Pilea: profiling bacterial growth dynamics from metagenomes with sketching

Abstract Background Quantifying bacteria’s growth rates is essential for understanding their ecological roles and for building predictive models in environmental and clinical settings. Peak-to-trough ratios (PTRs) derived from shotgun metagenomes offer a culture-independent proxy for in situ growth rates of bacterial species, yet their reliable computation remains challenging. Results We introduce Pilea ( https://github.com/xinehc/pilea ), an alignment-free, sketching-based method that incorporates statistical models for robust PTR estimation. Pilea achieves speed improvements over existing methods while also enhancing accuracy, as demonstrated on both simulated and real datasets. Conclusions By scaling efficiently to comprehensive reference collections such as the Genome Taxonomy Database (GTDB), Pilea enables large-scale analyses of bacterial growth dynamics across biomes, unlocking new insights for ecological research. Video Abstract