Periodic Reporting for period 4 - macroevolution.abc (Abiota, Biota, Constraints in Macroevolutionary Processes)
Okres sprawozdawczy: 2022-07-01 do 2023-06-30
The objective of this project was to understand evolution on different timescales. On the generation-to-generation timescale, we are able to predict evolution given information about parental phenotypes and the environment. However, on a longer time scale, such as those that represented in historical records and the fossil record, we are much less able to predict the diversity of forms that we observe. Macroevolution.abc hopes to bridge our gap in understanding evolutionary processes by elevating an understudied group called bryozoans, to a model system in which to link short-term and long-term evolutionary processes. Why are bryozoans especially good for this purpose?
Bryozoans are widely distributed in today's oceans, meaning that even when we study fossils from the deep past, we have contemporary analogues to help us understand ancient forms. They are sessile, benthic and calcify, characteristics that make them very fossilizable in the (shallow and deep) past. They are also clonal and polymorphic, traits that allow us to infer how fecund they are (or have been in the past), how they compete for space (or have competed in the past). This duality (contemporary and fossils) of observation of factors important for evolution (fitness and competition) renders bryozoans a perfect model system for the purpose of macroevolution.abc.
Bryozoans are, however, severely understudied. They are marine foulers and are an annoyance, perhaps, to leisure craft owners and companies that own marine installations. Some species have been prospected for chemicals that may be useful in cancer treatments. This said, bryozoans themselves do not have a high direct utilitarian value for the general public. However, what they can teach us about evolution is potentially applicable to problems as diverse as pandemics (e.g. using tools of inferences developed in evolutionary biology) to conservation management and climate change ecology (e.g. using ecological inference to suggest management practices).
There were two main themes in macroevolution.abc. The first was to infer past population and diversification dynamics. To do so, we developed machine-learning approaches to compile past occurrences of bryozoans available in the literature; collected new field data; and developed new statistical models to understand the contribution of ecological competition in the past to such dynamics. The second was to understand what contributed to the evolutionary success (or failure) of different groups of species with varying characteristics. To do so, we sequenced more 1000 different species of bryozoans, inferred their phylogenetic relationships, developed deep-learning approaches to be able to measure phenotypic traits, including those pertaining to fitness (e.g. fecundity).
In summary, to use bryozoans for answering big evolutionary questions, this project seeks to infer the evolutionary relationships among species of bryozoans all over the world and to study how they have evolved by collecting and analysing data on several different time scales (generational, decadal, tens of thousands of years and millions of years).
Bryozoans are widely distributed in today's oceans, meaning that even when we study fossils from the deep past, we have contemporary analogues to help us understand ancient forms. They are sessile, benthic and calcify, characteristics that make them very fossilizable in the (shallow and deep) past. They are also clonal and polymorphic, traits that allow us to infer how fecund they are (or have been in the past), how they compete for space (or have competed in the past). This duality (contemporary and fossils) of observation of factors important for evolution (fitness and competition) renders bryozoans a perfect model system for the purpose of macroevolution.abc.
Bryozoans are, however, severely understudied. They are marine foulers and are an annoyance, perhaps, to leisure craft owners and companies that own marine installations. Some species have been prospected for chemicals that may be useful in cancer treatments. This said, bryozoans themselves do not have a high direct utilitarian value for the general public. However, what they can teach us about evolution is potentially applicable to problems as diverse as pandemics (e.g. using tools of inferences developed in evolutionary biology) to conservation management and climate change ecology (e.g. using ecological inference to suggest management practices).
There were two main themes in macroevolution.abc. The first was to infer past population and diversification dynamics. To do so, we developed machine-learning approaches to compile past occurrences of bryozoans available in the literature; collected new field data; and developed new statistical models to understand the contribution of ecological competition in the past to such dynamics. The second was to understand what contributed to the evolutionary success (or failure) of different groups of species with varying characteristics. To do so, we sequenced more 1000 different species of bryozoans, inferred their phylogenetic relationships, developed deep-learning approaches to be able to measure phenotypic traits, including those pertaining to fitness (e.g. fecundity).
In summary, to use bryozoans for answering big evolutionary questions, this project seeks to infer the evolutionary relationships among species of bryozoans all over the world and to study how they have evolved by collecting and analysing data on several different time scales (generational, decadal, tens of thousands of years and millions of years).
Macroevolution.abc has now elevated bryozoans to a model system in which different ecological and evolutionary questions can continue to be answered. We have done this by inferring the genealogical relationships of c. 1000 species of living bryozoans by obtaining their molecular sequences (now publicly available). International teams of marine biologists and museum curators have contributed to specimens required for this work. We also estimated the temporal sequence of major evolutionary events in the group. We have collected, made publicly available and analysed data on ecological interactions (spatial combats) that are preserved in the fossil record and observed in contemporary benthic marine communities. Field data for ecological interactions was mainly collected in field expeditions in New Zealand, carried out by European and New Zealand scientists. Through this, we gained understanding on how trait combinations can predict outcomes of spatial combats and how these change through time. By capitalizing on Natural Language Processing (NLP) tools, including machine-learning approaches, we explored automated approaches to compiling data from the literature on the observations of different species at different geological ages or locations. Using such observations, we have estimated the past diversification rates of bryozoans and inferred the lack of evidence of the influence of local species distributions on long-term lineage-level evolution. Earlier on in the project, we manually measured huge volumes of bryozoan specimens and their phenotypic traits. Later in this project, we contributed to the development of deep-learning tools to automate such measurements. Through these developments, we were able to estimate phenotypic selection in the deep past and the consequences thereof for long-term trait evolution.
In general, we find that short-term processes/outcomes do not seem to influence the long-term survival of a species, but that phenotypic evolution is constrained, but less so than expected given canonical microevolutionary models.
The results of the project have been presented at many scientific conferences in North America, Europe and Australasia, and also shared on social media and with artists interested in the development (geometry) of bryozoans.
In general, we find that short-term processes/outcomes do not seem to influence the long-term survival of a species, but that phenotypic evolution is constrained, but less so than expected given canonical microevolutionary models.
The results of the project have been presented at many scientific conferences in North America, Europe and Australasia, and also shared on social media and with artists interested in the development (geometry) of bryozoans.
The data and insights compiled over the course of this project continue to be analysed and understood both in the context of the original goals of the project, but beyond. For example, in the course of sequencing diverse bryozoan species, we have successfully inferred the whole circularized mitochondrion of a 150-year-old sample of a bryozoan species that was collected in the Arctic even before the double helix was known. This was a serendipitous result from our exploration and marks a new era for sequencing projects, especially for understudied groups with little tissue and a life-style prone to contamination. While we did not anticipate using deep-learning approaches when we started this project, we now have the DEEPBRYO automated phenotyping pipeline to aid ourselves and other researchers in quickly accumulating morphological and count data for analyses. We anticipate the further integration of diverse data types, including molecular sequences, phenotypic data, fossil data, in giving us an integrated understanding of the differences in the dynamics and driving forces of evolution on vastly different time scales.