Open poloarol opened 2 years ago
Tested it, and it does not model the interactions as I would have liked it. Would probably consider it, for an ML ready dataset.
Hi @poloarol - what did you not like? Happy to make improvements.
@a-r-j - I was working with complexes and needed to separate them. However, I do not think graphein can do that hence opted for the biopython approach. Maybe I'm not just well versed with graphein and it is possible.
Move to graphein to represent the peptides, as it is built on networkx and much more robust. Provides a good interface for Geometric Deep Learning downstream