Welcome to PiRaNhA

PiRaNhA is a server for predicting RNA-interacting residues in protein sequences. The method is based on a Support Vector Machine created using a combination of features including: position specific scoring matrices, residue interface propensity, predicted residue accessibility and residue hydrophobicity, in overlapping subsequences. The method achieves a Matthews Correlation Coefficient (MCC) of 0.50 in a 5-fold cross-validation on a dataset of 81 known RNA-binding proteins. This server enables users to upload (single or multiple) protein sequences and obtain information on potential RNA-binding residues.

PiRaNhA requires Netscape v4.0 or Internet Explorer v4.0 and Java script enabled.
This website is free and open to all users and there is no login requirement. You can receive results live without an email address but as the calculations take >5 minutes in most cases it is recommended you enter an email address to receive results as attached files.