HI everyone
I m working on Blast-2.2.28+ version to generate a PSSM using multiple sequence alignment file in PSI-BLAST.
I have used makeblastdb to generate my own database.
I successfully generated the PSSM using arguments (-out_pssm and -out_ascii_pssm).
I would like to know is there any way by which i can use this pssm and run against my own database to retrive the sequences which was used to make this pssm.Or how to check whether the pssm is correct or not?
To perform a search with PSI-BLAST using a PSSM as input (assuming NCBI BLAST+):
psiblast -in_pssm checkpoint.pssm -db database
Where 'checkpoint.pssm' is the checkpoint file you created with PSI-BLAST using '-out_pssm'.
The sequences used to create the PSSM were indicated in the PSI-BLAST run which generated the PSSM, they are also recorded in the checkpoint, however I am not aware of any easy way to access that information.
As for assessing the accuracy of the PSSM, the simplest way is to perform searches against various annotated databases (e.g. UniProtKB/SwissProt) and look for unrelated sequences. If the unrelated sequences are occurring above the inclusion threshold, it suggests that the PSSM is either contaminated (see PubMed:20064877 or has not been sufficiently refined.
