Principal Investigators: Kathrin Otte and Chris Biemann

PhD students: Stylianos Mavrianos; Karim el Akoum

This project investigates rapid evolutionary processes in natural and experimental populations of Daphnia, a key genus of freshwater crustaceans that plays a central role in lake food webs. Daphnia populations are highly responsive to environmental change, making them ideal model systems for studying adaptation to factors such as warming, novel predators, and cyanobacterial blooms. Using large-scale genomic time series, the project aims to detect subtle genetic shifts that drive adaptation and resilience. 

The core innovation lies in combining population genomics with machine learning (ML). The project will develop new computational approaches capable of identifying evolutionary signals in complex genomic datasets. Several neural network architectures—CNNs, RNNs, GNNs, and Transformer models—will be tested, with a strong focus on improving interpretability through techniques such as saliency mapping. Models will be trained using simulated datasets and benchmarked against empirical genomic data. 

Experimental evolution trials will expose replicated Daphnia pulex populations to 20°C and 23°C over at least 60 generations, simulating a 3°C warming scenario. Regular genomic and phenotypic sampling will allow the ML models to detect genomic signatures of temperature adaptation. Complementary field data from lakes in northern Germany, including genomic time series and detailed environmental measurements, will be used to evaluate the models’ ability to identify evolutionary patterns in fluctuating ecosystems. 

Ultimately, the project aims to create powerful, generalizable ML tools for analyzing population genomic time series and to improve understanding of how rapidly organisms can adapt to global change. Close collaboration between experts in genomics, ecology, and data science will ensure methodological innovation and broad applicability across different biological systems.

Foto: Vera van Santvoort