How to specify Cut-off/parameters of sequence identity > 70% and query length > 40% in BLAST command line?

Question

[Kumar]

I have been searching homologues sequences of my target gene families in my concerned genome sequences. I need to search the homologues with a score of > 70% identity and > 40% query length to the target gene families. Is it possible to include these cut-off/parameters in the BLAST command line. Thank you in advance.

[Eugene]

Blastn v. 2.13.0 support % query length coverage; from blastn -help:

-qcov_hsp_perc <Real, 0..100> Percent query coverage per hsp (high scoring pair, aka a continuous alignment between query & subject)

in addition to perc_identity cutoff.

Answer 1

[gb]

-perc_identity

There is no coverage cutoff parameter, so you need to do that yourself afterwards. (It is called qcovs in de outfmt options)

Here you can find all the parameters: https://www.ncbi.nlm.nih.gov/books/NBK279684/

Or use:

blastn -help

[Kumar]

Thank you gb, if it possible could you please elaborate qcovs in de outfmt options.

[gb]

I believe there is no filter parameter for that but you can output it and filter it afterwards, these are the option:

qlen means Query sequence length
slen means Subject sequence length
length means Alignment length

EDIT you changed the text of your comment:

qcovs is the percentage of basepairs from your query that is used for the alignment against the target. So with an coverage of 90%, the of your query that is used for the alignment with the target that is 90% of the total amount of basepairs or your query.

[priyankamowlali]

hello gb, Can you help me out with this? Whenever I use perc_identity it shows this "Error: (CArgException::eInvalidArg) Unknown argument: "perc_identity", Im using tblastn

[gb]

You sure you use blastn? Also don't forget the dash. You can double check the parameter with the blastn -help command.

EDIT: The error is correct. tblastn does not have such option. https://www.ncbi.nlm.nih.gov/books/NBK279684/

[priyankamowlali]

yeah, just checked it, thank you. So, if I have to get the sequences with >=50%, i have to manually check them right?

and just a confirmation pident is used for identity percentage, right?