I have 5 strains of P.falciparum. Each FASTA file has all its annotated CDSs. After a first pre-processing phase, where I eliminated the strangest sequences (perhaps the longest or shorter ones, which did not start with methionine or which did not have a length multiple of three) I wanted to align the CDS of these strains to evaluate how much every single gene was prone to acquire structural polymorphisms (insertions, SNPs deletions, etc.). A first step I thought could be BLAST to find the orthologs and associate the identifiers of the genes. The problem is that BLASTn has a limit on the number of sequences it can use. How can I go about getting multiple alignments? Also, is there any useful statistic that you can apply that you know of in the literature that can provide a mathematical criterion for obtaining a statistic about it? thanks a lot

  • $\begingroup$ 1) why are you using blastn and not tblastx, for CDS? 2) are you using the blast web portal or a binary on HPC or laptop? I would be surprised if a binary had this limit. 3) have you considered other methods, such as using HMMER to annotate ORFs with Pfam and then matching Pfams to find orthologs? or translating to protein sequences, which would be easier? 4) Have you considered googling for multiple sequence alignment packages? $\endgroup$ Jan 24, 2021 at 21:56
  • $\begingroup$ I Split the comment in two part :I was using the portal. It is the first time that I have to use BLAST massively so I am quite bewildered. Since the web portal has limits in the bases it can use I downloaded the standalone BLAST and was studying how to use it. As for the packages you mention: are you referring to R? Since I have to align 5 proteomes with more than 5000 sequences to find the orthologues among these isolates, if I used R would the computation times be acceptable? $\endgroup$ Jan 25, 2021 at 10:56
  • $\begingroup$ Otherwise what are you referring to? No Pfam I had not considered for now. Aside from the things mentioned above I am adding it to the list of tools to use to get around the problem. Can you link me some reports from which to draw information in such a way as to make it faster to use? thanks a lot $\endgroup$ Jan 25, 2021 at 10:56
  • 1
    $\begingroup$ 1) Standalone is likely to work. You will have to use makeblastdb to make a database per proteome. 2) No, these are not in R. tblastx (or blastp if you translate) is a part of the blast suite (nebc.nox.ac.uk/bioinformatics/docs/tblastx.html). I suppose that directly searching with BLASTP or whatever will yield best hits, but annotating with Pfam is still something that you will likely have to do anyways. Using hmmsearch (in HMMER) to search against Pfam is probably simplest and fastest, if you choose that route (eddylab.org/software/hmmer3/3.1b2/Userguide.pdf). $\endgroup$ Jan 25, 2021 at 20:40
  • $\begingroup$ STALN is what I'm currently trying to use now. Indeed, I am creating a DB per proteome. Once I will have studied the material I will post my doubts. Thank you in advance! $\endgroup$ Jan 26, 2021 at 9:19


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.