tricking alphafold2 docker to use the same reference sequences as mmseq2

sokrypton / ColabFold

Making Protein folding accessible to all!

MIT License

2.01k stars 504 forks source link

tricking alphafold2 docker to use the same reference sequences as mmseq2 #21

Open avilella opened 3 years ago

avilella commented 3 years ago

Hi, I've been running the MMseq2 ColabFold for a specific type of protein sequences, and I always get the same set of ~4000 sequences in the .a3m file. Would it be possible to put these sequences from the .a3m file in a file in one of the folders for the local installation of AlphaFold2, and tricking AlphaFold2 to always look at these when running?

sokrypton commented 3 years ago

Are the protein sequences always the same length? If so, should be possible to hardcode to always use same a3m file.