ARIEL

Overview

ARIEL (Affinity Resource for Interactions of RNA & Ligands) is an openly accessible, manually curated database that links experimentally measured RNA–ligand dissociation constants (K_d) to the high-resolution three-dimensional structures of the same complexes. By pairing validated binding affinities with experimental structures, ARIEL brings together two kinds of information that are usually held apart — affinity values scattered across the primary literature, and structures deposited in the Protein Data Bank — in a single, consistently curated resource for RNA-targeted drug discovery and computational modeling.

RNA has become an important target for small-molecule drugs, offering therapeutic routes to bacterial, viral, and neurological diseases that are difficult to address through protein-targeted approaches. Experimentally measured RNA–ligand binding affinities, however, are scattered across the primary literature and are rarely linked to the three-dimensional structures that show how binding occurs. This fragmentation has limited the structure-based study of RNA–small-molecule recognition and the computational models built on it. ARIEL was created to bring affinity and structure together in one openly accessible, consistently curated resource.

Records are presented through searchable tables, affinity color-encoding, downloadable datasets, statistical summaries, and 3D structure viewers — allowing users to explore RNA–ligand binding in both human-readable and machine-parsable formats.

Key Features

Database Construction

ARIEL was assembled through a systematic, manually verified curation pipeline applied to all RNA-containing structures in the Protein Data Bank (accessed through the Nucleic Acid Knowledge Base, NAKB) up to the 10th of July 2026:

We began with 6,954 RNA–ligand-containing structures determined by X-ray crystallography, cryo-EM, or NMR. For X-ray and cryo-EM structures, only those with a resolution of 3.5 Å or better were included, while no resolution cutoff was applied to NMR structures. Associated ligands were identified through automated screening, capturing a broad range of species including water (HOH), metal ions, and organic ligands. Solvent molecules (including HOH) and simple salts were subsequently excluded. Structures containing only ligands from a predefined blacklist (primarily charged ions) were also removed. However, monoatomic ions that act as bona fide cognate ligands — such as fluoride or Mn²⁺ in metal-sensing riboswitches — were retained. After this filtering, the dataset comprised approximately 2,843 RNA–ligand complex structures.

For each structure, the primary literature was screened manually for an experimentally measured dissociation constant. Only true K_d values were retained; IC₅₀ and K_i values were excluded. Each value was recorded with its measurement method, buffer, ionic conditions, pH, and temperature where reported, and with explicit provenance distinguishing the structure’s primary citation from the publication reporting the affinity.

Redundancy was controlled at the level of the biological complex. Where several depositions represent the same RNA–ligand pair, the highest-resolution structure (then occupancy nearest 1, then lower pocket B-factor) was selected as the primary record and the others retained as related entries; where a single structure contained multiple crystallographic copies of the complex, one representative copy was kept and the remainder recorded as alternate copies (see the Help page for more details).

This pipeline produced 278 non-redundant RNA–ligand complexes.

Content & Coverage

ARIEL contains 278 non-redundant RNA–ligand complexes spanning riboswitches, ribosomal RNA and intact ribosomes, aptamers, ribozymes, and other structured RNAs. The entries represent a wide range of ligand chemistries and binding affinities spanning several orders of magnitude (pK_d from 2.14 to 10.96, across 179 distinct ligands).

Updates & Versioning

ARIEL is updated timely. Each cycle applies the identical curation pipeline to newly deposited RNA–ligand structures and newly published affinity measurements. Releases are versioned and archived so that analyses performed on any prior version remain fully reproducible. Where a new, higher-resolution determination of an existing pair appears, the primary record is updated and the earlier structures are retained as related entries.

Availability and License

ARIEL is freely accessible to all users, including commercial users. Detailed documentation and usage instructions are available in the Help section of the website.