A hybrid knowledge-based and empirical scoring function for protein–ligand interaction: SMoG2016
T Debroise, EI Shakhnovich… - Journal of chemical …, 2017 - ACS Publications
… the choice of parameters did not depend on the training set, we … Errors were estimated with
the standard deviation (SD), … is the use of a single-point approach to estimate a value that is a …
the standard deviation (SD), … is the use of a single-point approach to estimate a value that is a …
Beware of Machine Learning-Based Scoring Functions On the Danger of Developing Black Boxes
… Training machine learning algorithms with protein–ligand … not learning any type of
protein–ligand interaction, since we do … of either positive or negative weights assigned to favorable …
protein–ligand interaction, since we do … of either positive or negative weights assigned to favorable …
[HTML][HTML] DLIGAND2: an improved knowledge-based energy function for protein–ligand interactions using the distance-scaled, finite, ideal-gas reference state
… but negative partial charge, which is repulsive to the negative … over-estimate due to protein
homologs between training and … DEKOIS 2.0 dataset to evaluate DLIGAND2 and RF-Score-…
homologs between training and … DEKOIS 2.0 dataset to evaluate DLIGAND2 and RF-Score-…
An accurate free energy estimator: based on MM/PBSA combined with interaction entropy for protein–ligand binding affinity
K Huang, S Luo, Y Cong, S Zhong, JZH Zhang, L Duan - Nanoscale, 2020 - pubs.rsc.org
… detailed energetic investigation of protein–ligand interaction. … the scoring function, which is
widely used to estimate quickly … for 10 protein–ligand complexes in the training set to 10 ns. …
widely used to estimate quickly … for 10 protein–ligand complexes in the training set to 10 ns. …
Convex-PL: a novel knowledge-based potential for protein-ligand interactions deduced from structural databases using convex optimization
M Kadukova, S Grudinin - Journal of computer-aided molecular design, 2017 - Springer
… derive our scoring function for protein-ligand interactions, … training set, we used randomly
chosen 80% of protein-ligand … with amide oxygen; c negatively charged oxygen with positively …
chosen 80% of protein-ligand … with amide oxygen; c negatively charged oxygen with positively …
[HTML][HTML] BgN-Score and BsN-Score: bagging and boosting based ensemble neural networks scoring functions for accurate binding affinity prediction of protein-ligand …
HM Ashtawy, NR Mahapatra - BMC bioinformatics, 2015 - Springer
… R p and/or R s values (minimum is negative one). Another measure of … scoring performances
of NN and RF SFs on the training set … protein-ligand interactions. Thus we find that BsN-…
of NN and RF SFs on the training set … protein-ligand interactions. Thus we find that BsN-…
[HTML][HTML] Multi-PLI: interpretable multi‐task deep learning model for unifying protein–ligand interaction datasets
F Hu, J Jiang, D Wang, M Zhu, P Yin - Journal of cheminformatics, 2021 - Springer
… negative ligands for targets, as exhibited in the last section. In addition, different datasets may
have their own protein-ligand interaction space, and models trained on … (a docking scoring …
have their own protein-ligand interaction space, and models trained on … (a docking scoring …
[HTML][HTML] SMPLIP-Score: predicting ligand binding affinity from simple and interpretable on-the-fly interaction fingerprint pattern descriptors
… been applied to encode protein–ligand interactions. The fingerprint … The DNN model was
trained with tunable parameters that … the same set of training, validation, and test data that had …
trained with tunable parameters that … the same set of training, validation, and test data that had …
KDEEP: Protein–Ligand Absolute Binding Affinity Prediction via 3D-Convolutional Neural Networks
… of use of K DEEP makes it already an attractive scoring function … In our case, the total number
of learnable parameters adds … this by using the PDBbind full minus core set as training and …
of learnable parameters adds … this by using the PDBbind full minus core set as training and …
[HTML][HTML] The impact of cross-docked poses on performance of machine learning classifier for protein–ligand binding pose prediction
… docking due to deficiency of scoring functions (SFs) and … that characterize protein–ligand
interactions and experimental … difficult set because most poses are marked as the negatives…
interactions and experimental … difficult set because most poses are marked as the negatives…