Protein Sequences MSA vs. BLASTp

Question

[Ronnie Ni]

Hello everyone here! I am recently working on ligand discovery. What I'm doing now is to find out if the ligand which targets the pathogen's specific enzyme will also have effect on human's.

I do BLAST to pathogen's protein to look for similar proteins in human. The identity and similarity given by BLAST are different from the MSA one(I use clustal omega). I wonder which algorithm is better to fit my work.

I am new to bioinformatics, and thank you for your patience.

[Dunois]

I wonder which algorithm is better to fit my work.

I don't suppose you're using a pairwise search tool (like BLAST) to find the sequences for your MSAs are you?

[Ronnie Ni]

Yes I use the results from BLAST to do MSA, but BLAST also gives me local alignment data(like local identity and positives). I do not quite understand which data I indeed need, the local one or the global one.

Answer 1

[cfos4698]

I assume it's the pairwise identity that's differing between tools? If so, it's potentially a difference in how gaps are scored. Try reading the info at this link: https://lh3.github.io/2018/11/25/on-the-definition-of-sequence-identity. There's some good info on Clustal here too: Make matrix of protein pairwise identities/similarities from multiple protein sequences. As for which is better? It's really context dependent, and how you interpret the results from either tool with respect to the question you're asking is the important thing.