I generated a high quality de novo genome assembly which has been annotated using previously published isoseq data from the species. This genome is now the Refseq for the species. I am currently writing a manuscript detailing the assembly. My PI suggested that I add a quick biologically relevent analysis to the text to show the utility of my genome. Specifically, she said I should investigate if any of a list of candidate genes we have for a specific trait are duplicated or expanded.

To do this, she said I simply need to blast the sequence of my gene of interest , as annotated in this assembly, against the whole assembly. I have looked for papers which do this to reference, and in sifting through them have only succeeded in confusing myself.

As far as I understand, the specific method to do this is the same as this method to identify paralogs using blastp (I am unsure why i would use blastp to identify duplicated genes, but the papers I am reading all seem to agree on blastp instead of blastn):


where I

  1. take the FASTA nucleotide sequence for my gene of interest (as determined by the annotation of my genome) and blast (do I use blastn? or blastp?) specifying the database as nr (nonredundant protein) and organism as my species of interest

  2. Once I get my blast results back, my top hit will be that same gene I blasted

  3. If any other results have a significant E value and score, those are potential paralogs/duplicated genes? How might I verify or validate that. or could I only say these are putative paralogs?

Am I missing anything? Is there a way to screen all annotated genes for duplication/paralogs that would make more sense then repeatedly blasting through the list? I am very lost as to how to proceed

  • $\begingroup$ This is typically done by blasting against translated coding sequences. Since you have annotated your genome you can extract and translate all CDSs and then use those as the blastdb to count how many "similar proteins" you have in the genome for your query. If there are more than in whatever you are comparing to it would probably be an expanded family. $\endgroup$
    – Niklas
    Commented Apr 16 at 17:24
  • $\begingroup$ got it, so I blast my translated nucleotides against my species whole genome using tblastn? $\endgroup$
    – BMM
    Commented Apr 17 at 18:29

1 Answer 1


Genome sequence means nucleotide. If you want to search for duplicated genes or paralogues in one go you better use protein query vs translated nucleotide tblastn. The main problem you may encounter are repetitive sequences which at times are included in the translated genes.

Apart from BLAST there are other programs. Check https://academic.oup.com/bioinformatics/article/39/1/btad014/6989621


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.