Navigating the pitfalls of applying machine learning in genomics

The scale of genetic, epigenomic, transcriptomic, cheminformatic and proteomic data available today, coupled with easy-to-use machine learning (ML) toolkits, has propelled the application of supervised learning in genomics research. However, the assumptions behind the statistical models and performance evaluations in ML software frequently are not met in biological systems. In this Review, we illustrate the impact of several common pitfalls encountered when applying supervised ML in genomics. We explore how the structure of genomics data can bias performance evaluations and predictions. To address the challenges associated with applying cutting-edge ML methods to genomics, we describe solutions and appropriate use cases where ML modelling shows great potential. Machine learning is widely applied in various fields of genomics and systems biology. In this Review, the authors describe how responsible application of machine learning requires an understanding of several common pitfalls that users should be aware of (and mitigate) to avoid unreliable results.

Automated summary

This review discusses common pitfalls encountered when applying supervised machine learning in genomics, emphasizing how the structure of genomics data can bias performance evaluations and predictions. It also presents solutions and appropriate use cases to address the challenges associated with applying cutting-edge ML methods to genomics.