Hi Milan,

Your idea is in principle interessting. If this would be possible, a lot of problems would have been solved. To mention only a few problems:

The proteome is simply too big as this would be possible even with ultra-high-throughput screenings. And its getting even harder if you are on the level of" "Individuals" So it is possible that they have a mutant in some protein, that is not known.

In general you can use a proteome database for matching amino-acid sequences. But:

1) PTMs are usually not included

2) you need to know the sequence of your protein

3) Sequence matching does not necessarily means, the epitopes are the same, they can have different 3-D structures.

Best regards

Peter

Thanks Peter Carl for your comment.

I would then like your opinion of this: We made a longlist of about 500 proteins that might be involved in the process that we're investigating. Would it then be possible to blast against these 500 and find out IF there are common epitopes (according to the amino acid sequence). Then, if there are matching aa sequences, investigate those further, controlling for PMTs and 3D structures?

Thanks in advance!

Milan

Hypothetically yes. But what do you mean: Do you only want to compute this or do you want to really "test" al of them fro cross reactivity in an e. g. Immunoassay?

This is all very hypothetical, I am not certain it will actually be helpful, you might find a matching sequence, but what does this tell you? Will not be so helpful.

For the first one, you need a bit programming skills: You need to know the exact sequence of each of the 500 proteins (and your target) of cause. Some might be found in Proteom databases e.g. UniProt. (https://www.uniprot.org/) There you can find the sequences of interests.

But should be careful: The sequence is not determined by protein sequencing it is from DNA Sequencing and then assumed, that the transcripts are the proteins (usually, there might be exceptions)

At least you can have a look where something is "matching". Maybe this is interessting?

For your outcome there might be some problems:

- Some Sequences are "highly conserved" and seems to be found in a lot of samples. They will for sure not be your target.

- PTMS, like glycosylation, which are very important for immuno-recognition are not included.

So in the end you might have a result that is "useless".

There are btw. also some epitope databases: Maybe you can search for it there?

https://www.iedb.org/

I think it would be more practically searching for this protein "in the patient" and then finding out what it is. You said it might be recognized by an antibody? Super! You can do an affinity enrichment. And then you can do mass spec for analysis.

Best Regards

Peter