Prepare centered sequences
Recommended input is 501 or 601 nt. Candidate sites sit at position 251 or 301.
CROSSMOTIF / ONLINE PREDICTION
Select a modification type and enter candidate-site centered sequences to obtain site-level predictions.
Scroll inside the result area to review records
READY FOR INPUT
Results will show modification probability, model decision and candidate center-site information.
CROSSMOTIF INFERENCE
The first prediction loads model weights. Please keep this page open.
| Record | Center Site | Modification Probability | Decision |
|---|
INPUT GUIDE
Recommended input is 501 or 601 nt. Candidate sites sit at position 251 or 301.
Adenosine for m6A/m1A, cytidine for m5C/f5C, guanosine for m7G, uridine for Ψ.
Enter a single sequence or paste standard FASTA; up to 500 records per request.
ABOUT THE MODEL
A sequence prediction model for RNA modification sites, supporting six common modification types.
The first model load takes longer; later predictions for the same type reuse the loaded weights.
Paper: CrossMotif: Cross-Modal Learning of Site-Centered Motifs and Global Context for Multi-Type RNA Modification Site Prediction
Repository: github.com/kaiyangli-nyist/CrossMotif
Email: kaiyangli@nyist.edu.cn