Periodic Reporting for period 1 - EVOTOOLS (EVOlutionary Trends in mOrphOLogical diverSity: new comparative tools for high-dimensional data)
Reporting period: 2018-10-01 to 2020-09-30
The availability of high-quality morphometric data has exploded in the past few years thanks to the increased development and availability of 3D imaging techniques (such as CT-scans). These new datasets offer a unique opportunity to better characterize and understand how phenotypic diversity has evolved in responses to changes in climates, environments and species ecology through time. However, the high dimensionality of these datasets (they are often described by a larger number of features/traits than species), prevents the use of state-of-the-art modelling and statistical phylogenetic techniques such as maximum likelihood traditionally used.
The development of methods that allow to model and to handle these bigdata in a robust statistical framework is extremely important to understand how a major part of the biodiversity linked to the functioning of species in ecosystems (their morphology) has evolved. For instance, how species have evolved during past climatic changes? How their responses were related to their ecology? How subtle variations in morphology can be linked to various biological factors (ecology, development, genetic diversity)?
The overall objectives of EVOTOOLS were then to (i) develop new phylogenetic comparative methods capable of handling high-dimensional datasets such as from 3D morphometrics, (ii) disseminate these new developments to the scientific community through implementations in free and open-source software, and (iii) apply them on comprehensive empirical datasets across tetrapods that are housed by the host institution. All together, these research objectives have the potential to shed light on major and still open questions in Evolutionary Biology.
The development of methods that allow to model and to handle these bigdata in a robust statistical framework is extremely important to understand how a major part of the biodiversity linked to the functioning of species in ecosystems (their morphology) has evolved. For instance, how species have evolved during past climatic changes? How their responses were related to their ecology? How subtle variations in morphology can be linked to various biological factors (ecology, development, genetic diversity)?
The overall objectives of EVOTOOLS were then to (i) develop new phylogenetic comparative methods capable of handling high-dimensional datasets such as from 3D morphometrics, (ii) disseminate these new developments to the scientific community through implementations in free and open-source software, and (iii) apply them on comprehensive empirical datasets across tetrapods that are housed by the host institution. All together, these research objectives have the potential to shed light on major and still open questions in Evolutionary Biology.
To reach the proposed objectives, EVOTOOLS started to evaluate the performances of alternative approaches that were currently in use to exploit high-dimensional 3D morphometrics. After identifying several limitations and biases with these approaches, the researcher proposed to develop a framework based on penalized likelihood to model high-dimensional traits on phylogenetic trees. With this new framework, statistical tools (such as multivariate phylogenetic regressions, MANOVA and MANCOVA) were extended to the difficult high-dimensional case. Further models were also developed to link morphological changes and rates of evolution to abiotic factors (such as past climate changes). These developments were incorporated into a software package for the R statistical environment. With these new developments, extensive studies were performed across major tetrapod radiations (birds, frogs, salamanders and mammals). Most results were published (or are in the process of being published) in peer-reviewed scientific journals and presented at international conferences.
Beyond studying macroevolution and the impact of past climatic and environmental changes, the new developments proposed by EVOTOOLS will be, for instance, also of tremendous interest for mapping the genotype to the phenotype using high-quality 3D morphometrics across species. This has the potential to lead to major interdisciplinary synergies in a close future, but also to bring new scientific insights that were not reachable before then. The preliminary efforts achieved in EVOTOOLS will be key for a better synthesis of the complex processes involved in phenotypic evolution with potential applications in diverse fields such as Evolutionary biology, Conservation, Paleobiology, Genetics, Computational Biology and Statistics.