Periodic Reporting for period 2 - OCP (Optimal Cellular Prediction)
Periodo di rendicontazione: 2022-10-01 al 2024-03-31
Living cells need to respond to changes in their environment. Yet, mounting a response takes time.Remarkably recent experiments have shown that living cells can detect correlations in the environmental fluctuations to anticipate changes in the environment and mount a response ahead of time. How living cells can predict future signals, how reliably they can do so, and what the fitness benefits of anticipation are, are question that are not understood. In this program, theory and experiments will be employed to address these questions.
Using measures from information theory and ideas from statistical physics, we will study network motifs and environmental stimuli of increasing complexity to derive the fundamental limit to the prediction accuracy as set by the information on the past. We will determine how close biochemical networks can come to this bound, and how this depends on the topology of the network and the resources to build and operate it – protein copies, time, and energy. We will elucidate how the features of the past signal that are most informative about the future signal are encoded in these optimal networks, and how the cell decodes these. The studies on these minimal model systems will uncover general principles of cellular prediction.
We will use our theoretical framework to set up experiments that allow us to test whether two specific biological systems – the E. coli chemotaxis system and the glucose sensing system of yeast – have implemented the uncovered design principles for optimal cellular prediction. We will measure how close these systems come to the fundamental bound on the prediction precision and how this constrains their fitness. We envision that this program will establish information transmission efficiency as a paradigm for understanding cellular function.
Using measures from information theory and ideas from statistical physics, we will study network motifs and environmental stimuli of increasing complexity to derive the fundamental limit to the prediction accuracy as set by the information on the past. We will determine how close biochemical networks can come to this bound, and how this depends on the topology of the network and the resources to build and operate it – protein copies, time, and energy. We will elucidate how the features of the past signal that are most informative about the future signal are encoded in these optimal networks, and how the cell decodes these. The studies on these minimal model systems will uncover general principles of cellular prediction.
We will use our theoretical framework to set up experiments that allow us to test whether two specific biological systems – the E. coli chemotaxis system and the glucose sensing system of yeast – have implemented the uncovered design principles for optimal cellular prediction. We will measure how close these systems come to the fundamental bound on the prediction precision and how this constrains their fitness. We envision that this program will establish information transmission efficiency as a paradigm for understanding cellular function.
Any system, be it a human, bacterium, or robot, must predict the future based on information from the past. That first of all requires this information to be stored. Whereas we do that in our brain, bacteria do it with the help of proteins and energy in the cell, a so-called biochemical network. Ultimately, how much information any system can possibly have about the future, the so-called predictive information, is always limited by the amount and quality of information that it extracts from the past signal. Our work showed that living cells like bacteria can, in principle, reach the fundamental limit on the predictive information as set by the past information: they can extract those bits of past information that are most informative about the future. Yet, our work also shows that reaching this fundamental bound is prohibitively costly, in terms of protein copies and energy. Optimal systems that maximize the predictive information for a given resource cost, as set by protein copies and energy, are indeed not at the information bound. They extract the bits of past information with the best price-to-quality ratio of cost versus predictive power.
We applied our theory to the chemotaxis system of the bacterium E. coli. Using experimental data, we showed that E. coli is indeed not at the information bound, as predicted by our theory. Moreover, our work revealed that for shallow ligand-concentration gradients where sensing becomes particularly challenging, the bacterium is very close to being optimal in terms of predictive information per resource cost. This analysis thus not only suggests that E. coli has been optimized for chemotaxis in challenging environments, but also provides firm support for our newly uncovered design principle that biological systems extract information with the best price-quality ratio.
In parallel, we have also developed a new algorithm, called PWS, that, for the first time, enables the exact computation of the information transmission rate for any stochastic system.
We applied our theory to the chemotaxis system of the bacterium E. coli. Using experimental data, we showed that E. coli is indeed not at the information bound, as predicted by our theory. Moreover, our work revealed that for shallow ligand-concentration gradients where sensing becomes particularly challenging, the bacterium is very close to being optimal in terms of predictive information per resource cost. This analysis thus not only suggests that E. coli has been optimized for chemotaxis in challenging environments, but also provides firm support for our newly uncovered design principle that biological systems extract information with the best price-quality ratio.
In parallel, we have also developed a new algorithm, called PWS, that, for the first time, enables the exact computation of the information transmission rate for any stochastic system.
It had been long known that making predictions requires extracting and storing information from the past signal. However, past studies focused on deriving and analyzing systems and schemes that maximize the predictive information per amount of past information: these studies focused on optimal compression schemes, which compress the past input as much as possible while retaining the information content, i.e. the information about the future input. In contrast, our work shows that while this is an important design criterion, it is equally vital to consider the biophysical cost of extracting information, in terms of protein copies and energy. Ultimately, one needs to consider the price-quality ratio of information: the ratio of the predictive content of the stored information versus the cost of obtaining and storing that information. Importantly, while we have focused on cellular prediction, this principle is far more general: any data compression scheme needs to consider both the information content and the physical cost of compressing it, in terms of material and energy cost. The design principle thus uncovered by our work is thus of very wide significance.
The PWS scheme to compute information transmission is truly a breakthrough. This year it is precisely 75 years ago that Shannon showed how information transmission should be quantified, namely via the mutual information. Yet, exactly computing this mutual information for time-varying signals, necessary to compute the information transmission rate, had been impossible: all existing schemes, except for the simplest systems, required approximations. Our scheme is the first scheme that enables the exact computation of the information transmission rate for any stochastic system: it can be applied to systems in all domains of physics, from biological systems, to optical, mechanical, and electrical systems. PWS is thus expected to have a big impact.
In the coming years, we will extend our theory of prediction to more complicated input signals, and to systems that encode the input into multiple outputs. Moreover, we will use PWS to develop a new scheme that makes it possible to obtain the information transmission rate directly from experiments.
The PWS scheme to compute information transmission is truly a breakthrough. This year it is precisely 75 years ago that Shannon showed how information transmission should be quantified, namely via the mutual information. Yet, exactly computing this mutual information for time-varying signals, necessary to compute the information transmission rate, had been impossible: all existing schemes, except for the simplest systems, required approximations. Our scheme is the first scheme that enables the exact computation of the information transmission rate for any stochastic system: it can be applied to systems in all domains of physics, from biological systems, to optical, mechanical, and electrical systems. PWS is thus expected to have a big impact.
In the coming years, we will extend our theory of prediction to more complicated input signals, and to systems that encode the input into multiple outputs. Moreover, we will use PWS to develop a new scheme that makes it possible to obtain the information transmission rate directly from experiments.