Predicting gene essentiality and drug response from perturbation screens in preclinical cancer models with LEAP: Layered Ensemble of Autoencoders and Predictors

2025-02-21

AI Summary

To address the limited generalizability of gene essentiality and drug response prediction across diverse preclinical cancer models—spanning cell lines, tissue types, and disease contexts—this paper proposes LEAP, a novel framework. LEAP first employs a Data-Augmented Masked Autoencoder (DAMAE) to learn robust, multi-initialization gene expression representations; it then integrates these representations via a LASSO regressor for interpretable prediction. Crucially, LEAP introduces a representation-level ensemble paradigm—aggregating diverse learned representations rather than merely ensembling predictors—thereby substantially enhancing cross-model generalizability. With low parameter count and high computational efficiency, LEAP achieves state-of-the-art performance across multiple benchmark datasets, demonstrating superior accuracy in predicting responses for unseen cancer cell lines, tissue types, and disease models. The framework exhibits strong transferability across preclinical models and is publicly available with open-source code and data.

Technology Category

• Machine Learning: Representation Learning
• Computer Vision: Representation Learning for Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Preclinical perturbation screens, where the effects of genetic, chemical, or environmental perturbations are systematically tested on disease models, hold significant promise for machine learning-enhanced drug discovery due to their scale and causal nature. Predictive models can infer perturbation responses for previously untested disease models based on molecular profiles. These in silico labels can expand databases and guide experimental prioritization. However, modelling perturbation-specific effects and generating robust prediction performances across diverse biological contexts remain elusive. We introduce LEAP (Layered Ensemble of Autoencoders and Predictors), a novel ensemble framework to improve robustness and generalization. LEAP leverages multiple DAMAE (Data Augmented Masked Autoencoder) representations and LASSO regressors. By combining diverse gene expression representation models learned from different random initializations, LEAP consistently outperforms state-of-the-art approaches in predicting gene essentiality or drug responses in unseen cell lines, tissues and disease models. Notably, our results show that ensembling representation models, rather than prediction models alone, yields superior predictive performance. Beyond its performance gains, LEAP is computationally efficient, requires minimal hyperparameter tuning and can therefore be readily incorporated into drug discovery pipelines to prioritize promising targets and support biomarker-driven stratification. The code and datasets used in this work are made publicly available.

Problem

• Predict gene essentiality and drug response
• Enhance robustness across biological contexts
• Improve computational efficiency in drug discovery

Innovation

• LEAP combines autoencoders and predictors
• Uses DAMAE and LASSO regressors
• Improves gene essentiality and drug response predictions