This repository contains Table 1 for the review Biologically-driven generative chemistry: using biological data to guide de novo drug design.
By Philip John Harrison and Rocío Mercado.
View the searchable/filterable table here:
The source CSV is available here:
- The model names enclosed in square brackets are not the official names, but those that we used in our review paper for convenience.
- SE(3)-eq. stands for SE(3)-equivariance. For 3D molecular models, E(3)-equivariance enforces rotation, translation and reflection symmetries, whereas SE(3)-equivariance does not enforce reflection symmetry, thus enabling enantiomer compounds to be differentiated.
| Acronym | Meaning |
|---|---|
| AAE | Adversarial autoencoder |
| AL | Active learning |
| BERT | Bidirectional encoder representations from transformers |
| CNN | Convolutional neural network |
| DGM | Deep generative model |
| EC50 | Half-maximal effective concentration |
| ECFP | Extended connectivity fingerprint |
| GAN | Generative adversarial network |
| GCN | Graph convolutional network |
| GEP | Gene expression profile |
| GNN | Graph neural network |
| GRU | Gated recurrent unit network |
| GVP | Geometric vector perceptron |
| IC50 | Half-maximal inhibitory concentration |
| JT-VAE | Junction tree variational autoencoder |
| LSTM | Long short-term memory network |
| MCMC | Markov chain Monte Carlo |
| MCTS | Monte Carlo tree search |
| MPNN | Message-passing neural network |
| NCI | Non-covalent interaction |
| PLC | Protein-ligand complex |
| QED | Quantitative estimate of drug-likeness |
| QSAR | Quantitative structure-activity relationship |
| RL | Reinforcement learning |
| RNN | Recurrent neural network |
| SA | Synthetic accessibility (score) |
| SDE | Stochastic differential equation |
| SELFIES | Self-referencing embedded strings |
| SMILES | Simplified molecular input line entry system |
| VAE | Variational autoencoder |
| VN | Vector neuron |