ASOCompass: Context- and Chemistry-Aware Activity Prediction for Transferable Antisense Oligonucleotide Screening
Antisense oligonucleotide (ASO) activity is jointly influenced by nucleotide sequence, chemical modification, target-RNA context, dose, delivery protocol, and cellular environment. Most existing computational screening methods model only a subset of these factors, limiting their ability to predict experimentally measured activity across heterogeneous screening conditions and previously unseen…
ASO activity is influenced by a combination of factors, including sequence, chemical modification, target-RNA context, dosage, delivery method, and cellular environment. Current computational screening methods often only consider a portion of these variables, which can hinder their capacity to accurately forecast activity in diverse screening scenarios and novel biological contexts.
ASOCompass is a novel, context- and chemistry-aware framework designed to enhance ASO activity prediction and rank candidates. This framework combines contextualized ASO and target-RNA sequence representations with position-specific molecular representations of chemical modifications. It also integrates dosage and delivery information alongside prototype-adapted transcriptomic representations of target genes and cell lines.
To ensure the model generates chemically and biophysically relevant representations, it is trained concurrently on auxiliary molecular-property and sequence-derived thermodynamic prediction tasks. ASOCompass was tested on ASO Atlas, a substantial dataset of RNase H-mediated gapmer ASOs derived from patents, under undisclosed drug, target-gene, cell line, and combined gene-cell line conditions.
The framework demonstrated a Spearman correlation of 0.5970, surpassing the most effective ASO-specific baseline by 0.0421. Importantly, ASOCompass consistently outperformed the baseline across all four distribution shifts.
When applied to previously unseen SOD1 and KLKB1 targets, ASOCompass provided more accurate candidate ranking across various annotation budgets, achieving correlations of 0.830 and 0.696, respectively. Additional analyses suggest that supervision based on molecular properties enhances modification-specific ranking, while the auxiliary thermodynamic task generates representations more closely aligned with experimentally measured inhibition.
These findings highlight the potential of integrating sequence, chemistry, and experimental-biological context for transferable ASO screening, paving the way for more efficient and accurate screening processes in the future.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
