The Wheeler lab has been awarded the CBSD COBRE Junior Investigator project grant to develop “methods for fast bio-sequence comparison with profile hidden Markov models”. The grant will provide $450K in funding over three years, starting … now!
Sequence database search is fundamental to modern molecular biology –it allows one sequence to be annotated based on detected similarity to other known sequences. Annotation of metagenomics and metatranscriptomics datasets is incomplete (many observed sequences can’t be identified) and also rate limited by computational analysis. The funded research addresses these conflicting needs for maximal sensitivity and high speed, through development of new algorithmic approaches designed to dramatically accelerate annotation methods based on highly sensitive probabilistic models called profile hidden Markov models (HMMs).
The resulting gains will enable broader application of profile HMMs to large-scale analysis projects (e.g. in metagenomics pipelines), improving the biological insights gleaned from these data sets. The Wheeler lab will also demonstrate the power of new heterogeneous FPGA/CPU environments.