I have a 1500+ residues sequence. Why alphafold2 colab is not predicting its strucrure and saying the session is crashed after using all available ram?
What exactly is 1500+? Is it 1502 of 2500? I am not being pedantic here - it makes a large difference. It should be easy enough to provide the exact number of residues. Generally speaking, your protein is in the range where it might not be possible for AF Colab to model it due to memory issues. We are talking GPU memory here rather than RAM, so it doesn't matter how much RAM or disk space you have. In such cases you may need to divide your protein into domains, or run it on a local GPU with more memory. Using a paid Colab account might also give you access to GPUs with more memory.
For these types of questions you will likely get a more informative answers by joining a ColabBio channel on Discord. The invitation is in the link below and it will expire in 7 days, starting from Feb 15.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
It sound like you got unlucky with the GPU you got assigned. The T4s they normally pair users up with can fold peptides up to ~2400 AAs currently (though my experience is with ColabFold rather than the a AF2 directly.
Have you tried more than once?
If so, please provide more information about the parameters you have used as we don't know anything other than the length of the peptide currently.
Yes, I have tried it multiple times but same problem occurs again and again. I have also tried it on a workstation with more than 1 TB space but same problem appeared there.
What kind of parameters are you looking for?
We don't know the length of the protein either, as 1500+ is not exact enough.