Date of Degree


Document Type


Degree Name





Ronald Koder

Committee Members

Marilyn Gunner

Raymond Tu

Ilona Kretzschmar

Latha Venkataraman

Ronald Koder

Subject Categories

Biological and Chemical Physics


Proteins, Information Theory, Peptide Charge Transport, Protein Design


This thesis is the beginning of an attempt to build a coherent theory of the properties of proteins based in information theory and the duality of information theory and nonequilibrium thermodynamics. Throughout, we will adopt the viewpoint that information can act as a thermodynamic potential, which is necessary to understand how biological processes are both enabled and constrained by the laws of thermodynamics. Understanding information as a form of thermodynamic potential also clarifies the description of proteins and other biological macromolecules as “molecular machines”: meso-scale structures with emergent causal powers which perform work on their environments by irreversibly dissipating energy and processing information. The theory of molecular machines is due to Schneider and is now more than 30 years old.[1] Here, we apply the theory, and the general framework of IT/non-eq thermo to: iv 1) Develop a method for the design of artificial proteins which perform basic computational tasks. The design strategy relies on the concept of information minimization-entropy maximization within functional constraints. 2) Construct a phase diagram for an “intrinsically disordered protein” and demonstrate that such a molecule is in fact an environmental sensor which performs a 1-bit computation 3) Define the channel capacity of a protein which participates in a charge-transfer pathway via non-adiabatic tunneling, and show that proteins are capable of controlling all parameters that determine charge transport rates, making them ideal materials for molecular electronics. Furthermore, each section has an associated experimental component. In the design section we show results obtained from a test study on the design of heme-binding helical bundles with the potential to perform a diverse array of machine tasks. We compare the predictions of section 2 to data obtained through characterization of a designed IDP “molecular switch,” and finally discuss experimental work to define the “beta parameter” or conductance decay with length constant, of peptides and the implications of this data for the practical design of proteins as single-molecule transistors. We stress that understanding the nature of biological proteins as molecular machines illustrates that proteins already play this role in biology, and that the idea of using them to perform computations is the opposite of far-fetched: it simply recognizes the work proteins do already in enabling our own existence and that of the rest of biology.