Binder/Non binder per site?
alejandrovr opened this issue · 6 comments
Hello,
I was wondering if the prediction has to be made by site or globably. That is, will we be asked to predict binders and not binders for S1, for S2, ... or just binders and not binders for the whole protein?
Best regards and thanks for organizing these events,
Alejandro.
Do you have any input? I believe the current thinking is to release which sites the fragments bind in when we release the list of binders, and then give folks the option of what they want to do from there, but we aren't 100% decided.
Also @MehtapIsik and @bergazin
Hi Alejandro,
The fragments were found to bind at these sites. Thus, it would make more sense to focus your efforts there. However, the presence of false negatives is a possibility and any information (including outside of the sites) that would enable us to identify those would be a plus!
I hope that helps,
Harold
Yes, sorry, I was not referring to the whole protein, but rather the difference between discriminating binder vs non binder at each site, OR discriminating binder vs non binder in any of the sites. So 4 lists of binder/not binder (S1,S2,S3,S4) or just one. I understand is the second option what you are looking for.
Yes. Sorry, I'm just submitting something to clarify this and also add the submission format.
Closed by #62