Open anton-bushuiev opened 1 year ago
It also seems that here the sign is swapped if I understand correctly (i.e. destabilizing mutations are a positive class): https://github.com/luost26/RDE-PPI/blob/f6d68a7331a9f65290f525f0a491848245b072b3/rde/utils/skempi.py#L127
Hi, @luost26! Great work! I have a question regarding the evaluation.
I am comparing the evaluation of RDE-Net and RDE-Linear. Why are there fewer entries in the case of RDE-Linear? Where are they filtered?
The "result" dataframes for RDE-Net and RDE-Linear which are the input to the final![image](https://github.com/luost26/RDE-PPI/assets/67932762/bffe304d-a4ec-4a3b-a390-9b53119fb165)
eval_skempi_three_modes
: