RBind      Zhao group at Central China Normal University

RBind: computational network method to predict RNA binding sites

 

Non-coding RNA molecules play essential roles by interacting with other molecules to perform various biological functions. However, it is difficult to determine RNA structures due to their flexibility. At present, the number of experimentally solved RNA-ligand and RNA-protein structures is still insufficient. Therefore, binding sites prediction of non-coding RNA is required to understand their functions. Current RNA binding site prediction algorithms produce many false positive nucleotides that are distance away from the binding sites. Here, we present a network approach, RBind, to predict the RNA binding sites with improved and reasonable accuracy in a large-scale dataset testing.

 

 

RBind is available:

 

 

Instructions of usage of RBind closeness calculation

 

Once RBind program downloaded, you can follow the steps below to use the closeness package:

 

INPUTS:

PDB file                   - Strcutre information

Distacne cutoff        - Contact parameter for generating matirx

 

OUTPUTS:

mapping.txt              - Output file nucleotides vs PDB file nucleotides

contact.dat               - Static contact network by cutoff

closeness.txt            - Closeness Centrality

 

 

Instructions of usage of RBind degree calculation

 

Once RBind program downloaded, you can follow the steps below to use the degree package:

 

INPUTS:

contact.dat               - Static contact network by cutoff

 

OUTPUTS:

degree.txt                - Degree values

 

 

Testing datasets:

 

 

In previous structure-based RNA-ligand docking study, Anna et al. used an RNA-ligand complex structure benchmark for testing(RNA 19, 1605-1616, 2013). Here, we excluded the stranded RNA helix structure due to simple helix topology. Therefore, our RNA-ligand dataset included 22 RNA-ligand complex structures with RNA length of 20 ~ 94 nucleotides. The experimental binding sites are defined as the nucleotides within 4 Å distance from the ligand.

 

 

The RNA-protein dataset consists of 72 diverse RNA-protein structures with RNA length of 20 ~ 157 nucleotides. Since the prediction is needed when complexes are unknown, we used the unbound RNA structures for binding site prediction. The RNA-protein dataset can be also downloaded from http://zoulab.dalton.missouri.edu/RNAbenchmark/.

 

 

We predicted RNA tertiary structures using RNAComposer (http://rnacomposer.cs.put.poznan.pl). For each RNA structure prediction, we used the corresponding sequence and secondary structure on the RNA structure modeling server . All tertiary structures were predicted automatically.

 

 

We compared RBind with existing computational RNA binding sites prediction method Rsite. For each RNA, Rsite identifies as binding sites when nucleotides are located within the extreme points in the distance curve derived from the tertiary structure. We used downloaded Rsite tool to predict the analyze the RNA-ligand and RNA-protein data (http://www.cuilab.cn/rsite).

 

 

We compared RBind with existing computational RNA binding sites prediction method Rsite2. For each RNA, Rsite2 identifies as binding sites when nucleotides are located within the extreme points in the distance curvederived from the secondary structure. We used the online server to analyze the RNA-ligand and RNA-protein data (http://www.cuilab.cn/rsite2/start).

 

 

 

Contact us:

Any questions about RBind program, please email to yjzhao.wh@gmail.com.

 

 

Copyright 2017, Lab of Biophysics