As in the title... I have a protein and I would like to know its secundary structure. I couldn't find it in uniprot, althought I tought they had annotations for it there. In the end I have used a predictor (jpred) but there it should be a database somewhere.
If you have the PDB file then you can use the standard tool called DSSP , it is supposed to be the gold standard for obtaining secondary structure. In case you just have sequence then I personally prefer PSIPRED , it takes evolutionary information into account to predict the secondary structure . According to CASP evaluation it is one of the best secondary structure predictor available.
I think you found the best answer yourself: use a predictor! There are several out there...
You suggest that there should be a Secondary Structure Database. I'm not sure that makes much sense, let me explain my point of view (which may not be that of everyone): most often, the data that is found in databases is the "state of knowledge" of the described object, based on experimentation.
That may be the case for secondary structures of proteins, but only in the case where the said proteins have been crystalized. In those cases, it is not only the secondary structures but also the tertiary structures (with the caveat that the crystal structure of a protein does not prove "all" states that a protein may take in real "dynamic" physiological conditions).
For all those proteins that have not been crystalized, then we can only rely on predictions. And I use them quite frequently: they are extremely useful! But as far as I know, no prediction is accepted as fact. They're "educated guesses" that are often correct, but sometimes wrong. The results may differ from one prediction method to another. Also they change each time the algorithms are improved...
If there was a database of predicted secondary structures, people would likely take them for granted (make the equation prediction = fact) which would be quite "unscientific".
I think such a resource would be more of a hindrance than an asset to the scientific community...
May be a little bit dated, but let me blow my own trumpet (collection of links):
If you want to obtain domains as well as the annotations that come along, you can do it locally with an RPS-BALST. Here for example to obtain Pfam annotations :
rpsblast -i ".$InputPath."/".$item." -d ~/Bioinfo/cdd/Pfam -e 0.000000000001 -o ".$elemt."_Pfam.rpsblast -T T -m 7
- i = the input path
- d = the database path
- e = the e-value cut-off value
- o = the output name
- T T and -m 7 = to have the output in XML format
You can download all the databases fromm CDD http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml. You'll obtain external source databases like Pfam, SMART, COG, PRK, TIGRFAM.
You have got all the answers needed for your query. The structure of the protein generally comes after the X-Ray crystallography (which crystallized 80% of the protein structure existed) or NMR technology.