Please contact us to submit larger jobs.
IDRBind provides three prediction scores/labels for each surface residue of the protein structure: CorePred, RimPred, and the IDRBind label from the conditional random field. The IDRBind label is generally the most accurate and easy to interpret, combining the results from CorePred and RimPred. However, CorePred and RimPred scores may provide additional information, predicting core and rim interface residues, respectively. For example, if the query protein has multiple putative interfaces, an interface may be overshadowed and suppressed in the IDRBind label but may still appear prominent in CorePred and RimPred scores.
For the classification output of IDRBind: 0 = non-interface, 1 = core, 2 = rim, -99.88 = buried residues or unrecognized residues
For CorePred and RimPred outputs range between 0 and 1. CorePred, for example, approximates the probability that the residue is a core residue if we ignored all the rim residues. CorePred and RimPred scores are only available in the csv file.
The class column in the csv file is for internal testing and should be ignored.The resulting PDB formatted output can be visualized in VMD by coloring the residues based on the B-factor field. It is best viewed with the color range set between 0 to 2. One can also set the range from -0.5 to 2 to differentiate the non-interface residues from the buried residues.
Yes. However, larger structures will take much longer to run. Structures that are too large may fail to run (i.e., greater 1000 residues; varies depending on the shape and geometry).
For structures large in size or surface area, it may take upwards of 2 hours. Splitting oligomers into individual chains would greatly reduce the processing time, but this will also change predicted interfaces. Please try the following in order:
1. Please check that the FASTA file is formatted correctly. See Documentation page.
2. Try formatting PDB file. We try to be flexible, but IDRBind would more likely work with files directly from the PDB.
..a) Add terminal lines at the end of each protein chain. e.g.
TER 4367 PRO A 590..b) Keep trailing spaces of each ATOM line. Many protein manipulation software do not output
.. the element column at the end of the line, but they do keep the trailing spaces.
.. However, some text editors would remove them by default. IDRBind expects the columns
.. even if they're empty.
..c) Remove hydrogen. This is not manditory, but the hydrogens will be striped from the structure and thus would not change the results anyway.
Please contact support for additional troubleshooting.