Computation and Biology Group (CBG)

The Computation & Biology Group comprises members from the Department of Mathematics and EECS at MIT, and the Theory of Computation group at the MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL). The group focuses on various areas of research within the field of Computational Biology.

The Computation and Biology group, led by Professor Bonnie Berger, works on a number of problems at the interface of algorithms and biology. Many of the advances in modern biology revolve around recent advances in automated data collection and the subsequent large data sets drawn from them. We design algorithms to gain biological insights from this data and from future sources. We work on a diverse set of problems, including Protein Folding, Network Inference, Genomics, and Disease Classification. Additionally, we collaborate closely with biologists implementing these new techniques in order to design experiments to maximally leverage the power of computation for biological explorations.

Public-Domain Computing Resources -- Work on protein structure prediction has led to a number of programs for predicting protein folds from primary sequence data, discovering novel sequence-structure motifs, generating near-native decoys, side-chain packing, and threading. Work in mathematical models of virus shell assembly introduced a local-rules model for explaining the self assembly of viral shells. Work in genomics has focused primarily on annotating genes, regulatory motifs, and conserved secondary structures through interspecies comparison. Work in systems biology has resulted in a program that determines the optimal sampling strategy for time-series expression experiments. All of our programs are freely available for academic use.

The group has continued its efforts in protein structure prediction by updating our coiled-coil programs (i.e., Paircoil2, Multicoil2) and ß-structure prediction and modeling programs (BetaWrapPro), in collaboration with the Keating and King labs (MIT Biology) respectively. BetaWrapPro has recently been awarded a Best Structure Poster award at RECOMB 2005. Our work on beta-helices has provided good computational models for prions and amyloids, which are being tested by the Lindquist (Whitehead and MIT) and Gusella (Harvard Med) labs. We are currently applying our expertise in structure prediction and comparative genomics to functional analysis of protein active sites and binding sites. A recent collaboration with the Perrimon lab (Harvard Med) involves the use of RNA interference (RNAi) for the purpose of elucidating protein-protein interaction networks and signalling pathways. We are also exploring the use of structure to enhance the network-centric perspective of biological systems.