Time Series Data Mining for Linking the Shape of Bacterial Growth Curves to Biological Functions

Connecting bacterial growth dynamics to biological functions is essential for understanding microbial systems, yet studies directly examining growth dynamics remain limited. Here, we analyzed over 10,000 growth curves from single-gene knockout Escherichia coli strains using combined dynamic time warping and derivative dynamic time warping methods, followed by hierarchical clustering based on shape similarity with and without considering experimental replicates. Clustering revealed groups enriched for specific gene categories and biological processes, particularly enzymes and biosynthesis pathways. Growth curves with high reproducibility were associated with conserved biosynthetic functions. These findings demonstrate that time series data mining can effectively link bacterial growth dynamics to biological functions, providing a framework for interpreting complex genetic effects on population behavior and advancing data-driven approaches in biotechnological and microbial research.