[stability-pipeline] sequence missmatch due to Rosetta conversion of unnatural AA within sequence
Closed this issue · 2 comments
j0kaso commented
If you have a structure that contains unnatural AA within their sequence, the pipeline currently doesn’t kick them out but ignores them in the sequence (only taking HETATM). But Rosetta relax converts them into standard AA (at least in this case). So when the fina
l prism file should be written, it has a mismatch with the sequence (which originally contains a gap and is 1AA shorter than the one from the relaxed structure).
As illustration:
initial sequence = Q-Q
relaxed sequence = QYQ
pdb-file:
ATOM 3545 N GLN A 529 25.701 43.573 88.678 1.00 86.82 A N
…
ATOM 3553 OE1 GLN A 529 30.082 41.865 92.239 1.00 82.80 A O
HETATM 3554 N PTR A 530 28.685 45.294 88.598 1.00 79.23 A N
…
HETATM 3566 P PTR A 530 24.278 46.108 82.041 1.00100.00 A P
…
HETATM 3569 O3P PTR A 530 22.883 46.452 82.450 1.00100.00 A O
ATOM 3570 N GLN A 531 32.082 45.905 88.540 1.00 79.98 A N
…
The pipeline will currently give you an error when you haven’t cleaned that file (for full saturation mutagenesis tested)
(file can be found on deic /groups/sbinlab/tiemann/projects/PRISM/debug-files/2020-11-23_Matteo/ )