Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development (2021-02-18T00:00:00.000000Z)

TL;DR

Therapeutics Data Commons is introduced, the first unifying platform to systematically access and evaluate machine learning across the entire range of therapeutics, and it is envisioned that TDC can facilitate algorithmic and scientific advances and considerably accelerate machine-learning model development, validation and transition into biomedical and clinical implementation.

Abstract

Therapeutics machine learning is an emerging field with incredible opportunities for innovatiaon and impact. However, advancement in this field requires formulation of meaningful learning tasks and careful curation of datasets. Here, we introduce Therapeutics Data Commons (TDC), the first unifying platform to systematically access and evaluate machine learning across the entire range of therapeutics. To date, TDC includes 66 AI-ready datasets spread across 22 learning tasks and spanning the discovery and development of safe and effective medicines. TDC also provides an ecosystem of tools and community resources, including 33 data functions and types of meaningful data splits, 23 strategies for systematic model evaluation, 17 molecule generation oracles, and 29 public leaderboards. All resources are integrated and accessible via an open Python library. We carry out extensive experiments on selected datasets, demonstrating that even the strongest algorithms fall short of solving key therapeutics challenges, including real dataset distributional shifts, multi-scale modeling of heterogeneous data, and robust generalization to novel data points. We envision that TDC can facilitate algorithmic and scientific advances and considerably accelerate machine-learning model development, validation and transition into biomedical and clinical implementation. TDC is an open-science initiative available at https://tdcommons.ai.

Authors

M. Zitnik

7 papers

J. Leskovec

21 papers

Kexin Huang

2 papers

TL;DR

Abstract

Authors

References181 items

MARS: Markov Molecular Sampling for Multi-objective Drug Discovery

Using a genetic algorithm to find molecules with good docking scores

Barking up the right tree: an approach to search over molecule synthesis DAGs

WILDS: A Benchmark of in-the-Wild Distribution Shifts

Accelerating high-throughput virtual screening through molecular pool-based active learning

DeepPurpose: a deep learning library for drug–target interaction prediction

RetroGNN: Approximating Retrosynthesis by Graph Neural Networks for De Novo Drug Design

Learning explanations that are hard to vary

Pushing the boundaries of molecular representation for drug discovery with graph attention mechanism.

Prediction of chemical reaction yields using deep learning

OpenChem: A Deep Learning Toolkit for Computational Chemistry and Drug Design

Deep Learning in Protein Structural Modeling and Design

In Search of Lost Domain Generalization

Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits

Predicting Antibody Developability from Sequence using Machine Learning

Towards FAIR principles for research software

Open Graph Benchmark: Datasets for Machine Learning on Graphs

SkipGNN: predicting molecular interactions with skip-graph networks

Learning To Navigate The Synthetically Accessible Chemical Space Using Reinforcement Learning

DeepPurpose: a Deep Learning Library for Drug-Target Interaction Prediction and Applications to Repurposing and Screening

Network medicine framework for identifying drug-repurposing opportunities for COVID-19

In vitro screening of a FDA approved chemical library reveals potential inhibitors of SARS-CoV-2 replication

A Deep Learning Approach to Antibiotic Discovery

Autonomous discovery in the chemical sciences part II: Outlook

A reference map of the human binary protein interactome

The Synthesizability of Molecules Proposed by Generative Models

Invariant Risk Minimization Games

Multi-Objective Molecule Generation using Interpretable Substructures

Predicting Retrosynthetic Reactions Using Self-Corrected Transformer Neural Networks

Deciphering interaction fingerprints from protein molecular surfaces using geometric deep learning

Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization

Therapeutic target database 2020: enriched resource for facilitating research and early development of targeted therapeutics

The DisGeNET knowledge platform for disease genomics: 2019 update

AqSolDB, a curated reference set of aqueous solubility and 2D descriptors for a diverse set of compounds

A robotic platform for flow synthesis of organic compounds informed by AI planning

ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations

Large dataset enables prediction of repair after CRISPR–Cas9 editing in primary T cells

ChemML: A machine learning and informatics program package for the analysis, mining, and modeling of chemical and materials data

Evaluating Protein Transfer Learning with TAPE

Self-referencing embedded strings (SELFIES): A 100% robust molecular string representation

Strategies for Pre-training Graph Neural Networks

The Potential for microRNA Therapeutics and Clinical Research

SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems

Analyzing Learned Molecular Representations for Property Prediction

Personalized Medicine and the Power of Electronic Health Records

DrugComb - an integrative cancer drug combination data portal

Five computational developability guidelines for therapeutic antibody profiling

Prediction of higher-selectivity catalysts by computer-driven workflow and machine learning

Ultra-large library docking for discovering new chemotypes

Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models

GuacaMol: Benchmarking Models for De Novo Molecular Design

miRBase: from microRNA sequences to function

ChEMBL: towards direct deposition of bioassay data

Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction

The Immune Epitope Database (IEDB): 2018 update

A graph-based genetic algorithm and generative model/Monte Carlo tree search for the exploration of chemical space

Optimization of Molecules via Deep Reinforcement Learning

Drug repurposing: progress, challenges and recommendations

How Powerful are Graph Neural Networks?

Learning Multimodal Graph-to-Graph Translation for Molecular Optimization

Parapred: antibody paratope prediction using convolutional and recurrent neural networks

Improved methods for predicting peptide binding affinity to MHC class II molecules

Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation

Domain Generalization with Adversarial Feature Learning

Predicting reaction performance in C–N cross-coupling using machine learning

Fréchet ChemNet Distance: A Metric for Generative Models for Molecules in Drug Discovery

Integrative omics for health and disease

Modeling polypharmacy side effects with graph convolutional networks

DeepDTA: deep drug–target binding affinity prediction

Generating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks

DeepSynergy: predicting anti-cancer drug synergy with Deep Learning

Domain Generalization by Marginal Transfer Learning

DrugBank 5.0: a major update to the DrugBank database for 2018

miRTarBase update 2018: a resource for experimentally validated microRNA-target interactions

Large-Scale Analysis of Disease Pathways in the Human Interactome

Predicting Organic Reaction Outcomes with Weisfeiler-Lehman Network