Projects
HHsuite
HHsuite is an open-source software package widely used in biochemistry and bioinformatics. It contains the most sensitive and fastest applications (HHsearch, HHblits) for the frequent task to search for homologous proteins. The method is based on the pairwise alignment of hidden Markov models (HMMs).
- Moved the project from a local SVN to GitHub
- Modernizing legacy (spaghetti) C++ code
- Refactoring the build pipeline with cmake
- Parallelization for HPC (Open MPI and OpenMP)
- Support scripts for a new protein structure format (mmcif)
PEnG
Fast discovery of statistically overrepresented motifs in a set of DNA sequences. Those sequence sets can be collected from ChIP-Seq experiments, so one protein binds to all these sequences. Therefore, all these sequences should have a binding motif for this protein. However, it is not known where the motif is, what length it has and if there are multiple motifs. In this simple but fast approach we test all possible k-mers, calculate their Pvalues and optimize the k-mers first to the degenerate IUPAC alphabet and later to position specific weight matrices (PWMs). It is supposed to be a fast prefilter for BaMMmotif.
- Based on Bayesian statistics
- Written in C++, Python, R