Atomistic Description

MoleculeNet is a large scale benchmark for molecular machine learning. MoleculeNet curates multiple public datasets, establishes metrics for evaluation, and offers high quality open-source implementations of multiple previously proposed molecular featurization and learning algorithms (released as part of the DeepChem open source library). MoleculeNet benchmarks demonstrate that learnable representations are powerful tools for molecular machine learning and broadly offer the best performance.

The Materials Project is a collection of chemical compounds labelled with different attributes. The labelling is performed by different simulations, most of them at DFT level of theory. The dataset links: * MP 2018.6.1 (69,239 materials) * MP 2019.4.1 (133,420 materials)

Energies and forces for molecular dynamics trajectories of eight organic molecules. Level of theory DFT: PBE+vdW-TS.

QM9 provides quantum chemical properties (at DFT level) for a relevant, consistent, and comprehensive chemical space of small organic molecules. This database may serve the benchmarking of existing methods, development of new methods, such as hybrid quantum mechanics/machine learning, and systematic identification of structure-property relationships. Source: QM9 Dataset Image Source: https://pubs.acs.org/doi/pdf/10.1021/ci300415d

The Tox21 data set comprises 12,060 training samples and 647 test samples that represent chemical compounds. There are 801 "dense features" that represent chemical descriptors, such as molecular weight, solubility or surface area, and 272,776 "sparse features" that represent chemical substructures (ECFP10, DFS6, DFS8; stored in Matrix Market Format ). Machine learning methods can either use sparse or dense data or combine them. For each sample there are 12 binary labels that represent the outcome (active/inactive) of 12 different toxicological experiments. Note that the label matrix contains many missing values (NAs). The original data source and Tox21 challenge site is https://tripod.nih.gov/tox21/challenge/. Source: Tox21 Machine Learning Data Set Image Source: https://www.frontiersin.org/articles/10.3389/fenvs.2015.00080/full

The BBBP dataset comes from a study focused on modeling and predicting the permeability of the blood-brain barrier. The BBBP dataset contains binary labels indicating whether a compound can penetrate the blood-brain barrier (BBB) or not. Researchers use this dataset to develop and evaluate machine learning methods for predicting BBB permeability. It’s a critical task because understanding which compounds can cross the BBB is essential for drug discovery and designing therapeutics for neurological conditions. Check https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8708321/

ESOL is a water solubility prediction dataset consisting of 1128 samples.

The FreeSolv database offers a curated collection of experimental and calculated hydration-free energies for small molecules in water. It includes both experimental values obtained from prior literature and calculated values based on simulations. The goal is to provide accurate hydration-free energy data, which is essential for understanding solvation properties and interactions of molecules in aqueous environments.

QM7 dataset is a subset of the GDB-13 database. GDB-13 contains nearly 1 billion stable and synthetically accessible organic molecules. In the QM7 subset, only molecules with up to 23 atoms are included. These atoms consist of carbon ©, nitrogen (N), oxygen (O), and sulfur (S). The total number of molecules in the QM7 dataset is 7165. Each molecule is represented using the Coulomb matrix, which captures the interactions between atoms.

The BACE dataset focuses on inhibitors of human beta-secretase 1 (BACE-1). It includes both quantitative (IC50 values) and qualitative (binary labels) binding results. The dataset comprises small molecule inhibitors across a wide range of affinities, spanning three orders of magnitude (from nanomolar to micromolar IC50 values). Specifically, it provides: 154 BACE inhibitors for affinity prediction. 20 BACE inhibitors for pose prediction. 34 BACE inhibitors for free energy prediction.

JARVIS-DFT is a repository of density functional theory based calculation data for materials.

