Details

Autor(en) / Beteiligte

• Li, Songling
• Yamashita, Kazuo
• Amada, Karlou Mar
• Standley, Daron M

Titel

Quantifying sequence and structural features of protein-RNA interactions

Ist Teil von

• Nucleic acids research, 2014-09, Vol.42 (15), p.10086-10098

Ort / Verlag

England

Erscheinungsjahr

2014

Quelle

MEDLINE

Beschreibungen/Notizen

• Increasing awareness of the importance of protein-RNA interactions has motivated many approaches to predict residue-level RNA binding sites in proteins based on sequence or structural characteristics. Sequence-based predictors are usually high in sensitivity but low in specificity; conversely structure-based predictors tend to have high specificity, but lower sensitivity. Here we quantified the contribution of both sequence- and structure-based features as indicators of RNA-binding propensity using a machine-learning approach. In order to capture structural information for proteins without a known structure, we used homology modeling to extract the relevant structural features. Several novel and modified features enhanced the accuracy of residue-level RNA-binding propensity beyond what has been reported previously, including by meta-prediction servers. These features include: hidden Markov model-based evolutionary conservation, surface deformations based on the Laplacian norm formalism, and relative solvent accessibility partitioned into backbone and side chain contributions. We constructed a web server called aaRNA that implements the proposed method and demonstrate its use in identifying putative RNA binding sites.