Problems with fuzzpro and ehmmer

Kimmo Mattila kkmattil at csc.fi
Thu Dec 13 13:08:52 UTC 2001


Dear EMBOSS people.

I have had few problems with fuzzpro, patmatdb and ehmmer. If anyone of
you have suggestions how to solve them, please tell.


FUZZPRO and PATMATDB

I am using fuzzpro and patmatdb with GCG formatted databases. If I run a
search against whole database (e.g. swiss:*), the programs do find the
right hit sequences, but pick wrong names for the found entries. With
plane sequence files or with sequence name lists, this error does not
occur. I have checked both the EMBOSS indexing and the GCG database files
and they should be OK. Other EMBOSS and GCG ­applications give correct
results, when same database files are used.

Has someone else had similar troubles? If the indexing of the
databases is in order, what might cause this?


EHMMER

We have successfully installed EMOBOSS-HMMER, however, unlike the native
HMMER, the emboss version is not able to use multiple processors (even
though ­cpu option is mentioned in the help data.) When I compared the
Makefile of EMBOSS-HMMER to the native one, in noticed that the EMBOSS
version lacks the settings for compiling multiprocessor version of HMMER.
Has someone managed to circumvent this with some simple trick like copying
some parts of the original HMMER Makefile to the Makefile of
EMBOSS-version?

Secondly, when I use ehmmsearch long output files are not complete.
After about 200 lines lines ehmmsearch starts writing the output to the
screen instead of the output file. The last line in the output file seems
to be

 Domain top hits:

And after this the alignments are printed to the screen.
What might cause this? Is there e.g. some limit in the output file
size.

Regards,

Kimmo Mattila

---------------------------------------------------------------
   Kimmo Mattila		Science Support
   kimmo.mattila at csc.fi		Center for Scientific Computing
   tel. +358 (0)9 457 2708	Tekniikantie 15a D, PL 405
   fax. +358 (0)9 457 2302	FIN-02101 Espoo, Finland
---------------------------------------------------------------







More information about the EMBOSS mailing list