A Federated Learning Benchmark for Drug-Target Interaction

TL;DR

This paper introduces a federated learning benchmark for drug-target interaction prediction, demonstrating improved performance and robustness on DTI datasets while respecting data privacy constraints.

Contribution

It presents a federated learning approach tailored for DTI data, showing its effectiveness and resilience against non-IID data issues in pharmaceutical applications.

Findings

Up to 15% performance improvement over non-private methods

Non-IID data does not significantly harm federated learning in DTI

Trade-off identified between data benefits and client costs

Abstract

Aggregating pharmaceutical data in the drug-target interaction (DTI) domain has the potential to deliver life-saving breakthroughs. It is, however, notoriously difficult due to regulatory constraints and commercial interests. This work proposes the application of federated learning, which we argue to be reconcilable with the industry's constraints, as it does not require sharing of any information that would reveal the entities' data or any other high-level summary of it. When used on a representative GraphDTA model and the KIBA dataset it achieves up to 15% improved performance relative to the best available non-privacy preserving alternative. Our extensive battery of experiments shows that, unlike in other domains, the non-IID data distribution in the DTI datasets does not deteriorate FL performance. Additionally, we identify a material trade-off between the benefits of adding new data, and the cost of adding more clients.