I have a set of n aligned protein sequences. I have constructed a profile hidden Markov model (HMM) from the alignment with hmmbuild in the HMMER package.
I would like to compare my constructed HMM with a database of standard HMMs (e.g. those in the PFAM or ESTHER database) to see which standard HMM is most similar to mine.
Is there a way I can do this (web server or command line)?
You can use your profile for search with hmmsearch at EBI HMMER service, and can choose among several databases. I don't know specifically about ESTHER db.
But most standardised analysis would be to download the database and conduct the analysis locally with the same hmmer package you are using.
Abhijeet Singh I think hmmsearch allows me to search my HMM against the sequences in the databases and it returns individual sequences as hits.
What I am interested in doing is searching my HMM against the HMMs of the protein families in the databases, rather than against individual sequences. I would like to determine which HMM of the database protein families is most similar to my HMM.
Or maybe you can get/download the db HMM and subset the profile of your interest and then format to sequence database on which you can try your HMM.
Just some thought. Assuming that you can translate your HMM profiles into position weight matrixes (PWMs), you can use motif similarity search tools like Tomtom in MEME suite to search PWM by your HMM against PWMs of HMM databases.
- Badges
- Science method