How to predict?

The Predict page is designed for sequence-based inference of RNA-binding protein (RBP)-RNA interactions across multiple species. Users are required to provide RNA sequences in a FASTA-like format, and then select the target Protein and Species for analysis from the 184 RBPs currently available for 11 species. The web server applies a default probability threshold of 0.9 to determine whether a query is classified as binding or non-binding, although this threshold can be adjusted by the user according to the desired level of stringency. An example input can also be loaded directly through the Example button.

A brief usage workflow is as follows. First, paste one or more RNA sequences into the input box using FASTA-like formatting, following the length and sequence-number constraints indicated in the placeholder text. Second, choose the RBP of interest in the Protein menu and the corresponding species model in the Species menu. Third, review or modify the probability threshold if needed. Finally, click Submit to launch prediction; alternatively, click Example to populate the form with a demonstration sequence before submission.

What’s the predicted results?

The Prediction results table summarizes the final prediction outcome for each input RNA sequence. Each row corresponds to one query sequence, where Read indicates the input sequence index, RNA name denotes the FASTA header or sequence identifier, and Protein and Species specify the selected RBP and species model used for inference. Probability represents the predicted likelihood of interaction between the query RNA and the selected protein, whereas Bind provides the final binary classification based on the user-defined probability threshold, with values above the threshold interpreted as binding and those below the threshold interpreted as non-binding. You can also click the Download button to download the results.

Reference