Tweet
Hi all!
I have generated genome assemblies for my samples and would like to find the location of coding sequences in the contigs. Using coding sequences from a reference genome, I tried using command line BLAST+ (blastn) to do this, similar to what this paper does: https://www.nature.com/articles/srep38952#Sec7. It works, but I don't get the number of hits that I expect.
The reference (Pichia pastoris) has 5291 coding sequences.
Setting e-value to 1e-10 and perc_identity = 95, I expect to find close to this many coding sequences however I usually get about half. I tried blastn (2176 hits), blastn-short (2286 hits) and megablast (2220 hits). The genomes were deeply sequenced and assemblies are high quality. I tried other parameters like -num_alignments but nothing I tried gets me close to 5291.
What's stranger is that if I take 1 of the 5291 genes that did not yield a match and blast it on its own, I then get a match.
What am I missing?
I have generated genome assemblies for my samples and would like to find the location of coding sequences in the contigs. Using coding sequences from a reference genome, I tried using command line BLAST+ (blastn) to do this, similar to what this paper does: https://www.nature.com/articles/srep38952#Sec7. It works, but I don't get the number of hits that I expect.
The reference (Pichia pastoris) has 5291 coding sequences.
Setting e-value to 1e-10 and perc_identity = 95, I expect to find close to this many coding sequences however I usually get about half. I tried blastn (2176 hits), blastn-short (2286 hits) and megablast (2220 hits). The genomes were deeply sequenced and assemblies are high quality. I tried other parameters like -num_alignments but nothing I tried gets me close to 5291.
What's stranger is that if I take 1 of the 5291 genes that did not yield a match and blast it on its own, I then get a match.
What am I missing?