Dr Mayetri Gupta
- Reader in Statistics (Statistics)
R 231 Mathematics Building
15 University Gardens
My research primarily involves the development of novel statistical, in particular Bayesian, methodology for scientific problems arising in the fields of computational biology and genetics. Detection of sparse signals from noisy discrete data is a significant challenge in many fields, but especially so in genomic data analysis, due to latent positional or structural constraints in such data. I have been involved in developing novel Bayesian statistical approaches for detecting recurrent conserved patterns (motifs) in DNA sequence data, and prediction of chromatin structure (positioning of nucleosomes in DNA) through new adaptations of hidden Markov and hidden semi-Markov models. Motifs are often the sites of active gene regulation, and tend to be prone to damage from the environment, thus locating them accurately is an important challenge for biologists and clinicians alike. I also work on the development of Bayesian regression mixture, and hidden Markov regression models and associated Monte-Carlo based estimation procedures. These provide a robust and efficient way of deciphering regulatory networks of genes and associated motifs, combining different types of genomic data, including genomic sequence, gene expression microarray and tiling array data. I am also interested in general Bayesian methodology for clustering, classification, and model selection with high-dimensional, correlated data. In previous work with collaborators, we have developed a new information-based prior framework that allows for efficient inference in high-dimensional regression-type models. I am currently working on extensions of these ideas to be applied to the discovery of causal genetic mutations behind complex disease phenotypes.
Selected publications | View all publications
Mitra, R., and Gupta, M. (2011) A continuous-index Bayesian hidden Markov model for prediction of nucleosome positioning in genomic DNA. Biostatistics, 12(3), pp. 462-477. (doi:10.1093/biostatistics/kxq077)
Gelfond, J.A.L., Gupta, M., and Ibrahim, J.G. (2009) A Bayesian hidden Markov model for motif discovery through joint modeling of genomic sequence and ChIP-chip data. Biometrics, 65(4), pp. 1087-1095. (doi:10.1111/j.1541-0420.2008.01180.x)
Cheng, F., Hartmann, S., Gupta, M., Ibrahim, J.G., and Vision, T.J. (2009) A hierarchical model for incomplete alignments in phylogenetic inference. Bioinformatics, 25(5), pp. 592-598. (doi:10.1093/bioinformatics/btp015)
Gupta, M., and Ibrahim, J.G. (2009) An information matrix prior for Bayesian analysis in generalized linear models with high dimensional data. Statistica Sinica, 19(4), pp. 1641-1663.
All publications | View selected publications
Current PhD students
Tusharkanti Ghosh (Bayesian modeling approaches for the analysis of genome-wide association studies)