I used a PSIBLAST (position specific scoring) to detect the similarities between specific proteins and the database of Clostridium.

The results did not reveal any similarities above the e value threshold of 0.001.

Could I continue iterations of the second and third round of PSIBLAST?


The first round of psiblast is the same as blastp and an evalue of 0.001 is low and below the recommended threshold of (e)-value = 0.005.

You require significant knowledge of composition bias and globular domains to move into the 2nd and 3rd iterations. I'd recommend reading NCBI's manual here.

From you response you are using psiblast as a tool to look at antibody cross-reactivity, which I presume is pan-Clostridium.

I recommend using psi-blast as investigative analysis only and switching to the Immune Epitope Database here. Everything will be explained on the IEDB server. On this server I suggest performing two separate analyses and seeing whether the results correlate,

  • B-cell epitope prediction on your strongest evalue
  • B-cell epitope prediction on your query sequence
  • Do the epitopes positions match?

If the epitope positions do not match then the hypothesis of a pan-Clostridium cross-reactive antibody can be immediately discarded. If there is a match it provides the basis for further investigation.

Please keep in mind Clostridium will use glycosylation to modulate antibody binding.

  • 2
    $\begingroup$ 0.005 isn't the "recommended threshold", the NCBI manual you linked to only mentions it as a "rule of thumb" for beginners. The threshold you choose will always depend on the specific question you are trying to answer, the distance between your query and the target database, the size of the target database, the specific features of your query etc. There's no single, global recommended threshold for an e-value. And why would you need significant knowledge to go into the 2nd iteration? The results from the first one are basically a normal blast, you need iterations for PSI-BLAST to be useful. $\endgroup$
    – terdon
    Aug 16 '19 at 8:35
  • $\begingroup$ Clostridium difficile is a major research front and has a large amount of associated genomic data. I would say given that 0.001 is low for an e-value in context. The evalue being a consequence of the database size. However, we don't know the query sequence or why it should be associated with C difficile. Ultimately we don't really know what the purpose of the psiblast is ... in this sense you might be right. $\endgroup$
    – M__
    Aug 16 '19 at 8:37
  • $\begingroup$ I am new to PSIBLAST and I need more resources to understand it. As I understand from your answers, I can continue the iterations with these sequences. $\endgroup$
    – S.Khalil
    Aug 16 '19 at 9:52
  • 1
    $\begingroup$ What is your query sequence? Are you querying putative commensal Clostridium against the (essentially) C. difficile database? What is your objective? Without that information it is impossible to say in my opinion $\endgroup$
    – M__
    Aug 16 '19 at 9:56
  • 1
    $\begingroup$ Wow, that really helped. Thank you. $\endgroup$
    – S.Khalil
    Aug 18 '19 at 6:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.