Methods of locating new proteins' location on genome?

[linnyl]

Is there a method of finding out where a unknown protein (doesn't give any good BLAST hits) would be located on human or virus genomes?

All I have are the protein sequences themselves, which were derived from MS/MS data.

[GenoMax]

Does the "no blast hits" include reverse translated blast searches? If the sequence is correct and you know the genome it came from you should be able to locate them (euchromatic sequence of the human genome is reasonably complete at this time)

[linnyl]

Does the "no blast hits" include reverse translated blast searches? If the sequence is correct and you know the genome it came from you should be able to locate them (euchromatic sequence of the human genome is reasonably complete at this time)

Thank for the suggestion. I don't think I included reverse translated BLAST, I only ran blastp because I'm unfamiliar with the tools. Is the reverse translate BLAST = tBLASTn?

[GenoMax]

That is correct.

These are peptides? Not full length proteins. Are they short?

I assume you have already tried a mascot search (http://www.matrixscience.com/cgi/sea...ER=2&SEARCH=SQ)

[linnyl]

In that case I did run a few tBLASTn for my sequences, but they only mapped to various regions of a known sequence, multiple times in the same gene in some cases. (e.g. one of them had 17 matches to the AHNAK nucleoprotein as it was only 76aa long)

These are predicted to be proteins, using ORFs and mass spec data. But they're fairly short sequences (>200 aa)

At the moment I'm only using the sequences, it's the only actual data I've been given so I haven't had the chance to use mascot.

[Richard Finney]

genome.ucsc.edu/cgi-bin/hgBlat does protein to genome alignments.
It is very convenient to use on an ad-hoc basis for several genomes
Command line blat can be used for many alignments or on a genome not supported by UCSC.