The HIV dataset was introduced by the Drug Therapeutics Program (DTP) AIDS Antiviral Screen, which tested the ability to inhibit HIV replication for over 40,000 compounds. Screening results were evaluated and placed into three categories: confirmed inactive (CI), confirmed active (CA), and confirmed moderately active (CM). Check https://link.springer.com/chapter/10.1007/978-3-540-89689-0_33

Using conservation of energy -- a fundamental property of closed classical and quantum mechanical systems -- we develop an efficient gradient-domain machine learning (GDML) approach to construct accurate molecular force fields using a restricted number of samples from ab initio molecular dynamics (AIMD) trajectories. The GDML implementation is able to reproduce global potential energy surfaces of intermediate-sized molecules with an accuracy of 0.3 kcal $\text{mol}^{-1}$ for energies and 1 kcal $\text{mol}^{-1}$ $\text{\AA}^{-1}$ for atomic forces using only 1000 conformational geometries for training. We demonstrate this accuracy for AIMD trajectories of molecules, including benzene, toluene, naphthalene, ethanol, uracil, and aspirin. The challenge of constructing conservative force fields is accomplished in our work by learning in a Hilbert space of vector-valued functions that obey the law of energy conservation. The GDML approach enables quantitative molecular dynamics simulations for molecules at a fraction of cost of explicit AIMD calculations, thereby allowing the construction of efficient force fields with the accuracy and transferability of high-level ab initio methods.

The ClinTox dataset compares drugs approved by the FDA and drugs that have failed clinical trials for toxicity reasons. The dataset includes two classification tasks for 1491 drug compounds with known chemical structures: (1) clinical trial toxicity (or absence of toxicity) and (2) FDA approval status.

SIDER contains information on marketed medicines and their recorded adverse drug reactions. The information is extracted from public documents and package inserts. The available information include side effect frequency, drug and side effect classifications as well as links to further information, for example drug–target relations. Source: Side Effect Resource Image Source: http://sideeffects.embl.de/drugs/2756/

Molecule3D is a new benchmark that includes a dataset with precise ground-state geometries of approximately 4 million molecules derived from density functional theory (DFT). It also provides a set of software tools for data processing, splitting, training, and evaluation, etc.

ToxCast is an initiative by the U.S. Environmental Protection Agency (EPA) aimed at predicting the potential toxicity of various chemical compounds. It involves high-throughput screening assays that evaluate thousands of chemicals across multiple biological endpoints. These endpoints cover a wide range of effects, including cell cycle disruptions, interactions with steroid receptors, and cytotoxicity.

The Maximum Unbiased Validation (MUV) dataset is a benchmark dataset selected from PubChem BioAssay. It was created by applying a refined nearest-neighbor analysis. The MUV dataset is specifically designed for the validation of virtual screening techniques.

The Matbench test suite v0.1 contains 13 supervised ML tasks from 10 datasets. Matbench’s data are sourced from various subdisciplines of materials science, such as experimental mechanical properties (alloy strength), computed elastic properties, computed and experimental electronic properties, optical and phonon properties, and thermodynamic stabilities for crystals, 2D materials, and disordered metals. The number of samples in each task ranges from 312 to 132,752, representing both relatively scarce experimental materials properties and comparatively abundant properties such as DFT-GGA formation energies. Each task is a self-contained dataset containing a single material primitive as input (either composition or composition plus crystal structure) and target property as output for each sample.

QM8 dataset is a collection of molecular data used for studying quantum mechanical calculations of electronic spectra and excited state energy of small molecules. The QM8 dataset consists of approximately 7,165 molecules. These molecules are a subset of the GDB-13 database, which contains nearly 1 billion stable and synthetically accessible organic molecules. The subset includes all molecules with up to 23 atoms, including 7 heavy atoms (C, N, O, and S).

Benchmark for de novo molecular design

DrivAerNet is a large-scale, high-fidelity CFD dataset of 3D industry-standard car shapes designed for data-driven aerodynamic design. It comprises 4000 high-quality 3D car meshes and their corresponding aerodynamic performance coefficients, alongside full 3D flow field information. It includes: - CFD Simulation Data: The raw dataset, including full 3D pressure, velocity fields, and wall-shear stresses, computed using 8-16 million mesh elements has a total size of $\sim$ 16TB. - Curated CFD Simulations: For ease of access and use, a streamlined version of the CFD simulation data is provided, refined to include key insights and data, reducing the size to $\sim$ 1TB. - 3D Car Meshes: A total of 4000 designs, showcasing a variety of conventional car shapes and emphasizing the impact of minor geometric modifications on aerodynamic efficiency. The 3D meshes and aerodynamic coefficients $\sim$ 84GB. - 2D slices include the car's wake in the $x$-direction and the symmetry plane in the $y$-direction $\sim$ 12GB.

