Liu Group
PRIME-3D2D(PDB)
PRIME-3D2D(yeast genome)
HELP
PRIME
RMalign

PRIME-3D2D is a method for predicting protein-RNA complex structures and binding sites using the proteins 3D structure and the RNA 2D structure. On the PDB scale, PRIME-3D2D can predict both binding sites on protein and RNA. On the genome scale, the method can predict the binding sites of proteins on RNA.

The RNABLOSUM80 can be downloaded:RNABLOSUM80
Input:
Input for PDB scale:
First, the protein and RNA structures that you want to predict must be provided by the following two ways:
  • 1. Provide a PDB id. If you provide structure by this way, the webserver will fetch the structure from PDB.
  • 2. Upload a file.
    • Second, the chain id must be assigned.

    Input for genome scale:
    First, the protein and RNA structures that you want to predict must be provided by the following way:
  • 1. For protein, you can provide a PDB id, then the webserver will fetch the structure from PDB. Besides, you can upload a file.
  • 2. For RNA, you must provide the 2D structure of RNA, such as 2D structure example
    • The secondary structure state ends with "#FS".
    Output:
    Output for PDB scale:
    *ResultOut, it's a result file sorted by 3D2Dscore. The higher the ranking, the target structure and the template structure are the more similar. Each column represents: template ID, TMScore, RNAsimilarity 3D2DScore, 5-9 and 10-14 columns are TP, FP, TN, FN and N for protein and RNA binding sites prediction, respectively.
    PRIME_out, it's a result file including the sequence and the predicted binding sites, in which ‘+’ indicates prediction as a binding site, '-' is predicted as a non-binding site, and ' ' indicates that residues/bases cannot be aligned between target and template.
    PRIME_err, it's the error file including error message.

    Output for genome scale:
    *resultout, it's a result file sorted by 3D2Dscore. The higher the ranking, the target structure and the template structure are the more similar. 1-4 columns represent template ID, TMScore, RNAsimilarity, 3D2DScore. If the binding sites of target RNA are provided, then 5-9 columns are represent TP, FP, TN, FN and N for RNA binding sites prediction.
    PRIME_out, it's a result file including the sequence and the predicted binding sites, in which ‘+’ indicates prediction as a binding site, '-' is predicted as a non-binding site, and ' ' indicates that residues/bases cannot be aligned between target and template.
    PRIME_err, it's the error file including error message.

    Note:
    If you use our server a few days ago and the results cannot be downloaded, this is because the IP of our server was changed a few days ago. It can be used normally, and the calculation resultscan also be downloaded normally (2020-07-28).

    References:
      1. Juan Xie, Jinfang Zheng, Xu Hong, Xiaoxue Tong & Shiyong Liu. PRIME-3D2D is a 3D2D model to predict binding sites of protein-RNA interaction. Communications Biology, 16 July 2020 [PubMed]
      2. Jinfang Zheng, Juan Xie, Xu Hong & Shiyong Liu. RMalign: an RNA structural alignment tool based on a novel scoring function RMscore. BMC Genomics, 8 Apr 2019 [PubMed]
    PHY logo