AqSolDB: Curated Aqueous Solubility (logS) of 9,982 Unique Compounds from Nine Public Datasets

Provider
Murat Cihan Sorkun, Abhishek Khetan, Suleyman Er (AMD group); mirror by Mao Lab, University of Michigan
License
CC0-1.0 (source)
Size
4 files
Format
CSV (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/maomlab/AqSolDB)
Access
Free and public. No account with the data provider is needed.
DOI
10.7910/DVN/OVHAW8
Category
Drug Discovery
Source
https://github.com/mcsorkun/AqSolDB

About this dataset

AqSolDB is a curated reference set of experimental aqueous solubility values for 9,982 unique compounds, merged from nine publicly available solubility datasets by the Autonomous Energy Materials Discovery (AMD) group. Each compound has a name, SMILES, InChI and InChIKey, the solubility as log mol/L (logS), the standard deviation across sources, the number of occurrences, a reliability group from the curation, and RDKit descriptors such as molecular weight, logP, molar refractivity, hydrogen bond donors and acceptors, rotatable bonds and ring counts. This Hugging Face copy from the Mao Lab at the University of Michigan mirrors the official mcsorkun/AqSolDB repository (2019) as a 7,488-row train and 2,494-row test split, with a few SMILES that RDKit could not parse replaced in version 1.1.0. Use it to train and benchmark solubility prediction for drug discovery, environmental fate and materials; load with datasets or pandas. Released under CC0 1.0 on Harvard Dataverse (DOI 10.7910/DVN/OVHAW8). Cite Sorkun, Khetan and Er, AqSolDB, a curated reference set of aqueous solubility and 2D descriptors for a diverse set of compounds, Scientific Data 6, 143 (2019).

How to load this dataset in Python

import pandas as pd
df = pd.read_csv("hf://datasets/maomlab/AqSolDB/AqSolDB/test.csv")
df.head()

Data structure

File types in sampled files: CSV

Files

Example file: AqSolDB/test.csv

Analyze this dataset without downloading it

On Scigantic, AqSolDB: Curated Aqueous Solubility (logS) of 9,982 Unique Compounds from Nine Public Datasets opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Drug Discovery datasets