graphein icon indicating copy to clipboard operation
graphein copied to clipboard

Bug in calculating edges based on k nearest neighbors

Open tsjain opened this issue 6 months ago • 0 comments

In order to exclude residues less than long_interaction_threshold, the code checks for residue_number value in the pdb dataframe. The residue_number value is identical for residues with insertion codes. Therefore, it will not create neighboring edges between residues that have the same number but differ in the insertion code.

This is very severe bug because insertion codes are very common in protein structures.

https://github.com/a-r-j/graphein/blob/2ebd21100c8a2d6c092eb23dd1f75154962857c8/graphein/protein/edges/distance.py#L1146-L1147

tsjain avatar Aug 23 '24 23:08 tsjain