Document Type

Conference Paper

Publication Date

2026

DOI

10.1145/3807503.3819481

Publication Title

BCB '26: Proceedings of the 17th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics

Pages

20 (9 pp.)

Conference Name

17th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics, 30 June-3 July 2026, New York, NY, United States

Abstract

Drug combination therapy in disease management gained popularity in the last few decades. Computational modeling of such combinations is an active area of research in the drug discovery domain. While earlier approaches solely emphasized on the structural features of participating drugs for designing synergistic models, they lack other crucial factors directly linked with drug administration - omics expressions. As differential omics expression is a downstream consequence of the administered drug combinations, utilizing such expressions while designing synergistic models promises robust and dynamic modeling. In this work, we propose SynergyLM that fuses multi-omics features with drug embeddings to build an omics-aware synergy model. Drug embeddings are extracted from a chemical language model fine-tuned on a vast chemical compound dataset. The omics expressions come from high-throughput drug screening studies for the NCI-60 cancer cell lines. The proposed model utilizes three omics expressions - mRNA, miRNA and proteomics. Attention-based fusion method is used to learn the inter-relations of those omics and generate unified hidden representations. Finally, those hidden omics representations are concatenated with drug embeddings and fed to a regression head to predict drug synergy. SynergyLM outperforms state-of-the-art models designed for pairwise drug synergy prediction. The proposed approach thus suggests multi-omics guided paths for practical and effective synergistic-drug formulation.

Rights

© 2026 Copyright held by the owner/authors.

This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.

Data Availability

Article states: "The source code to reproduce the results presented in this paper can be accessed at https://github.com/debnathk/drug-synergy. Drug embeddings are generated using MolFormer-XL and ChemBERTa-2, both are open source chemical language models. The model cards along with usage examples can be found in Hugging Face. Drug Comb dataset can be used from Therapeutics Data Commons(TDC), a free resource for deep learning-ready datasets in therapeutic do main."

Original Publication Citation

Debnath, K., Rana, P., & Ghosh, P. (2026). Attention-based multi-omics fusion for drug synergy prediction. In Proceedings of the 17th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (pp. 1-9). Association for Computing Machinery. https://doi.org/10.1145/3807503.3819481

ORCID

0000-0001-9199-2479 (Rana)

Download

Share

COinS