Frederic PLEWNIAK wrote in message <34C4793D.919 at igbmc.u-strasbg.fr>...
>> I think it's not only an expected GCG ouput, it is also because
>> different version of blast gives different output. We experienced such
>> problems with DBWatcher...
>> If I remember correctly, Fred found a way to parse easily
>> both outputs, but, Fred is out this week, so he will confirm next
>> week if I'm right.
>>>Oh! my... Just comin' back from a sweet quiet week off lab and now...
>François, you're too fast to jump onto your keyboard ;o)... Sorry to
>let you down but DBWatcher doesn't really parse title lines in BLAST
>output. Not the way SRS does anyway! And it is certainly not able
>to tell the accession number from the ID... Sorry...
well, too bad! I think the only way to get around this is to create separate
versions of .i and .is for different versions of blast.
This would be reasonable since not only the format of the output but also
the program options change
parsing the title line for entry names and accession numbers is another
problem and could be alleviated by including productions for both in the
syntax where only one is active at a time. I used a similar solution for
switching between gcg and native format for most sequence database .