Periodic Reporting for period 3 - AdaptiveResponse (The evolution of adaptive response mechanisms)
Berichtszeitraum: 2021-12-01 bis 2023-05-31
All organisms have the remarkable ability to adapt to their local conditions and to cope with environmental change. However, organisms differ largely in how they respond to their environment. These response patterns have been shaped by natural selection, but the underlying evolutionary design principles are not well understood. To study the evolution of responsive traits, one needs to consider the machinery that regulates the expression of these traits in relation to the local environment. This requires a major rethink. Adaptive responses to environmental change are not mediated by single regulatory elements, but by whole networks: gene regulatory networks determine the expression of genes; metabolic networks orchestrate a huge number of metabolic reactions; the immune network responds to the attacks of pathogens; the hormone system launches physiological responses; and neural networks steer behaviour and allow to learn from experience. The evolution of such networks is governed by different principles than the evolution of single traits. The reason is that the network as a whole is evolving, making it impossible to study its evolution from a gene-centred perspective. The project will develop theoretical models for the evolution of a wide spectrum of response networks, with the hope of distilling general and overarching design principles. Pilot studies show that such networks have intriguing features: they evolve in the background without changing the phenotype; they allow the coexistence of different response strategies; and their evolution can be much faster than in traditional models. These insights have many implications: in the medical sciences they shed new light on the arms race between pathogens and the immune system; in ecology they challenge the belief that ecology and evolution act on different timescales; in the behavioural sciences, they provide a new explanation for the universal finding that individuals differ consistently in the way they respond to their environment.
In the initial phase of the project, we focussed on the development of new modelling frameworks and simulation methods. Subsequently, these were employed in modelling studies covering a wide range of research questions. The goal is to lay the foundation of a new, mechanism-based theory of evolution, and to convince the scientific community that a mechanistic approach is feasible and that it provides profound new insights into the adaptive capacity of organisms. In the final phase of the project, the insights of the various studies will be synthesized.
Development of new methods:
[M1] To study the interplay of ecology and evolution, we developed a new method for modelling the evolution of animal movement. This method produces realistic movement patterns; it is versatile and computationally efficient.
[M2] To study the evolution of learning, we developed a new modelling approach. Here, learning is mediated by an evolving neural network. Learning differs considerably from machine learning, as it takes place in a decentralized ‘per neuron’ manner.
[M3] We developed a new modelling framework for the transmission of cultural information via social learning. In contrast to traditional ‘cultural evolution’ models, cultural traits are not copied (like genes); instead cultural information is reconstructed in the neural networks of socially learning individuals.
Progress and main results:
[WP1] Evolvability. A main theme of our project is to determine the scope and limitations of ‘evolvability’, the capacity to adapt to environmental challenges. Studies on this include (a) evolutionary robustness of phenotypic plasticity; (b) evolution of mutation rates; (c) evolution of mutational switches; (d) non-genetic inheritance. Two leading journals invited us to write reviews on our novel insight that ‘evolvability’ is intimately related to ‘phenotypic plasticity’. Articles on (a) and (d) were published in 2020 and 2021; (b) resulted in an award-winning MSc thesis; various ms are in preparation.
[WP2] Evolution of learning. We studied (a) evolution of associative learning; (b) evolution of trial-and-error learning; (c) joint evolution of emotions and learning; (d) social learning and cultural evolution. In (a), we demonstrated that evolving neural networks can be highly efficient in solving associative learning tasks (ms under review). The other studies are still in progress, but pilots resulting in 5 MSc theses show that our approach works. Various ms are in preparation.
[WP3A] Interplay of ecology and evolution. Our new method for studying movement was used in various studies: (a) complex pattern formation in predator-prey coevolution; (b) animal personalities and the spatial distribution of foragers; (c) movement in a landscape of fear; (d) the joint evolution of movement and competition strategies; (e) movement and social network structure. These resulted in 3 MSc theses, 1 accepted ms and 2 ms under review, and various ms in preparation. Major findings are: evolution is much faster than previously thought; in a competitive setting, evolution fosters the emergence of individual variation.
[WP3B] Responsiveness in social evolution. By definition, social behaviour is interactive, and one would expect that responsiveness is key element. Yet, most evolutionary models neglect responsiveness. In several studies we have shown that including responsive strategies makes a huge difference for the evolutionary outcome: (a) evolution of group-living in primates; (b) evolution of parental care; (c) joint evolution of sex ratios and parental investment; (d) effects on sexual selection on evolvability. All studies are well-advanced and producing new insights; until now, they led to one publication (2020), 2 ms under review, and various ms in prep.
We organized 4 workshops:
[WS1] Eco-evolutionary implications of non-genetic inheritance (2019). Resulted in an opinion paper, published in a leading journal (2020).
[WS2] Foundations of cultural evolution (2019). Resulted in a theme issue of an interdisciplinary journal (2021).
