Chemical space visual navigation in the era of deep learning and Big Data
Sergey Sosnin
Abstract
• The size of chemical libraries has grown to millions of compounds, while human analytical capabilities remain constant, highlighting the necessity for efficient visualization methods. • Recently, new methods and tools have been developed for the large-scale visualization of chemical space. • Today, chemical space visualization extends beyond chemical compounds to include chemical reactions, chemical libraries, and even artistic representations. • Integrating parametric models with deep generative modeling can create a platform for interactive, Human-in-the-Loop exploration of chemical space. The ‘Big Data’ era in medicinal chemistry presents new challenges for analysis. While modern computers can store and process millions of molecular structures, final decisions in medicinal chemistry remain in human hands. However, the ability of humans to analyze large chemical data sets is limited by cognitive constraints, creating a demand for methods and tools to visualize chemical space. In this review, I highlight recent advances in algorithms and tools for visual navigation in chemical space. I explore how these methods are evolving to address the ‘Big Data’ challenge and discuss unconventional applications, including the visual validation of quantitative structure–activity relationship (QSAR)/quantitative structure–property relationship (QSPR) models, interactive generative approaches, and even the use of chemical space maps as digital art.