Search for a command to run...
<ns3:p>The comparison of biomolecular 3D structures is essential to understand functional relationships, evolutionary conservation, and molecular recognition mechanisms. While a rich set of methods has been developed for proteins, the comparison of RNA 3D structures remains less explored, and few approaches are able to handle both RNAs and RNA–protein complexes within a unified framework.</ns3:p> <ns3:p/> <ns3:p>We introduce STAlign, a proximity-based computational method that compares RNA and protein 3D structures using spatial information. Each structure is represented as an Arc-Annotated Sequence (AAS), where arcs connect pairs of residues that are spatially closer than a given threshold. In addition to purely geometric proximity, AAS representations can also be derived from known chemical or base-pairing interactions extracted by dedicated annotation tools. This representation is then processed through a regular tree grammar, which uniquely associates each structure with an algebraic tree. By abstracting the algebraic tree into a structural tree, we define a new and computationally efficient Algebraic Structural Alignment (ASA) distance based on classical tree alignment.</ns3:p> <ns3:p/> <ns3:p>Unlike conventional atom-level metrics such as RMSD, ASA captures the topological organization of spatial proximities, naturally handling cases where structures have different sizes or missing residues. Moreover, since it operates at a symbolic level, it can efficiently detect both global and local rearrangements, such as helical crossings or triple-helix formation.</ns3:p> <ns3:p/> <ns3:p>The ASA distance quantifies the structural dissimilarity between molecules, thus enabling cross-comparison among RNAs, proteins, and complexes. We implemented this framework in the open-source Java tool STAlign (https://github.com/bdslab/STAlign), which automatically converts PDB entries into AASs and computes ASA distances. We demonstrate that STAlign can capture subtle spatial variations between RNA molecules that show similar secondary structures (tRNA, riboswitches), as well as protein structural similarities among subsets of tandem repeat proteins (TRPs) from the REPEATS DB. In the latter case, the ASA-based comparison outperforms the TM-align tool in clustering proteins according to their Class and Topology. Future developments will include large-scale benchmarking on RNA–protein complexes and the integration of ASA into pipelines for structural clustering, classification, and evolutionary analysis. By unifying RNA and protein comparison under a single algebraic framework, STAlign contributes to bridging the gap between formal modeling and practical structural bioinformatics.</ns3:p>