Okay, I come with new stuff. I realised that the bug happens only when I used fasta from NCBI, so I was assuming blast_formatter downloads unuseful stuff from NCBI before performing the formatting.
Actually formatter's behaviour seems correct. The problems comes from the ASN.1 file I use for query. I am generating these ASN files with blastx or blastp using fasta from NCBI so with classical labeling of each fasta sequence with GI number and everything else. I'm also use "home-made" fasta files I obtained from sequencing. With the home-made sequenced fasta I have no problem, only with blast results using NCBI's fasta. So I take a look in the the ASN.1 files : - with my sequenced fasta, it contains the different sequences of the fasta, encoded in hexadecimal or amino-acids (depending if I used fna and blastx or faa and blastp), after a "seq-data" label - with NCBI fasta, it contains only the GI number after a "id gi" label, and no sequence at all So I assume formatter need to get the sequence from NCBI before being able to convert my file ! And conclude that the bug come from the blast algorithm (at least the output formatter), not from the formatter, but only when ASN.1 format is used. Is there any way to by-pass this ? Telling blast to write sequence in the ASN file systematically ? Is it a expected behaviour ? -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
