Design effective antisense oligonucleotides
This tool predicts the efficacy of RNase H1-mediated antisense oligonucleotides (ASOs) based on their sequence and chemical modifications using the OligoAI deep learning model.
OligoAI was trained on the ASO Atlas dataset comprising 188,521 gapmer ASOs targeting 334 genes, extracted from published patents. The model achieved a 5.72-fold reduction in experimental screening effort compared to random selection.
Enter a gene name, select a transcript, define the ASO chemistry (sugar/backbone), and view predicted efficacy scores ranked by the model.
Publication: Hill et al. (2025). "Accurately modelling RNase H-mediated antisense oligonucleotide efficacy"
Resources:
• GitHub: github.com/barneyhill/aso_atlas
• Model: huggingface.co/barneyhill/OligoAI