MoreRONN Disorder Prediction Server
Powered by MoreRONN v{{moreronnVersion}} Server Online Server Offline
MoreRONN is the spiritual successor of RONN and is useful for surveying disorder in proteins as well as designing expressible constructs for X-ray crystallography. This web server only accepts 3500 amino acids at a time. If you have any questions, please .
NOTE: The standalone MoreRONN source code is available on Github for download. This version runs locally on GNU/Linux without prediction limits. Please click here for the latest release.
Predictor and web service built by Varun Ramraj. Publication currently in preparation.
How does MoreRONN work?
At the heart of the MoreRONN predictor is a neural network called the Bio-Basis Function Neural Network (BBFNN) (Thompson and Esnouf, 2004), also the basis of the precursor RONN predictor (Yang et al., 2005). MoreRONN is different in that its input library of known disordered sequences was first clustered on the basis of weak sequence similarities. MoreRONN was then trained on this pre-clustered database using the same ten-fold cross-validation approach used in RONN. The input sequence is compared against the resulting clustered database. Rather than comparing full sequences, MoreRONN uses a "sliding window" approach of comparing overlapping 15-residue windows.
MoreRONN's only "knowledge" is the set of curated prototype disordered sequences. It does not rely on any other information (amino-acid propensities, secondary structure predictions, charge-hydropathy etc.), making it a fast and consistent predictor. On a modern processor, MoreRONN analyses approximately 150 residues per second.
Why "MoreRONN"?!
Just for a little fun. The predictor is similar to RONN but has "more" to it.
Privacy and Usage Terms
Your query sequences, and MoreRONN's corresponding output, are not stored anywhere on our servers. All communication is SSL-encrypted.
Results
Your predictions will appear below. Click "View Graph" to generate a line graph of your prediction (note that the graph may take a few seconds to render for long sequences). Disordered residues have scores >= 0.5. Right-click on the graph to save the image. Click "Download raw data" to download the raw MoreRONN prediction in plain-text format.
Each result is displayed in a format that makes it easier to identify potential constructs. Per-residue scores < 0.4 are denoted as blank spaces, scores >= 0.4 and < 0.5 are shown as '-', scores >= 0.5 and < 0.6 are identified as '=', and scores >= 0.6 are denoted by '#'. Residues predictions for disorder (i.e. scores >= 0.5) are coloured in red. Hovering over a blue residue in the sequence shows the raw score for that residue.