Yann Dumont <[email protected]> writes: > Just one point, as I read your answer : the GI number mentioned in my > message is from the query file, not from the subject file. It is just > a precision, since the work-around will still work.
It looks like this behavior occurs only when explicitly using -parse_deflines (or -J, if you're going through a wrapper emulating legacy BLAST), at which point it comes down to library code with no override facility at present. As such, a simpler workaround should be to stop passing that option, though you might still need to postprocess the result to reinstate the original ID. At any rate, it should be possible to add a mechanism to force that library code to embed the literal query sequence even if it carries an ID from NCBI, EMBL, or the like. Having this mechanism arrange to consult a new command-line option would be cleanest, but require extra plumbing, so I am considering having it instead consult a new environment variable for simplicity. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/[email protected]
