AqSolDB: Aqueous Solubility Data for 9,982 Unique Compounds

AqSolDB contains curated experimental aqueous solubility values for 9,982 unique compounds, aggregated from 9 publicly available sources. The dataset was created by the Autonomous Energy Materials Discovery research group and includes calculated topological and physico-chemical 2D descriptors alongside validated molecular representations. It is described as the largest dataset of its kind for training data-driven models.

Use Cases

• Training machine learning models to predict aqueous solubility based on molecular descriptors.
• Benchmarking new solubility prediction algorithms against a large, curated reference dataset.
• Studying the relationship between molecular structure and solubility using calculated 2D descriptors.
• Validating computational chemistry methods against a broad set of experimental solubility measurements.

Strengths

• Contains 9,982 unique compounds, described as the largest dataset of its kind.
• Data is curated from 9 different publicly available aqueous solubility datasets, suggesting consolidation of sources.
• Includes calculated topological and physico-chemical 2D descriptors in addition to experimental values.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• Row count is unknown, which may limit suitability assessment for specific modeling tasks.

Provenance

Source

Aggregated from 9 publicly available datasets including eChemPortal and sources cited in the Nature Scientific Data paper.

Collection Method

Curated by the Autonomous Energy Materials Discovery (AMD) research group.