Use the prosite scan

http://prosite.expasy.org/scanprosite/

use option 2, you can input a sequence of amino acids for each postition

Ctrl + F ?...

No kidding, that's the easy solution... If not convenient due to an enormous db that can't be loaded in RAM, then google for "pattern search".

I'm sure the most convenient way to do it would be "grep" functions on a multi-fasta file...

Dear Alexander Streng, as much I understood your query, you want  only the sequence which matched your query sequence with 100%identity?

First of  all I would like to tell you, that with the help of BLAST you can do this you have to choose BLAStp. If you have seen the blast result clearly you will find that the outcome of BLAST shows three main things 1. Sequence length, 2. Query covered, 3. %identity. Here i would like to suggest you that, when you rung BLAST with your query sequence you should see that the result will be like that (query covered 100%, Identity 100%) if these two things will be 100% then it seems that the query sequence or peptide sequence is exactly same.

I think this information will be helpful for you.

Best of Luck

protein BLAST

PHI-BLAST option

greetings

http://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastp&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome

I'm thinking that you could use fuzzpro from the EMBOSS package. What you are trying to do is actualy finding a motif in a set of protein. Try to have a look.

Best

http://emboss.sourceforge.net/apps/release/6.6/emboss/apps/fuzzpro.html

Hi! Try HMMER, it is  great  search engine for protein sequences.

http://hmmer.janelia.org/

http://hmmer.janelia.org/

You can use the LinDa. It is a database of peptide fragments extracted from 13000 proteins. You can search based on Amino Acide sequence, Polarity pattern, Secondary Structure and Hydrophobisity pattern.

http://bioinf.modares.ac.ir/software/linda/

I hope you find if useful.

http://bioinf.modares.ac.ir/software/linda/

Like Joseph I would go for grep. However, I already experienced RAM problem with grep as well... hehe... bloody large datasets. Then I turned toward awk.

As long as you know what the file you are browsing looks like, you can specify which pattern you want to get, and if the pattern is found, which column, line, group of lines you want to get in the output (for example if it is a fasta file, you would want to get the protein sequence which is the current line, and the protein name, which is the line before the matching pattern, assuming a non interspaced fasta).

Try a BLAT search of the sequence on UCSD Genome Browser page

It is very good and has some great tools added.

http://genome.ucsc.edu/cgi-bin/hgBlat?command=start

go http://pacsy.nmrfam.wisc.edu, and send query at the bottom.

For instance, this is for finding GXXXG motif.

select * from SEQ_DB where SEQUENCE LIKE '%G___G%'

You can find exactly matching sequence like this instead:

select * from SEQ_DB where SEQUENCE LIKE '%GEQLG%'

However, this is for NMR structures...

Try conserved domain database:

http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml

Thank you all for your answers!

I am going to give all suggestions a try and report back what worked best in my opinion.

To Yogesh Kumar; Blast would be the preferred algorithm to use. However, what Blast does is Query a lot of different sequences based on the sequence you input. If you input a sequence in Blast, you can get a 100% identity hit which does not match your sequence. In that case blast identified a protein that matched your entire sequence minus 1 or 2 amino acids on either terminus. For some reason it still calls it a 100% identity.

If this is not what you meant then please elaborate. But I found that doing a blast search,sorting your results on identity and picking only the 100% identities still did not produce the results I was hoping for.

Regards, Alexander

Use the prosite scan

http://prosite.expasy.org/scanprosite/

use option 2, you can input a sequence of amino acids for each postition

You can use blast+ suite quickly.

If you have an access to a computational cluster (with blast+ suite installed on it):

-> blastp -task blastp-short -db nr -outfmt "6 qseqid qlen sallgi sallacc nident positive qstart qend sstart send bitscore qseq sseq" -query input_fasta.txt -out res.txt

Or on your computer with NCBI database:

- install blast+ suite on your computer (ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/)

- open a shell (invit command for windows).

- select a directory with your sequence (in fasta) file: input_fasta.txt

(cmd: cd my directory)

- paste this command line: blastp -task blastp-short -db nr -outfmt "6 qseqid qlen sallgi sallacc nident positive qstart qend sstart send bitscore qseq sseq" -query input_fasta.txt -out res.txt -remote

- open the results on Excel or R for exemple.

- Select the best bit score lines, and after calculate the identity percent of your query by dividing nident/qlen*100.

The expasy web site has quite a few tools on peptide protein motifs. I guess that it is one of these programs that you want to look at.

Best wishes.

Adding to what i said before here are some sites that may be of help:

http://pir.georgetown.edu/pirwww/search/

http://www.uniprot.org/help/uniref

http://www.expasy.org/tools/

http://molbiol-tools.ca/Motifs.htm

just to name a few options.

I hope that they will be helpful to you

Nikos

First of all, I want to thank all of you for your contributions.

In my opinion the two tools that work best (get a comprehensive list of proteins containing the exact sequence you query) with minimal amount of effort are:

Prosite, using the second option:

http://prosite.expasy.org/scanprosite/ (thanks Alexander)

and the peptide match tool from PIR:

http://pir.georgetown.edu/pirwww/search/ (thanks Nikos)

Or download the complete FASTA-file of your database of interest and use CTR+F, of course ;-) (thanks Joseph)

I actually tried it and it does work. Although it kills your computer and it is quite tedious to do with 20+ sequences.

Alexander Streng, have you tried the PHI-BLAST algorithm for BlastP?

It does a similarity search and returns only sequences that contain your pattern ....is it what you were looking for? 

Hello,

If you're using linux, try grep command ...  grep "your string" filename ... if you have multiple, make a file of the sequences, without headers, and put it in loop... Grep doesn't loads the file to memory. else you can write a code in the language of your choice...

@Anubhav Das - while using grep has a lot going for it, esp. due to simplicity,

a care should be taken in cases where the query sequence stretches across a line break (or several) - as such breaks will not be found :). The obvious solution will be to use FASTA without line breaks, but that adds an additional step, so not so convenient anymore.

MS Homology on the Protein Prospector does exactly this from the DB and species of your choice.

http://prospector.ucsf.edu/prospector/cgi-bin/msform.cgi?form=mshomology

you can paste just the peptide sequences for exact matches.

You can use Peptide Match from PIR

http://research.bioinformatics.udel.edu/peptidematch/index.jsp

Visit Uniprot.org

Use the peptide search option

http://crdd.osdd.net/raghava/satpdb/

Most of the therapeutic peptides have been curated and stored in this database, this must be helpful authentic source for your query.

Use UniProtKB website: https://www.uniprot.org/help/uniprotkb

then use peptide search section.