[WS3] Evolutionary principles in industrial engineering (2019). Exchange of ideas with engineers working on machines for rubber production, resulting in >20% energy gain in rubber mixing procedures (2021).
[WS4] Evolutionary design principles of gene regulatory networks (2020).
Development of new methods:
[M1] To study the interplay of ecology and evolution, we developed a new method for modelling the evolution of animal movement. This method produces realistic movement patterns; it is versatile and computationally efficient.
[M2] To study the evolution of learning, we developed a new modelling approach. Here, learning is mediated by an evolving neural network. Learning differs considerably from machine learning, as it takes place in a decentralized ‘per neuron’ manner.
[M3] We developed a new modelling framework for the transmission of cultural information via social learning. In contrast to traditional ‘cultural evolution’ models, cultural traits are not copied (like genes); instead cultural information is reconstructed in the neural networks of socially learning individuals.
Progress and main results:
[WP1] Evolvability. A main theme of our project is to determine the scope and limitations of ‘evolvability’, the capacity to adapt to environmental challenges. Studies on this include (a) evolutionary robustness of phenotypic plasticity; (b) evolution of mutation rates; (c) evolution of mutational switches; (d) non-genetic inheritance. Two leading journals invited us to write reviews on our novel insight that ‘evolvability’ is intimately related to ‘phenotypic plasticity’. Articles on (a) and (d) were published in 2020 and 2021; (b) resulted in an award-winning MSc thesis; various ms are in preparation.
[WP2] Evolution of learning. We studied (a) evolution of associative learning; (b) evolution of trial-and-error learning; (c) joint evolution of emotions and learning; (d) social learning and cultural evolution. In (a), we demonstrated that evolving neural networks can be highly efficient in solving associative learning tasks (ms under review). The other studies are still in progress, but pilots resulting in 5 MSc theses show that our approach works. Various ms are in preparation.
[WP3A] Interplay of ecology and evolution. Our new method for studying movement was used in various studies: (a) complex pattern formation in predator-prey coevolution; (b) animal personalities and the spatial distribution of foragers; (c) movement in a landscape of fear; (d) the joint evolution of movement and competition strategies; (e) movement and social network structure. These resulted in 3 MSc theses, 1 accepted ms and 2 ms under review, and various ms in preparation. Major findings are: evolution is much faster than previously thought; in a competitive setting, evolution fosters the emergence of individual variation.
[WP3B] Responsiveness in social evolution. By definition, social behaviour is interactive, and one would expect that responsiveness is key element. Yet, most evolutionary models neglect responsiveness. In several studies we have shown that including responsive strategies makes a huge difference for the evolutionary outcome: (a) evolution of group-living in primates; (b) evolution of parental care; (c) joint evolution of sex ratios and parental investment; (d) effects on sexual selection on evolvability. All studies are well-advanced and producing new insights; until now, they led to one publication (2020), 2 ms under review, and various ms in prep.
We organized 4 workshops:
[WS1] Eco-evolutionary implications of non-genetic inheritance (2019). Resulted in an opinion paper, published in a leading journal (2020).
[WS2] Foundations of cultural evolution (2019). Resulted in a theme issue of an interdisciplinary journal (2021).
[WS3] Evolutionary principles in industrial engineering (2019). Exchange of ideas with engineers working on machines for rubber production, resulting in >20% energy gain in rubber mixing procedures (2021).
[WS4] Evolutionary design principles of gene regulatory networks (2020).
* New modelling paradigm for the evolution of movement, allowing to study the interplay of ecology and evolution in a range of modelling studies.
* A new class of neural network models for the evolution of learning, allowing to study realistic scenarios for the evolutionary emergence of learning.
* A new approach to cultural evolution, allowing to study the transmission of cultural information without assuming cultural replication.
* A mechanistic approach to plasticity and evolvability, showing that both concepts are closely intertwined and producing novel insights such as the ability of evolution to generalize over environments.
* The finding that mechanistic adaptive response models universally lead to the emergence of individual differences, leading to a new theory for the evolution of ‘animal personalities’.
* The finding that the widely applied ‘selection gradient’ method fails when populations exhibit individual differences (what they typically do in mechanistic models).
* New insights into the evolution of parental care and into sex ratio evolution.
* A new class of neural network models for the evolution of learning, allowing to study realistic scenarios for the evolutionary emergence of learning.
* A new approach to cultural evolution, allowing to study the transmission of cultural information without assuming cultural replication.
* A mechanistic approach to plasticity and evolvability, showing that both concepts are closely intertwined and producing novel insights such as the ability of evolution to generalize over environments.
* The finding that mechanistic adaptive response models universally lead to the emergence of individual differences, leading to a new theory for the evolution of ‘animal personalities’.
* The finding that the widely applied ‘selection gradient’ method fails when populations exhibit individual differences (what they typically do in mechanistic models).
* New insights into the evolution of parental care and into sex ratio evolution.