PCBA dataset 11 is a collection of high-quality dose-response data, formulated as a multitask learning benchmark from 128 high-throughput screening (HTS) assays.

The OQMD is a database of DFT calculated thermodynamic and structural properties of one million materials, created in Chris Wolverton's group at Northwestern University. The OQMD v1.2 dataset for CGNN is downloadable from this link, which contains 561,888 materials. Its format is described in here. The original data is available at the OQMD website.

A benchmark for molecular machine learning where improvements in model performance can be immediately observed in the throughput of promising molecules synthesized in the lab. Photoswitches are a versatile class of molecule for medical and renewable energy applications where a molecule's efficacy is governed by its electronic transition wavelengths. Source: The Photoswitch Dataset: A Molecular Machine Learning Benchmark for the Advancement of Synthetic Chemistry

This is a large-scale dataset of quantum-mechanically calculated properties (DFT level) of crystalline materials for graph representation learning that contains approximately 900k entries (OQM9HK). This dataset is constructed on the basis of the Open Quantum Materials Database (OQMD) v1.5 containing more than one million entries, and is the successor to the OQMD v1.2 dataset containing approximately 600k entries (OQM6HK). * Technical Report * CGNN v1.1 Histograms of materials properties

This dataset is a multi-labelled SMILES odor dataset with 138 odor descriptors. This dataset was created for replicating the paper: A principal odor map unifies diverse tasks in olfactory perception. The complete replication of the paper (dataset curation + model) can be found in the OpenPOM GitHub repository. The dataset contains 4983 molecules, each described by multiple odor labels (e.g. creamy, grassy), was made by combining the GoodScents and Leffingwell PMP 2001 datasets each containing odorant molecules and corresponding odor descriptors.

A.2.1 AN OPEN, LARGE-SCALE DATASET FOR ZERO-SHOT DRUG DISCOVERY DERIVED FROM PUBCHEM We constructed a large public dataset extracted from PubChem (Kim et al., 2019; Preuer et al., 2018), an open chemistry database, and the largest collection of readily available chemical data. We take assays ranging from 2004 to 2018-05. It initially comprises 224,290,250 records of molecule-bioassay activity, corresponding to 2,120,854 unique molecules and 21,003 unique bioassays. We find that some molecule-bioassay pairs have multiple activity records, which may not all agree. We reduce every molecule-bioassay pair to exactly one activity measurement by applying majority voting. Molecule-bioassay pairs with ties are discarded. This step yields our final bioactivity dataset, which features 223,219,241 records of molecule-bioassay activity, corresponding to 2,120,811 unique molecules and 21,002 unique bioassays ranging from AID 1 to AID 1259411. Molecules range up to CID 132472079. The dataset has 3 different splitting schemes.

We propose an efficient high-throughput scheme for the discovery of stable crystalline phases. Our approach is based on the transmutation of known compounds, through the substitution of atoms in the crystal structure with chemically similar ones. The concept of similarity is defined quantitatively using a measure of chemical replaceability, extracted by data-mining experimental databases. In this way we build 189,981 possible crystal phases, including 18,479 that are on the convex hull of stability. The resulting success rate of 9.72% is at least one order of magnitude better than the usual success rate of systematic high-throughput calculations for a specific family of materials, and comparable with speed-up factors of machine learning filtering procedures. As a characterization of the set of 18,479 stable compounds, we calculate their electronic band gaps, magnetic moments, and hardness. Our approach, that can be used as a filter on top of any high-throughput scheme, enables us to efficiently extract stable compounds from tremendously large initial sets, without any initial assumption on their crystal structures or chemical compositions.