Abstract

Advances in structural and cell biology – in parallel with laboratory evolution – indicate complex relationships between sequences, structures and functions of proteins, which may not be described by deterministic principles. Adaptive transitions in conformational landscapes simultaneously affect ambiguous as well as context-specific interactions, the functional outcomes of which are hard to predict. To describe dynamical behaviour of proteins under stochastic cellular conditions, I propose to include multiple functionally relevant states and their uncertainties in structure-function relationships. In analogy to artificially intelligent (AI) systems, the fuzzy set theory offers a mathematical framework to describe the appropriate sequence, structure and function sets and their uncertainties; whereas their connections could be treated by a fuzzy inference model. I discuss a series of results supporting the fuzzy model, from enzymatic catalysis to dynamic protein complexes and higher-order structures. I will also present various approaches/formalisms how uncertainty could be estimated from sequence and forecast cellular behaviours.