Thank you for your insightful and beautiful work!
I've reviewed the current preprocessing script. While it successfully extracts the ground truth data , the output .pkl files lack the scaffold_prior and arms_prior information required for model training.It appears that a subsequent preprocessing file/script is missing—one that should load these .pkl files and calculate the statistical or geometric priors (e.g., computing the vectors from scaffold anchors to pocket centers).
Is there a missing script intended to run after this one?Thanks for your valuable time and efforts.