1
$\begingroup$

I have a 32-bp query that I want to perform BLAST+. my query represents different RNA viruses and I do expect that my hits get more than 500,000. but in practice, when I perform it I get around only 700 hits. would you please kindly help me how to increase the efficacy of my BLAST+? I leave my syntax here.

blastn -task megablast -db nr -outfmt 7 -remote -max_target_seqs 99999999 -perc_identity 80 -query D:/Query.fas  -out D:/result.txt
Improve this question
$\endgroup$
3
  • $\begingroup$ Can you explain why you expect many queries? More importantly, what is the biological question you are trying to answer here? Are you looking for distant homologs? Close homologs? Why did you choose to use megablast? That's an odd choice for such a short sequence. $\endgroup$
    – terdon ♦
    Commented Oct 26, 2023 at 12:49
  • $\begingroup$ Thanks for your comment. I expect many hits because there are many hits for my query as I know my query is common among RNA viruses, it is such a conservative seq. Yes, I want to extract all sequences in the NR database that are similar to my query. I think megablst was a bad choice, because when I choose -task blastn my hits were significantly improved (+500,000), but still I encounter an error by blast+ stating "search cannot be formatted" resulting in finding just some of those hits. $\endgroup$
    – Farzad Beikpour
    Commented Oct 27, 2023 at 19:53
  • $\begingroup$ Please edit your question and explain this. Megablast is only useful for very similar sequences. And long ones, usually. It doesn't really make sense to use it for a 32bp query, that will be smaller than the smallest thing it can find. So please edit and explain what you are trying to do. You don't want similar sequences, presumably you are looking for homologs, right? I mean explain the biological question you are trying to answer, so we know how to help. $\endgroup$
    – terdon ♦
    Commented Oct 27, 2023 at 21:04

1 Answer 1

Reset to default
1
$\begingroup$

Megablast is designed to find very similar sequences. It is highly specific but has a very low sensitivity. This is because it has a very large word size. By default, I think it will use a word size of 28 nt. That's almost your entire sequence, so if your target sequence doesn't have 28 contiguous nucleotides that are identical to 28 contiguous nucleotides of your query, no hit will be found.

If you are looking for distant homologous sequences, you should use simple blastn which defaults to a word size of 11. That will still likely be too small, you might want to change it to something lower (your blast tool should have an option for that), but it will give you more results than megablast.

Finally, if these 32nt are translated into protein, you can consider using tblastn instead.

Improve this answer
$\endgroup$

Your Answer

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

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