pQSAR: build massively multitask, two-step machine learning models with unprecedented scope, accuracy, and applicability domain
ZairaChem: Automated ML-based (Q)SAR
MolPipeline:
CNN
conv_qsar_fast: QSAR/QSPR using descriptor-free molecular embedding
Complete Package
SPOC: A tool for calculating spatial and physicochemical descriptors from molecular dynamics simulations.
DeepPurpose: A Deep Learning Library for Compound and Protein Modeling, DTI, Drug Property, PPI, DDI, Protein Function Prediction
Oloren ChemEngine: unified API for the development and use of molecular property predictors
DNN
DeepNeuralNet-QSAR: For training a multi-task DNN with dense QSAR dataset(s)
Ensemble
AIS-Ensemble:
Few-Shot
FewGS: This repository contains source code and datasets for "Few-Shot Graph and SMILES Learning for Molecular Property Prediction."
Few-Shot-Learning-for-Low-Data-Drug-Discovery: .)
FS-Mol: A Few-Shot Learning Dataset of Molecules
Meta-MGNN:
MHNfs: Context-enriched molecule representations improve few-shot drug discovery, available on HuggingFace
MolFesCue: molecular property prediction in Data-Limited and imbalanced contexts using Few-Shot and contrastive learning
PG-DERN:
MolecularGPT:
GAT
LGGA:
GNN
MolE: Molecular representations through redundancy reduction of Embeddings
CheMixNet: A mixed DNN architecture that predicts chemical properties using multiple molecular representations.
DeepDelta: A pairwise deep learning approach predicting property differences between two molecules.
MMGX:
MolCLR: Molecular Contrastive Learning of Representations via Graph Neural Networks
molgraph: Offers graph neural networks with TensorFlow and Keras for molecular machine learning, focusing on compatibility and ease of use.
Graph-Fusion
MolProp:
Hybrid
FGNN:
LLM
MolPMoFiT: transfer learning method based on self-supervised pre-training + task-specific fine-tuning for QSPR/QSAR modeling
MolecularGPT:
MPGNN
Graphormer: deep learning package that allows researchers and developers to train custom models for molecule modeling tasks
Other NN
ChemProp: Features a deep learning approach for molecular property prediction, focusing on scalability and fast uncertainty quantification.
GeminiMol: Incorporates conformational space profile into molecular representation learning, enhancing drug discovery including virtual screening, target identification, and QSAR.
Pretrained Models
OPERA: Open-source QSAR models for pKa prediction using multiple machine learning approaches. Also suite of QSAR models (windows, linux), recent implementation (CATMoS Acute Toxicity Modeling Suite, acute oral toxicity) (standalone).
Reviews
A Survey of Graph Neural Network Architectures in Ligand Binding Affinity Prediction Models:
Several architectures
DeltaClassifiers: a novel molecular pairing approach to process this data. This creates a new classification task of predicting which one of two paired molecules is more potent.
Transfer Learning
PGM:
Transformer
KnoMol:
ChemBERTa: BERT-like models applied to chemical SMILES data for drug design, chemical modelling, and property prediction
GPT-MolBERTa: A text-based molecular property prediction model utilizing a novel approach to represent SMILES molecules.
PointGAT:
TOML-BERT:
X-MOL: large-scale pre-training for molecular understanding and diverse molecular analysis