Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

New methods for drug synergy prediction: a mini-review

Fatemeh Abbasi, Juho Rousu

TL;DR

This mini-review surveys more than thirty post-2021 methods for predicting drug combination synergy, with a focus on deep learning and multi-omics data assembled from high-throughput screens. It unifies the field by mapping prediction tasks across synergy scores, data types, and evaluation protocols, and by categorizing predictive models into neural networks, forest-based, factorization, and Bayesian approaches. The review finds that the top methods excel in predicting synergy for known drug–cell-line triplets (LTO/LPO scenarios) but struggle to accurately extrapolate to new drugs or cell lines (LDO/LCO), highlighting a gap in generalization and data integration. It emphasizes the need for standardized benchmark datasets, consistent evaluation practices, and broader, richer input representations to advance extrapolative predictive accuracy and practical applicability in precision oncology.

Abstract

In this mini-review, we explore the new prediction methods for drug combination synergy relying on high-throughput combinatorial screens. The fast progress of the field is witnessed in the more than thirty original machine learning methods published since 2021, a clear majority of them based on deep learning techniques. We aim to put these papers under a unifying lens by highlighting the core technologies, the data sources, the input data types and synergy scores used in the methods, as well as the prediction scenarios and evaluation protocols that the papers deal with. Our finding is that the best methods accurately solve the synergy prediction scenarios involving known drugs or cell lines while the scenarios involving new drugs or cell lines still fall short of an accurate prediction level.

New methods for drug synergy prediction: a mini-review

TL;DR

This mini-review surveys more than thirty post-2021 methods for predicting drug combination synergy, with a focus on deep learning and multi-omics data assembled from high-throughput screens. It unifies the field by mapping prediction tasks across synergy scores, data types, and evaluation protocols, and by categorizing predictive models into neural networks, forest-based, factorization, and Bayesian approaches. The review finds that the top methods excel in predicting synergy for known drug–cell-line triplets (LTO/LPO scenarios) but struggle to accurately extrapolate to new drugs or cell lines (LDO/LCO), highlighting a gap in generalization and data integration. It emphasizes the need for standardized benchmark datasets, consistent evaluation practices, and broader, richer input representations to advance extrapolative predictive accuracy and practical applicability in precision oncology.

Abstract

In this mini-review, we explore the new prediction methods for drug combination synergy relying on high-throughput combinatorial screens. The fast progress of the field is witnessed in the more than thirty original machine learning methods published since 2021, a clear majority of them based on deep learning techniques. We aim to put these papers under a unifying lens by highlighting the core technologies, the data sources, the input data types and synergy scores used in the methods, as well as the prediction scenarios and evaluation protocols that the papers deal with. Our finding is that the best methods accurately solve the synergy prediction scenarios involving known drugs or cell lines while the scenarios involving new drugs or cell lines still fall short of an accurate prediction level.
Paper Structure (14 sections, 1 figure, 3 tables)

This paper contains 14 sections, 1 figure, 3 tables.

Table of Contents

  1. Introduction
  2. Prediction Setups
  3. Synergy scores
  4. Input data
  5. Prediction scenarios.
  6. Evaluation protocols and hyperparameter tuning
  7. Predictive models for drug synergy
  8. Neural Network Models
  9. Forest-based models
  10. Factorization models
  11. Bayesian models
  12. Prediction performance of the models
  13. Discussion
  14. Acknowledgements

Figures (1)

  • Figure 1: Summary of predictive performance of synergy prediction models: (a) Classification performance by method, (b) Regression performance by method (c) Classification performance by scenario, (d) Regression performance by scenario (e) Classification performance by dataset.