Quantifying sequence and structural features of protein-RNA interactions.
Bottom Line: 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.
Affiliation: Laboratory of Systems Immunology, WPI Immunology Frontier Research Center, Osaka University, Osaka 565-0871, Japan email@example.com.Show MeSH
Mentions: In addition to the benchmark tests presented above, we provide an illustrative example in Figure 5, the Csy4-crRNA complex. In general, sequence-based predictors were more likely to predict charged, polar or aromatic residues on the surface as positive binding sites regardless of their local structural environment. In contrast, due to spatial features introduced here, aaRNA gave more priority to such residues when localized in characteristic RNA binding sites, as learned from the training set. Hence, the aaRNA method could effectively decrease the number of FP predictions, as compared to Figure 5C, E, and F. Importantly, these structural features could facilitate identification of RNA binding sites that are invisible from the point of view of the linear amino acid sequence. As a result, more residues in actual RNA contact could be predicted by aaRNA, and the resulting binding patch appeared more native-like, as illustrated in Figure 5B and C.
Affiliation: Laboratory of Systems Immunology, WPI Immunology Frontier Research Center, Osaka University, Osaka 565-0871, Japan firstname.lastname@example.org.