Research

My lab works on a variety of projects in the general area of computational biology. Two particular areas of interest are computational genetics, particularly with regard to genetic variation analysis and phylogenetics, and modeling and simulation methods, especially with respect to molecular self-assembly systems. We are also regularly involved in various collaborative projects, bringing our expertise in modeling and algorithms for biology to applications drawn from many biological domains.

Our largest current project in computational genetics consists of studies of cancer heterogeneity and progression. Our group has worked for a number of years on evolution of tumor cell populations, with particular interest in phylogenetic algorithms for reconstructing evolutionary histories of tumor cell lineages. One particular contribution of our group to the field was the introduction of methods for reconstructing tumor phylogenies from cell-to-cell variation data, primarily using fluorescence in situ hybridization (FISH) data and now extending to various other data types. We continue to work on improving models and algorithms for these studies and extending them to other single-cell data types. Another contribution of ours was the development of mixture modeling methods as a way to resolve clonal populations and their evolution from bulk tumor data. We continue to work on improving these methods through a focus on more advanced algorithms for reconstructing complex mixtures from high-dimensional data sets. These projects in turn were built on a long-standing interest in algorithms for genetic variation analysis, population genetics, and phylogenetics more generally, topics that remain interests of the lab.

Our largest project in modeling and simulation has long been study of viral assembly as a model system for complex self-assembly more generally. We work on a combination of problems ranging from theoretical contributions to fundamental algorithms and data structures for fast stochastic sampling to application of such methods to ask basic questions about the nature of pathway spaces of self-assembly systems and other similarly complex reaction networks. Our major interest in recent years have been attempting to bridge the gap between simulation and experiment in this field by developing methods for model-fitting, to learn coarse-grained rule-based models of viral capsid assembly to fit bulk kinetic data produced by monitoring in vitro assembly systems. We are currently working to improve such methods through a combination of more advanced algorithms for learning models through simulation-based data fitting and extending such models to handle newer experimental data types. A further interest is the application of coarse-grained biophysical models to these systems to yield a more accurate picture of how complex assembly processes might work in the cell compared to the test-tube in vitro models by which they are usually studied.