CRUMp is based on a kernel-based learning method called Classification Relevance Units Machine (CRUM). Given an input set of protein sequences in FASTA format, the system outputs the position, residue type (S, T, or Y), and the estimated probability of each tested site being phosphorylatable.
Latest downloadable files:
- crump-0.2.0.tar.gz: CRUMp GNU Octave package
- crump-0.2.0.zip: CRUMp MATLAB script
- crumptestset.fasta: A testing dataset in FASTA format. The sequence headers list the accession number of the protein sequence and the position numbers of known phosphorylation sites. Note that CRUMp may predict additional phosphorylation sites that have not been experimentally verified yet.
The testing dataset is from Biswas et al. 2010, http://www.biomedcentral.com/1471-2105/11/273/additional.
Be the first to post a review of CRUMp!