Technion Study Finds Evolution Favors More Stable and Resilient Paths Over Random Choice

A new study by researchers at the Technion reveals that evolution does not always proceed randomly when multiple evolutionary paths yield the same fitness level. The research, published in the journal PNAS, was led by Razi Fakhr al-Din from the Faculty of Applied Mathematics and Professor Naama Brenner from the Wolfson Faculty of Chemical Engineering. They examined evolutionary dynamics and found that populations tend to evolve toward more stable and resilient states rather than randomly selecting among equivalent fitness options.

The study explains that while evolution is often described as climbing fitness peaks that improve an organism's survival, when several paths offer similar fitness, evolution preferentially moves toward "flatter" regions in the evolutionary landscape. These flatter areas provide greater tolerance to changes and disturbances, offering a survival advantage. This suggests that biological robustness and resilience can emerge spontaneously in evolution, even without direct selective pressure favoring these traits.

The findings have broad implications for understanding the development of complex biological systems and may help explain biological variability observed in nature. Additionally, the research could bridge evolutionary biology with fields like deep learning, where neural networks similarly gravitate toward stable, flat minima that enhance generalization. The study received partial funding from the US-Israel Binational Science Foundation (BSF).