Finding Short Near Perfect Amino Acid Sequence Matches

Question

[annegr]

Dear everyone,

I hope you can help me with a question that I have.

I have an antibody that I believe cross-reacts with a protein other than its intended target. I have performed an immunoprecipitation using this antibody and send the immunoprecipitate for mass spec, and think it likely that this spurious protein will be among the list of IP'ed proteins.

What I would now like to do is see if any of these candidate proteins have any short (near) perfect amino acid sequence matches to the full length protein that was used to raise the antibody, and would therefore explain the cross-reactivity. As an epitope for antibody recognition doesn't have to be long, I am not looking to align full length proteins but rather am looking for very short regions of very high identity.

It seems to me this is quite a straightforward query but I have not figured out yet how to do this. I tried using alignment programmes but they just try to align the full length of the proteins. I am not very bioinformatically literate in the sense I don't use command lines etc. so please help me in the style of "for dummies."

Thank you so much for your help and advice.

Best wishes,

Anne

Answer 1

[fridhackery]

Use protein blast. Only enter the short epitope sequence. Consider limiting the search to your favorite organism, if there are enough protien entries.

http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins

[fridhackery]

The epitope sequence should be known by the maker of the antibody, including "homemade" antibodies.

[annegr]

Hi,

Thank you for your help. The problem is, I used the full length protein (500 amino acids) to raise the polyclonal antibody against - therefore I do not know what the epitope is! And if I use the whole protein then I get it trying to match any candidate proteins to the whole protein. I really appreciate your help!

[fridhackery]

So this has really become a molecular biology thread, rather than BioFX. You should do the BLASTP exercise that julieN suggests but with the intent of discovering several realtively unique regions of your favorite protein (like three). Then raise new antibodies to these epitopes. The new antibodies will have a chance of being specific.

Answer 2

[julieN]

Take your most abundant proteins with high-quality hits from your MS run, and BLAST each one individually against your original protein using the two-protein BLAST at NCBI:

http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins&PROGRAM=blastp&BLAST_PROGRAMS=blastp&PAGE_TYPE=BlastSearch&SHOW_DEFAULTS=on&BLAST_SPEC=blast2seq

Under the algorithm parameters, use:

expect threshhold = 100 (or higher)

word size = 2

compositional adjustments = none

low complexity filter = off (unchecked)

You're not looking for a significant alignment in terms of evolutionary relatedness, just whether short sequences between two proteins match. Once you've got your candidates, you might want to try blocking with synthetic peptides to the sequences to see if they are really what your antibody recognizes.