What we learn from the learning rate

Cells need to sense their environment in order to survive. For example, some cells measure the concentration of food or the presence of signalling molecules. We are interested in studying the physical limits to sensing with limited resources, to understand the challenges faced by cells and to design synthetic sensors.

We have recently published a paper (arxiv version) where we explore the interpretation of a metric called the learning rate that has been used to measure the quality of a sensor (New J. Phys. 16 103024, Phys. Rev E 93 022116). Our motivation is that in this field a number of metrics (a metric is a number you can calculate from the properties of the sensor that, ideally, tells you how good the sensor is) have been applied to make some statement about the quality of sensing, or limits to sensory performance. For example, a limit of particular interest is the energy required for sensing. However, it is not always clear how to interpret these metrics. We want to find out what the learning rate means. If one sensor has a higher learning rate than another what does that tell you? 

The learning rate is defined as the rate at which changes in the sensor increase the information the sensor has about the signal. The information the sensor has about the signal is how much your uncertainty about the state of the signal is reduced by knowing the state of the sensor (this is known as the mutual information). From this definition, it seems plausible that the learning rate could be a measure of sensing quality, but it is not clear. Our approach is a test to destruction – challenge the learning rate in a variety of circumstances, and try to understand how it behaves and why

To do this we need a framework to model a general signal and sensor system. The signal hops between discrete states and the sensor also hops between discrete states in a way that follows the signal. A simple example is a cell using a surface receptor to detect the concentration of a molecule in its environment.

The figure shows such a system. The circles represent the states and the arrows represent transitions between the states. The signal is the concentration of a molecule in the cell’s environment. It can be in two states; high or low, where high is double the concentration of low. The sensor is a single cell surface receptor, which can be either unbound or bound to a molecule. Therefore, the joint system can be in four different states. The concentration jumps between its states with rates that don’t depend on the state of the sensor. The receptor becomes unbound at a constant rate and is bound at a rate proportional to the molecule concentration. 

We calculated the learning rate for several systems, including the one above, and compared it to the mutual information between the signal and the sensor. We found that in the simplest case, shown in the figure, the learning rate essentially reports the correlation between the sensor and the signal and so it is showing you the same thing as the mutual information. In more complicated systems the learning rate and mutual information show qualitatively different behaviour. This is because the learning rate actually reflects the rate at which the sensor must change in response to the signal, which is not, in general, the equivalent to the strength of correlations between the signal and sensor. Therefore, we do not think that the learning rate is useful as a general metric for the quality of a sensor.

Becoming more certain about uncertainty in molecular systems

By Jenny Poulton

Due to the unpredictability of motion at the microscopic scale, molecular processes have randomness associated with them, exhibiting what we call thermodynamic fluctuations. A group in Germany lead by Barato and Seifert have written a series of papers, beginning with "Thermodynamic uncertainty relation for biomolecular processes" (preprint here), exploring how uncertainty in the number of reaction steps taken by a molecular process is related to the degree to which the system is constantly consuming energy.

To be more precise, Barato and Seifert consider the number of times a system completes a cycle in a given time window. A good example of this kind of setup is the rotary motor F0F1-ATPsynthase (below, image taken from Wikipedia).

This motor is used to create the chemical fuel source of the cell (ATP) from its components (ADP and inorganic phosphate P). In order to drive this process, a current of hydrogen ions flows through the top half of the motor, causing it to systematically rotate in one direction with respect to the bottom half. This rotation is physically linked to the reaction ADP + P -> ATP, and so ATP is created. This one-directional rotational motion only arises because the current of hydrogen ions continuously supplies more energy (more technically, free energy) to the system than is needed to create the ATP. We say that the current of ions drives the system.

In general, small driven systems have a bias towards stepping forward, but there is still a non-zero probability of stepping backwards due to thermodynamic fluctuations. We also cannot predict exactly how long the system will take to complete each step of the cycle, and so the time taken per step is variable. Thus the number of cycles completed in a given time is uncertain. It is, however, possible to define an average of the net number of cycles in a time window µ and a variance σ², which is a mathematical measure of the typical deviation from the average due to fluctuations. The Fano factor F = σ²/µ gives a measure of the relative importance of the random fluctuations about the average.

In the paper "Thermodynamic uncertainty relation for biomolecular processes", Barato and Seifert relate the energy consumption and the Fano factor via F ≤ 2kT /E. Here E is the energy consumed per cycle, T is the temperature and k is Boltzmann’s constant. This expression means that the Fano factor is at least as big as the quantity 2kT /E. Thus a cycle which uses a certain amount of fuel E has an upper limit to its precision, and there is an evident trade-off between the amount of energy dissipated per cycle and the Fano factor.

In the original paper, the authors only prove their relation for very simple processes. However, it has since been generalised in this paper (preprint here). The result is actually based on very deep statements about the types of fluctuating processes that are possible in physical systems. One of the challenges now is to take this fundamental insight and apply it to gain a better understanding of practical systems. Fortunately, the F0F1-ATPsynthase rotary motor is not the only example of an interesting biological system that  undergoes driven cycles; the cell contains a huge variety of molecular motors that can also be understood in this way (preprint here). Molecular timekeepers that are vital to the cellular life cycle also depend on driven cycles. Understanding the trade-offs between unwanted variability and energy consumption will be vital in engineering such systems.