([14],[15]).

For gene expression data two main projects have been carried on by the groups: 1) Bayesian Gene Expression (BGX), where the aim is to develop a series of flexible Bayesian models that can be used to obtain more information from microarray experiments; 2) Biological Atlas of Insulin Resistance (BAIR), where the aim is to build an atlas to understand the genes responsible for insulin resistance and diabetes. Output from both the projects have been published on international journals (see [16]-[22])

Selected bibliography

  1. S. Richardson, W.R. Gilks. A Bayesian approach to measurement error problems in epidemiology using conditional independence models. American Journal of Epidemiology, 138, 430-442 (1993).
  2. M.P.Little, I. Deltour, S. Richardson. Projection of cancer risks from the Japanese atomic bomb survivors to the England and Wales population taking into account uncertainty in risk parameters. Radiat Environ Biophys, 39, 241-252, (2000)
  3. J. Bennett; MP Little, S. Richardson. Flexible dose-response models for Japanese atomic bomb survivor data: Bayesian estimation and prediction of cancer risk. Radiat Environ Biophys, 43(4) :233-45, (2004).
  4. V. Lasserre, C. Guihenneuc-Jouyaux, S. Richardson. Biases in ecological studies: utility of including within-area distribution of confounders. Statistics in Medicine, 19, 45-59 (2000).
  5. S. Richardson and N. Best. (2003) Bayesian hierarchical models in ecological studies of health-environment effects. Environmetrics, 14: 129-147.
  6. N. Cressie, S. Richardson and I. Jaussent (2004). Ecological bias: use of maximum- entropy approximations. Australian and New Zealand Journal of Statistics, 46: 233-
  7. C. Jackson, N. Best and S. Richardson. Improving ecological inference using individual-level data. Statistics in Medicine, 25: 2136-2159, (2006)
  8. S. Richardson, P.J. Green. On Bayesian analysis of mixtures with an unknown number of components (with discussion). Journal of the Royal Statistical Society, Series B, 59, 731-792 (1997).
  9. P.J. Green, S. Richardson Modelling heterogeneity with and without the Dirichlet process. Scandinavian Journal of Statistics, 28, 355-375, (2001)
  10. P.J. Green and S. Richardson (2002). Hidden Markov models and disease mapping. Journal of the American Statistical Association, 97: 1055-1070.
  11. Viallefont V, Raftery A E and Richardson S. Variable selection and Bayesian model averaging in case-control studies. Statistics in Medicine, 20: 3215-3230, (2001)
  12. S. Richardson, L. Leblond , I. Jaussent and P.J. Green Mixture models in measurement error problems, with reference to epidemiological studies. Journal of the Royal Statistical Society, Series A, 165: 549-566, (2002).
  13. W.R. Gilks, S. Richardson and D.J. Spiegelhalter (eds): Markov Chain Monte Carlo in Practice, Chapman & Hall, 1996.
  14. S. Richardson, A. Thomson, N. Best and P. Elliott (2004). Interpreting posterior relative risk estimates in disease mapping studies. Environmental Health Perspectives, 112: 1016-25.
  15. P. Elliott; D. Briggs; S. Morris; Hoogh; C. Hurt; TK Jensen; I. Maitland; S. Richardson; Wakefield J; Jarup L. (2001) Risk of adverse birth outcomes in populations living near landfill sites. British Medical Journal, 323(7309), 363-8.
  16. AM. Hein, S Richardson, H Causton, G Ambler and PJ Green. BGX: a fully Bayesian gene expression index for Affymetrix GeneChip data. Biostatistics, 6(3): 349-373, (2005).
  17. AM. Hein and S Richardson A powerful methods for detection differentially expressed genes from GenChip arrays that does not require replicates. BMC Bioinformatics 7: 353, (2006)
  18. P. Broët and S. Richardson. Detection of gene copy number changes in CGH microarrays using a spatially correlated mixture model. Bioinformatics, 22: 911-918, (2006)
  19. A. Lewin, S Richardson, C Marshall, A. Glazier and T Aitman. Bayesian modelling of differential gene expression. Biometrics, 62:1-9, (2006)
  20. S. Parry, D. Hadaschik, C. Blancher, M.K. Kumaran, N. Bochkina, H.R. Morris, S. Richardson, T.J. Aitman, D. Gauguier, K. Siddle et al. Glycomics investigation into insulin action. Biochim Biophy Acta, 1760: 652-668, (2006)
  21. M. Blangiardo and S. Richardson. Statistical tools for synthesizing list of differentially expressed features from related experiments. Genome Biology, 8, R54, (2007).
  22. N. Bochkina and S. Richardson. Tail posterior probability for inference in pairwise and multiclass gene expression data. Biometrics, in press, (2007)