Blastx -max_target_seqs

Question

[sansan]

Hello everyone,

I want to annotate a transcriptome that I assemble de novo following the Trinotate workflow, however, I am not clear about the -max_target_seqs command, what is the ideal value?

blastx -db uniprot_sprot.fasta -query Trinity.fasta -num_threads 4 -max_target_seqs 1 -outfmt 6 -evalue 1e-5 > swissprot.blastx.outfmt6

Answer 1

[GenoMax]

Recommended value is now 5. There was a follow-up to the referenced paper from NCBI: https://academic.oup.com/bioinformatics/article/35/15/2699/5259186

The newest BLAST+ release (scheduled for December 2018) will issue a warning if the user requests fewer than five matches. Future releases may also issue a warning if further equivalent matches are not shown. However, we want to emphasize that, as with any database search, there is always the possibility of equivalent results being returned and it rests upon the user to determine which one(s) to choose depending on the user’s purpose.

[sansan]

Thanks so much for your answer.

Answer 2

[biofalconch]

I think it is important to read this paper: https://academic.oup.com/bioinformatics/article/35/9/1613/5106166

In short:

The invocation using the parameter ‘-max_target_seqs 1’ simply returns the first good hit found in the database, not the best hit as one would assume. Worse yet, the output produced depends on the order in which the sequences occur in the database

Not sure what the ideal value is, but it is definitely not 1