NEW! Source code for standalone (local) RONN is available on GitHub here. We suggest compiling and running it on GNU/Linux.
Please note that this version is not actively maintained and has been superseded by MoreRONN, a brand new predictor. We recommend you use MoreRONN for future predictions.
Therefore, bugs and string handling have not been fixed in this version and so, please follow these key instructions to obtain a proper prediction:
1) Please provide only one input sequence!
2) Please use a FASTA-header. After the header, insert a newline ('Return' key).
3) The rest of the sequence should not have any newline characters.
RONN will try its best to handle inconsistencies in the input, but I cannot guarantee a valid prediction in these cases. Furthermore, it is probably better to wait for the new predictor if you need multiple sequences handled quickly. Alternatively, please email us using the link below if you need large-scale predictions done with the new algorithm.
On that note, RONN is much slower than MoreRONN, so this version is mainly for posterity without any optimisations or improvements (save for a new data set that was curated in 2012, which provides a modest improvement in prediction).
Please email us with any other questions or comments.
Finally, if you used RONN, please cite:
Yang, ZR, Thomson, R, McNeil, P, and Esnouf, RM (2005). RONN: the bio-basis function neural network technique applied to the detection of natively disordered regions in proteins. (Link to full text)