Development of a New Scoring Function for Virtual Screening: APBScore

Bao, Jingxiao; He, Xiao; Zhang, John Z. H.

doi:10.1021/acs.jcim.0c00474

Cited by 11 publications

(13 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the past few decades, a large number of additive scoring functions have been developed and some of them, such as the Vina score of Autodock Vina, GoldScore, and ChemPLP of GOLD, GlideScore-SP, and GlideScore-XP of Schrodinger, Affinity-dG, and GBVI/WSA-dG of MOE, are widely used with corresponding docking software. In addition, some other methods, such as the X-Score, SFCscore, MedusaScore, LISA, KECSA, SMoG2016, PBSA_E, DrugScore 2018 , and APBScore, are used as scoring functions alone. In more recent years, machine-learning-based scoring functions, such as TNet-BP, Δ vina RF 20 , k DEEP , TopBP-ML, Δ vina XGB, AGL-Score, OnionNet, AK-Score, and IP-SF, have shown great improvement in the accuracy of binding free energy prediction on experimentally determined structures compared to other traditional scoring functions.…”

Section: Introductionmentioning

confidence: 99%

DeepBSP—a Machine Learning Method for Accurate Prediction of Protein–Ligand Docking Structures

Bao

Zhang

2021

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

In recent years, machine-learning-based scoring functions have significantly improved the scoring power. However, many of these methods do not perform well in distinguishing the native structure from docked decoy poses due to the lack of decoy structural information in their training data. Here, we developed a machine-learning model, named DeepBSP, that can directly predict the root mean square deviation (rmsd) of a ligand docking pose with reference to its native binding pose. Unlike the binding affinity, the rmsd between the docking poses with reference to their native structures can be straightforwardly determined. By training on a generated data set with 11,925 native complexes and more than 165,000 docked poses, our model shows excellent docking power on our test set and also on the CASF-2016 docking decoy set compared to other major scoring functions. Thus, by combining molecular dockings that generate many poses with the application of DeepBSP, one can more accurately predict the best binding pose that is closest to the native complex structure. This DeepBSP model shall be very useful in picking out poses close to their natives from many poses generated from a dock application.

show abstract

Section: Introductionmentioning

confidence: 99%

DeepBSP—a Machine Learning Method for Accurate Prediction of Protein–Ligand Docking Structures

Bao

Zhang

2021

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

show abstract

“… 50 Specifically, the binding affinity prediction in metalloenzymes is challenging due to the complex interactions between the ligand, metals, and the protein environment. 15 , 51 As such, extensive efforts have been devoted to improve the prediction accuracy for metallocomplexes. 51 , 52 In terms of these key issues, we wonder if the selection of more suitable features and ML methods could significantly improve the performance and transferability of SFs when compared to the previous results.…”

Section: Introductionmentioning

confidence: 99%

XLPFE: A Simple and Effective Machine Learning Scoring Function for Protein–Ligand Scoring and Ranking

Dong

Wang

2022

ACS Omega

View full text Add to dashboard Cite

Prediction of protein–ligand binding affinities is a central issue in structure-based computer-aided drug design. In recent years, much effort has been devoted to the prediction of the binding affinity in protein–ligand complexes using machine learning (ML). Due to the remarkable ability of ML methods in nonlinear fitting, ML-based scoring functions (SFs) can deliver much improved performance on a selected test set, such as the comparative assessment of scoring functions (CASF), when compared to the classical SFs. However, the performance of ML-based SFs heavily relies on the overall similarity of the training set and the test set. To improve the performance and transferability of an SF, we have tried to combine various features including energy terms from X-score and AutoDock Vina, the properties of ligands, and the statistical sequence-related information from either the binding site or the full protein. In conjunction with extreme trees (ET), an ML model, we have developed XLPFE, a new SF. Compared with other tested methods such as X-score, AutoDock Vina, ΔvinaXGB, PSH-ML, or CNN-score, XLPFE achieves consistently better scoring and ranking power for various types of protein–ligand complex structures beyond the CASF, suggesting that XLPFE has superior transferability. In particular, XLPFE performs better with metalloenzymes. With its faster speed, improved accuracy, and better transferability, XLPFE could be usefully applied to a diverse range of protein–ligand complexes.

show abstract

“…In recent years, hydrogen bond energies have been introduced into modified binding free energy calculations. Hydrogen bond energy usually consists of the distance function and angle function between hydrogen bond donors and hydrogen bond acceptors (Bao et al, 2020 ; Zhao & Huang, 2011 ). Solvent exposed moieties of ligand may also form hydrogen bonds with surrounding residues, but these interaction sites are competed by the solvent, consequently, estimating whether these moieties contribute to protein–ligand binding or given a weight value is necessary, which is commonly characterized by the ratio of solvent accessible surface area (SASA) (Mahmoud et al, 2020 ).…”

Section: Introductionmentioning

confidence: 99%

“…Lewater takes into account both the polar and hydrophobic character of the binding site residues and ligands to predict the possible water molecules' binding points in the structure and makes a reasonable treatment for the sites where hydrogen bonds formed. Additionally, if the hydrogen bonds between receptors and donors are formed within solvent inaccessible regions, nonhydrogen bond penalty is created (Bao et al, 2020 ; Wang et al, 2002 ; Zhao & Huang, 2011 ).…”

Section: Introductionmentioning

confidence: 99%

Hydrogen bonding penalty used for virtual screening to discover potent inhibitors for Papain‐Like cysteine proteases of SARS‐CoV‐2

et al. 2022

View full text Add to dashboard Cite

The Papain‐Like proteases (PLpro) of SARS‐CoV‐2 play a crucial role in viral replication and the formation of nonstructural proteins. To find available inhibitors, the 3D structure of PLpro of SARS2 was obtained by homologous modelling, and we used this structure as a target to search for inhibitors through molecular docking and MM/GBSA binding free energy rescoring. A novel hydrogen bonding penalty was applied to the screening process, which meanwhile took desolvation into account. Finally, 61 compounds were acquired and 4 of them with IC 50 at micromolar level tested in vitro enzyme activity assay, which includes clinical drugs tegaserod. Considering the importance of crystal water molecules, the 4 compounds were re‐docked and considered bound waters in the active site as a part of PLpro. The binding modes of these 4 compounds were further explored with metadynamics simulations. The hits will provide a starting point for future key interactions identified and lead optimization targetting PLpro.

show abstract

Development of a New Scoring Function for Virtual Screening: APBScore

Cited by 11 publications

References 75 publications

DeepBSP—a Machine Learning Method for Accurate Prediction of Protein–Ligand Docking Structures

DeepBSP—a Machine Learning Method for Accurate Prediction of Protein–Ligand Docking Structures

XLPFE: A Simple and Effective Machine Learning Scoring Function for Protein–Ligand Scoring and Ranking

Hydrogen bonding penalty used for virtual screening to discover potent inhibitors for Papain‐Like cysteine proteases of SARS‐CoV‐2

Contact Info

Product

Resources

About