How OligoAI Works
4-step deep learning pipeline for ASO candidate generation
STEP 01
Input & Retrieval
Provide a gene ID (e.g., ENSG00000...) or raw mRNA sequence. Gene IDs are resolved via Ensembl API.
STEP 02
LightGBM Prescreen
Traditional biophysical features (GC%, Tm, ΔG, position) used to rapidly pre-filter candidate windows.
STEP 03
RINALMo Embeddings
Deep RNA language model generates contextual sequence embeddings capturing structural accessibility.
STEP 04
Ranked Output
Candidates ranked by inhibition score. Filter by GC%, Tm, region, and pass/fail status. Export to CSV.
Target Input
Enter a gene ID or paste a raw mRNA sequence
⏱ Prediction typically takes 30–90 seconds
Ctrl+Enter to run
Ctrl+Enter to run
Initialising…
0%
Gene Retrieval
Window Generation
LightGBM Screen
RINALMo Embed
Scoring & Ranking