Design considerations for in silico drug discovery platform that can accelerate targeted drug repurposing: a focus on...

Over the past decades, considerable debate has arisen regarding the decline in productivity within the pharmaceutical industry. In this context, drug repurposing has become an increasingly important strategy. Because of its cost- and time-efficient nature, it offers a particularly attractive avenue for identifying therapeutic solutions for rare, neglected, and emerging diseases. However, the implementation of rational drug repurposing strategies in these fields is hindered by data scarcity.

The author focuses on the challenge of data availability in the context of rare, emerging, and neglected diseases. Strategies that can be implemented to mitigate these limitations are discussed: open data initiatives, federated learning, data augmentation, data synthesis, and data-centric strategies for AI-guided drug repurposing. Scopus was searched for relevant literature published in the last 10 years.

Drug repurposing increasingly relies on knowledge-based strategies to identify new therapeutic opportunities. The implementation of these approaches depends on the availability of high-quality data. The consolidation of data augmentation and data synthesis technologies as well as open-knowledge and federated learning paradigms could help mitigate some of the data limitations that currently constrain computational drug repurposing. Their impact remains to be established through real-world implementation.