Things to do

[Hrovatin]

List of TODOs:

• Add functional submission form.
• Maybe add tie-breaking for Allosteric-Orthosteric Ligand Prediction Challenge

[Hrovatin]

List of questions:

• Q1: Can there be multiple ligands per protein @tobiasploetz ? Since it is a list
• Q2: Was jsonl renamed? The one I downloaded does not have "_dataset" at the end
• Q4: What is the point of 5 predictions if you only asses the first one (from the list?) ?

The evaluation script will compute the Capri-Q docking assessment classification scores (high, medium, acceptable, incorrect, error) for each of your top 5 predictions per data point.
It will then print the distribution of classifications for the top 1 predictions across all data points.
Additionally, it will compute the number of "successful" predictions, i.e., the number of data points for which the top 1 prediction is classified as "acceptable" or better.

[tobiasploetz]

Q1: We initially thought about having multiple ligands, but we now ended up with just a single ligand per data point. I added some clarifications.
Q2: fixed, thank you
Q4: imho, it is interesting to have 5 predictions to calculate some more stats