Dimensional Reduction in Complex Living Systems: Where, Why, and How

TL;DR

This paper explores how living systems achieve effective dimensional reduction through evolution and learning, using geometric insights to distinguish meaningful biological features from generic data properties, exemplified by protein evolution.

Contribution

It introduces a geometric framework for understanding dimensional reduction in living systems, linking evolution, learning, and physical reality to identify biological hallmarks.

Findings

Living systems perform dimensional reduction via evolution and phenotypic plasticity.

Geometric methods help distinguish biological features from generic data.

Application to protein evolution demonstrates the framework's utility.

Abstract

The unprecedented prowess of measurement techniques provides a detailed, multi-scale look into the depths of living systems. Understanding these avalanches of high-dimensional data -- by distilling underlying principles and mechanisms -- necessitates dimensional reduction. We propose that living systems achieve exquisite dimensional reduction, originating from their capacity to learn, through evolution and phenotypic plasticity, the relevant aspects of a non-random, smooth physical reality. We explain how geometric insights by mathematicians allow one to identify these genuine hallmarks of life and distinguish them from universal properties of generic data sets. We illustrate these principles in a concrete example of protein evolution, suggesting a simple general recipe that can be applied to understand other biological systems.