Matthew Stephens FRS
My lab works on a wide variety of problems at the interface of Statistics and Genetics. We often tackle problems where novel statistical methods are required, or can learn something new compared with existing approaches. Thus, much of our research involves developing new statistical methodology, many of which have a non-trivial computational component. And because data sets are getting larger and larger our work often involves modern methods for “high-dimensional statistics”. Our work often makes extensive use of Bayesian hierarchical models to borrow information across data sets or sampling units.
My lab has also been increasingly focused on making its research more open, reproducible and extensible. This is because I see this as the first step towards greater cooperation of scientists to achieve common goals.
See http://github.com/stephenslab/ash for an example of a project I conducted “in the open”. And see https://jdblischak.com/workflowr.html for an R package we developed to help students and others make research websites of their analyses. In learning to do research this way, my lab uses git for version control, github for sharing code, and Rmarkdown to help make our R analyses clear and share-able.
Current research interests include:
- Sparsity, shrinkage, and false discovery rates, particularly for complex inter-related datasets.
- Factor Analysis, dimension reduction, and estimation of large covariance matrices.
- Clustering methods, and generalizations (eg grade of membership)
- Applications of multi-scale and wavelet methods to genomic data
- Reproducible research and open science