How to do BLASTP for short sequences (<20 aa) effectively?

Question

[Riq]

I am trying to find sequence homology between viral sequences and my protein of interest. I have the sequences of their epitopes which varies from 5 to 500 amino acids long. For shorter sequences, it is recommended to use blastp-short but I am confused until which length it is okay to run blastp?

For shorter sequences, what are the parameters needs to be added for effective results?

blastp -task blastp-short -query POI.fasta -db virus.fasta -out output.txt -outfmt 6 -evalue 20000

Answer 1

[guillaume.rbt]

As you mentionned, you have to use the option blastp-short when blasting short sequence.

This option will modify several blastp parameters, see here : https://www.ncbi.nlm.nih.gov/books/NBK279684/table/appendices.T.blastp_application_options/

It is explained there that this option should be activated for sequences shorter than 30 residues.

The important point is that the word size will be set to 2. This means that you query sequence will be split into "words" of 2 amino acids, to find alignement in the selected database. Hence the parameters set with blastp-short should be ok for your sequences from 5 to 30 residues.