Blastall vs Blastn

Question

[sicat.paolo20]

Good day everyone, I am currently trying a workflow that was done with blastall version 2.2.26. I wanted to update it and use the 2.8.1 version of blast that we have. I have this one parameter that I am confused with. I'm currently trying to produce a tabular output (blastall -m 8 , blastn -outfmt 6), with the parameter -v (Number of database sequences to show one-line descriptions for (V)) for blastall and -num_descriptions in blast 2.8.1 is not compatible with 2.8.1 version of blast with a tabular format. I seem to not understand what exactly this does and how it affects my output. Can anyone help me understand this?

Answer 1

[lieven.sterck]

outfmt 6, does not provide descriptions summary, so only the -num_alignments is relevant.

The old blastall was a bit more lenient towards those two parameters, the blast+ suite is more strict on it

and I personally prefer -max_target_seqs in tabluar context

[Bastien Hervé]

If you did not see it yet, use -max_target_seqs with causion

https://blastedbio.blogspot.com/2015/12/blast-max-target-sequences-bug.html

[sicat.paolo20]

Thanks for the heads up. My supervisor sent me this https://academic.oup.com/bioinformatics/advance-article-abstract/doi/10.1093/bioinformatics/bty833/5106166 to warn me about -max_target_seqs

[lieven.sterck]

I see, has been extensively discussed in this forum as well

there is a follow-up on that though: from the NCBI blast devs themself (thx @genomax) https://academic.oup.com/bioinformatics/advance-article/doi/10.1093/bioinformatics/bty1026/5259186

[sicat.paolo20]

Hi, Thank you very much for this. This was really helpful and its amazing how the devs explained the issue and resolution simple terms. Thanks!

[sicat.paolo20]

Thanks. I'm currently being cautious just to make sure I understand each parameter completely. Would -max_target_seq be the same as having both -num_alignments and -num_descriptions?

[lieven.sterck]

roughly yes.

the only (slight) difference is that -max_target_seqs is summarized per DB entry ID, while -num_alginments is more directly related to the number of HSPs. As such they might, and will, differ if there are multiple HSPs per hit sequence, in that case -max_target_seqs 1 will report all HSPs for that hit sequence , while -num_alignments 1 will only report 1 HSP

[sicat.paolo20]

Thank you very much for the assistance. I think I'll just stick with -num_alignments. Thanks!