Scoring Functions

Carbohydrates

• VinaCarb: Carbohydrate specific scoring function
  

ElectroStatic Potential-based

• ECScore: Modified Electrostatic Complementary Score Function
  

Electron Density

• ExptGMS: Experimental Electron Density based Grid Matching Score
  

Empirical

• AA-Score: An empirical protein-ligand scoring function focusing on amino acid-specific interaction components for improved virtual screening and lead optimization.
  
• Cyscore: An empirical scoring function for accurate protein-ligand binding affinty prediction (linux command line) (standalone).
  
• LinF9: Presents Lin_F9, a linear empirical scoring function for protein-ligand docking, available within a fork of the Smina docking suite.
  
• Vinardo: Vinardo is a scoring function based on Autodock Vina that improves scoring, docking, and virtual screening capabilities. It was trained through a combination of scoring, minimization, and re-docking on curated datasets for optimum docking performance. Vinardo is available within Smina, a fork of Vina.
  

For predicted structures

• DENOISer: Physics-inspired accuracy estimator for model-docked ligand complexes
  

Hybrid

• SQM-ML: Hybrid Semiempirical Quantum Chemical - Machine Learning (SQM-ML) scoring function for protein-ligand interactions
  

Hydration

• HydraMapSF: Contains content related to HydraMap-based features generation, but specific details about its functionality or applications were not provided.
  
• WatSite3:
  

Knowledge-based

• Convex-PL: Convex-PL is a knowledge-based scoring function for protein-ligand interactions that augments Convex-PL with conformational flexibility for better performance in affinity prediction and virtual screening tasks.
  
• DLIGAND2: DLIGAND2 is a knowledge-based method for predicting protein-ligand binding affinity, utilizing a distance-scaled, finite, ideal-gas reference state.
  
• ITScoreAff: iterative knowledge-based scoring function for protein-ligand interactions by considering binding affinity information
  
• KORP-PL: KORP-PL is a novel coarse-grained knowledge-based scoring function for protein-ligand interactions, focusing on the relative orientation of a ligand molecule to a protein residue.
  

Machine-learning scoring functions

• AEScore: Learning protein-ligand binding affinity using atomic environment vectors.
  
• AKScore: This model uses an ensemble of multiple independently trained 3-D convolutional neural networks to predict protein-ligand complex binding affinity, showing high correlation with experimental data.
  
• APBScore: Atom Pair Based scoring function.
  
• CAPLA: Improved prediction of protein-ligand binding affinity by a deep learning approach based on a cross-attention mechanism.
  
• Censible: Predicts small-molecule binding affinities using deep-learning context explanation networks for interpretable output.
  
• CNN: Files for repeating a study on convolutional neural networks and atomic contact features for binding affinity prediction.
  
• ConBAP: Protein-Ligand Binding Affinity Prediction with Multimodal Contrastive Learning
  
• DAAP: Distance plus attention for binding affinity prediction
  
• DeepAffinity: Protein-compound affinity prediction through unified RNN-CNN.
  
• DeepBindGCN: Predicts protein ligand binding affinity non-dependent on docking conformation.
  
• DeepBindRG: A deep learning based method for estimating effective protein-ligand affinity.
  
• DeepGLSTM: A model predicting binding affinity values between FDA-approved drugs and viral proteins of SARS-CoV-2.
  
• ECIF: Extended Connectivity Interaction Features for molecular analysis.
  
• EGGNet: Source code for EGGNet, a framework for protein complex pose scoring.
  
• ENS-Score: An ensemble‐based approach to estimate confidence of predicted protein–ligand binding affinity values
  
• ENS-Score GUI:
  
• EquiScore: A scoring function for virtual screening and eval with interpretable output.
  
• ET-Score: A scoring function based on Extra Trees algorithm for predicting ligand-protein binding affinity.
  
• FABind: A fast and accurate protein-ligand binding model for NeurIPS 2023.
  
• FAST: Fusion models for Atomic and molecular STructures aiming at predicting accurate protein-ligand binding affinity.
  
• FGNN: Predicts ligand binding affinity with graph neural networks and 3D structure-based complex graph.
  
• FitScore: FitScore: a fast machine learning-based score for 3D virtual screening enrichment
  
• FitScore: FitScore: A Fast Machine Learning-Based Score for 3D Virtual Screening Enrichment
  
• FusionDTA: A framework focusing on protein-ligand binding affinity prediction with deep learning techniques.
  
• GBScore: A scoring function based on Gradient Boosting Trees algorithm for predicting ligand-protein binding affinity.
  
• GenScore: A scoring framework for predicting protein-ligand binding affinity.
  
• GGL-ETA-Score: This code computes features for both the SYBYL-GGL and ECIF-GGL models.
  
• GIGN: Contains GNN-based models for protein-ligand binding affinity prediction.
  
• GraphBAR: deep-learning-based prediction model based on a graph convolutional neural network.
  
• graphLambda: A deep learning model to score protein-ligand binding affinity using PyTorch and PyTorch Geometric, focusing on the application of graphLambda model for training and result replication.
  
• GraphscoreDTA: A graph neural network for protein-ligand binding affinity prediction.
  
• HAC-Net: A Hybrid Attention-Based Convolutional Neural Network for Highly Accurate Protein-Ligand Binding Affinity Prediction.
  
• HaPPy: Utilizes multi-perspective graphs for protein-ligand binding affinity prediction.
  
