… between good and bad docking solutions. Our results … here because the protein-ligand
interaction networks constructed … of the interacting atom type pairs for best fit of the training data. …

… and target are unseen in the training set. Yamanishi et al. … false negative data for some
ligands could actually interact … The AUC score of BLM-NII is 85.8% in our data set, while it is …

… with the construction of training data of protein–ligand complexes, … After training, weights and
parameters of the step function are … more negative binding scores (high binding affinity). For …

… and rationally engineering protein–ligand interactions. … to rigorously estimate biophysical
parameters from massively … , we developed the quality score S training , which measures model …

… data of protein–ligand complexes, which allow the training of … for protein–ligand complexes
in each target-specific scoring … Fine-tuning the parameters in model-training stage using a …

… site, and scoring, the quality assessment of docked ligands. … fall back on benchmarks and
training sets for evaluating new … and this type of negative data could be equally informative. …

… models in scoring protein–ligand interactions. Finally, to gain … , and thus errors will be
accumulated, leading to bad scoring. … useful to the study of protein–ligand interactions. In practice, …

… binding affinities of original training dataset with very high … posing and scoring abilities on
the large dataset of protein–ligand … useful quality parameter is the percentage of protein–ligand …

… affinity data of a training set of protein−ligand complexes with … Normally, the solvation
parameter σ i has a negative value … a reliable estimation of eq 10 requires the use of a set of well-…

… true binders versus negative binders) using the PDBbind 2013 … the training set and used
to evaluate the scoring power of the … file using the fixmol2 tag, then scoring the protein–ligand …