How to predict a structure if it is less than 30 aminoacids long?

Question

[camilacv]

Hi, there! I'm a soon-to-be graduate biology student and, for my final project in my course of bioinformatics, I decided to work with protein interactions, comparing a wildtype and a mutated aminoacid sequence. Everything went well until I had to find a model for my mutated aminoacid sequence. The thing is, the protein I'm working with has 25 aminoacids. And the databases require at least 30 aminoacids for this analysis.

Fortunately, the wildtype protein is easily found on the Protein Data Bank. But I would like to search for models for the mutated aminoacid sequences I generated according to some findings in a paper. I edited the wildtype sequences so that it fits the description.

I'm wondering if there is a DataBank for small aminoacid chains. Or maybe I'm missing something.

Regards

[Buffo]

That is not a mutation, is a deletion, furthermore "protein" of 25 aminoacids is a functional protein? I think you can't predict 3D structure because it doesn't exist, it is more a polypeptide than a protein.

[WouterDeCoster]

A deletion is also a mutation.

Answer 1

[Joe]

If there is a known good PDB structure, use a threading program like Phyre or ITASSER. I've used ITasser on peptides as short as 10AAs in the past. It should use the closest reference sequence/structure in NCBI for the template, so I'd expect you to get a reasonably decent structure back out.

The input to these programs is just a fasta sequence (with no ambiguous amino acids), so you can 'hand write' your mutated fasta sequences and just let it run.