Validating Interpretability in siRNA Efficacy Prediction: A Perturbation-Based, Dataset-Aware Protocol

TL;DR

This paper introduces a validation protocol for saliency maps in siRNA efficacy prediction, revealing failure modes and proposing a regularizer to improve interpretability, emphasizing its importance for therapeutic design.

Contribution

The paper presents a novel counterfactual sensitivity faithfulness protocol for validating saliency maps in siRNA models and introduces a biology-informed regularizer to enhance interpretability.

Findings

Cross-dataset transfer reveals failure modes like faithful-but-wrong and inverted saliency.

Models trained on mRNA assays collapse on luciferase data, showing protocol shifts can invalidate deployment.

The regularizer improves saliency faithfulness with modest predictive trade-offs.

Abstract

Saliency maps are increasingly used as design guidance in siRNA efficacy prediction, yet attribution methods are rarely validated before motivating sequence edits. We introduce a pre-synthesis gate: a protocol for counterfactual sensitivity faithfulness that tests whether mutating high-saliency positions changes model output more than composition-matched controls. Cross-dataset transfer reveals two failure modes that would otherwise go undetected: faithful-but-wrong (saliency valid, predictions fail) and inverted saliency (top-saliency edits less impactful than random). Strikingly, models trained on mRNA-level assays collapse on a luciferase reporter dataset, demonstrating that protocol shifts can silently invalidate deployment. Across four benchmarks, 19/20 fold instances pass; the single failure shows inverted saliency. A biology-informed regularizer (BioPrior) strengthens saliency faithfulness with modest, dataset-dependent predictive trade-offs. Our results establish saliency validation as essential pre-deployment practice for explanation-guided therapeutic design. Code is available at https://github.com/shadi97kh/BioPrior.