• HGScore: A Heterogeneous Graph Convolutional Neural Network to score a protein-ligand complex.
  
• HydraScreen: utilizes a state-of-the-art 3D convolutional neural network designed for the effective representation of molecular structures and interactions in protein–ligand binding
  
• KDeep: protein-ligand affinity predictor based on DCNNs (Deep Convolutional Neural Networks).
  
• KIDA: Process for extracting protein pockets and generating training data for KIDA.
  
• Lee2023a: Meta-modeling of ligand-protein binding affinity.
  
• MBI-Score: Utilizes graph representation for protein-ligand interfacial atoms in native complexes for scoring.
  
• MGraphDTA: Implements a high-efficiency concordance index for performance evaluation of drug-target affinity prediction.
  
• Milbinding: binary binding prediction
  
• ML-PLIC: ML-PLIC: a web platform for characterizing protein–ligand interactions and developing machine learning-based scoring functions
  
• MSECIFv2: Improved version of extended connectivity interaction features (ECIF) by enabling to take the atomic distance into account.
  
• OctSurf: Efficient hierarchical voxel-based molecular surface representation for protein-ligand affinity prediction (scoring) (standalone).
  
• OnionNet: A multiple-layer inter-molecular contact based deep neural network for protein-ligand binding affinity prediction.
  
• OnionNet-2: OnionNet-2 is constructed based on convolutional neural network (CNN) to predict the protein-ligand binding affinity.
  
• OnionNet-SFCT: Incorporates a scoring function correction term for improved docking and screening accuracies.
  
• PAMNet: Official implementation of PAMNet in the paper "A universal framework for accurate and efficient geometric deep learning of molecular systems" accepted by Nature Scientific Reports. PAMNet improves upon MXMNet and achieves high performance in tasks like small molecule property prediction, RNA 3D structure prediction, and protein-ligand binding affinity prediction with both accuracy and efficiency.
  
• PIGNet: physics-informed deep learning model toward generalized drug–target interaction predictions
  
• PIGNet2: A versatile deep learning-based model for protein-ligand interaction prediction.
  
• PLANET: A graph neural network model for predicting protein-ligand binding affinity and virtual screening.
  
• PLAPT: A state-of-the-art protein-ligand binding affinity predictor leveraging transfer learning from pre-trained transformers to predict binding affinities with high accuracy.
  
• ProSmith: A multimodal transformer network for protein-small molecule interactions.
  
• ProtMD: Pre-training Protein Geometric Models via Molecular Dynamics Trajectories
  
• QDL-DTA: A hybrid quantum-classical deep learning algorithm for protein-ligand binding affinity prediction.
  
• RFL-Score: RFL-Score: Random Forest with Lasso Scoring Function for Protein-Ligand Molecular Docking
  
• RoseNet: Predicts absolute binding affinity using molecular mechanics energies.
  
• RTMScore: A novel scoring function based on residue-atom distance likelihood potential and graph transformer for predicting protein-ligand interactions efficiently.
  
• SadNet: SadNet: a novel multimodal fusion network for protein–ligand binding affinity prediction
  
• SCORCH: Houses deployable code for the SCORCH scoring function and docking pipeline.
  
• SIEVE-Score: A virtual screening method based on random forest using interaction energy-based scoring.
  
• SMPLIP-Score: Substructural Molecular and Protein–Ligand Interaction Pattern Score, a direct interpretable predictor of absolute binding affinity.
  
• SSnet: Secondary Structure based End-to-End Learning model for Protein-Ligand Interaction Prediction
  
• StackCPA: A stacking model for compound-protein binding affinity prediction based on pocket multi-scale features (scoring - 2023) (standalone).
  
• TankBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction
  
• TB-IEC-Score: Meta-modeling of ligand-protein binding affinity.
  
• TrustAffinity: A sequence-based deep learning framework addressing challenges in protein-ligand binding affinity prediction including out-of-distribution generalizations, uncertainty quantification, and scalability.
  
• XLPFE: a Simple and Effective Machine Learning Scoring Function for Protein-ligand Scoring and Ranking (standalone).
  

Metalloproteins Specific

• MetalProGNet: MetalProGNet is a structure-based deep graph model specifically designed for metalloprotein-ligand interaction predictions, developed based on the IGN framework.
  

Negative Image based

• O-LAP: Building shape-focused pharmacophore models for effective docking screening
  
• PANTHER: Negative Image based docking and scoring
  

Protein-ligand interaction fingerprints

• ECIFGraph: Introduces a water network-augmented two-state model for protein-ligand binding affinity prediction, incorporating extended connectivity interaction features and graph transformer operators.
  

Quantum-mechanical

• SQM2.20: Semiempirical quantum-mechanical scoring function yields DFT-quality protein–ligand binding affinity predictions in minutes
  

RNA-specific

• RLAffinity: Contrastive pre-training and 3D convolution neural network for RNA and small molecule binding affinity prediction
  
• SPRank: Knowledge-Based Scoring Function for RNA-Ligand Pose Prediction and Virtual Screening
  

Relative Binding Affinity

• PBCNet: Webserver for relative binding affinity calculation.
  

SF Optimisation

• SANDRES: Statistical Analysis of Docking Results and Scoring functions
  
• SANDRES2.0: Statistical Analysis of Docking Results and Scoring functions
  
• SFSXplorer: Computational tool to explore the scoring function space (standalone).