Here we have two peptide sequences, one of them is not fully annotated, and the function is unknown.
We want to know if there is interaction between peptides in silicon. Is there any tools?
Thanks!
Hi, you might want to first predict the function of the peptide, there are few sequence/structure homology based tools.
PHYRE2 http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index
Hi. One standard method is to use some form of machine learning to predict binding affinity. Given a group binding affinities (from protein-protein databases), you would use supervised learning with a classifier such as a Random Forest or SVM to learn which sequences bind. The proteins could be encoded with AAindex matrices or you could use could devise a more simple feature vector based upon the frequency of AA in your peptide sequences. This is very much related to epitope binding problems (see http://www.ncbi.nlm.nih.gov/pubmed/15542369) for early methods based on neural networks.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics