Best practices in machine learning for chemistry. Article review

Chemistry has experienced a fundamental shift toward the use of statistical modeling and analysis in the constantly changing world of scientific research. These technologies have proven to be priceless resources that go beyond the limitations of human intuition and help researchers discover brand-new, elusive links. Chemspace offers a comprehensive service that combines DNA-encoded libraries (DEL), machine-learning models, and large chemical spaces to enhance early Drug Discovery. This approach, initially introduced by Kevin McCloskey et al., has shown promising results in obtaining active compounds for screening. Chemspace’s full-service approach includes DEL screening, machine-learning model development, and the provision of potentially active compounds, making it ideal for projects with limited structural or activity data.