Hi,

Besides Uniprot, which is an extremely rich protein centered database, you can use Ensembl:

http://fungi.ensembl.org/Neurospora_crassa/Info/Index

Ensembl is dedicated to gene annotation and integrates manual as well as automatic annotation.

You can get access to the latest gene/transcript translations (.fa file) here:

ftp://ftp.ensemblgenomes.org/pub/release-26/fungi/fasta/neurospora_crassa/pep/

Most of the encoded proteins seem to be predicted or putative characterized.

The main problem with uncurated databases is that it is becoming increasingly difficult to separate the wheat from chaff, i.e. tracking down which annotation can be trusted as most of it suffers heavily from the "chinese whispers" problem.

If you want to get a lead what a certain "hypothetical protein" is doing, I would start searching against the PFAM and TIGRFAM databases. This might give you some leads on conserved domains and thus on the possible function.

NCBI nr database? Or  use tool like HMMER to do the alignment.

Thank you very much Ilian Atanassov, Christian Rückert and  Guojun Wu!! I'm going to try to use your recommendations