10-14 March 2025

To foster international participation, this course will be held online

Techniques to explore the evolution of multidimensional traits

 The study of multidimensional traits has become integral to ecological and evolutionary research. These encompass both morphometric (e.g., geometric morphometrics) and non-morphometric (e.g., gene expression across multiple genes) phenotypic traits. While analysing multidimensional phenotypes poses unique challenges compared to univariate traits (e.g., body size), it provides unparalleled insights into the processes producing the diversity of life forms. This course will cover a range of techniques—from widely used methods to more specialized approaches—to examine variation in multidimensional traits at both microevolutionary and macroevolutionary scales.

This five-day course combines lectures and hands-on sessions. Lectures will address both foundational knowledge and its practical applications in research. During hands-on sessions, participants will work with example datasets and will be able to apply their learning to their own data. While the hands-on sessions will primarily use geometric morphometric example datasets, the course will also provide substantial information on applying these techniques to non-morphometric data. To maximize learning, we will use user-friendly software with graphical interfaces whenever possible. However, information on R implementations and scripts will be provided for all techniques.

This course is designed for beginners and intermediate users—researchers who wish to analyze multidimensional phenotypic data (e.g., geometric morphometric data) or those who have started analyses but seek a more structured foundation.

Participants should have a basic understanding of statistical concepts and experimental design. Familiarity with R is not essential but beneficial, as some of the more specialised techniques are not available in point-and-click software.

From 2018 to 2024, Physalia Courses offered a course in Geometric Morphometrics for beginner and intermediate users. The first two editions were held in Berlin, Germany, and participants contributed to a study published in 2020 (Fruciano et al., 2020 - Zoological Journal of the Linnean Society; https://doi.org/10.1093/zoolinnean/zlz069). Subsequent editions were successfully conducted online due to the COVID-19 pandemic. Starting in 2025, Physalia offers this course focused on multidimensional phenotypic data analysis and a companion course on Foundations in Geometric Morphometrics  focusing on geometric morphometric data acquisition.

Monday– Classes 1-7 pm Berlin time

 Introduction to multidimensional phenotypes
•    Overview of morphometric (traditional and geometric) and non-morphometric (e.g., gene expression) phenotypes
•    Challenges and opportunities provided by multidimensional phenotypes
•    Key data transformation techniques
 
Principal component analysis (PCA)
•    Fundamentals of PCA
•    Unconventional uses of PCA (e.g., to model and remove artefactual variation)
•    Understanding the limitations of PCA
•    Strategies for handling large datasets (including regularization)
 
Tuesday– Classes 1-7 pm Berlin time

Comparing and exploring group differences:
•    Between-group PCA
•    Tests of difference in mean phenotype (parametric and permutational methods)
•    Classification: discriminant function analysis (DFA)/canonical variate analysis (CVA)
•    Notes on machine learning approaches to classification

 Relationships between multidimensional phenotypes and predictors:
•    Regression and general linear models
•    Testing approaches (distance-based, regularization-based)
•    Allometry
 
Wednesday– Classes 1-7 pm Berlin time

Multidimensional phenotypes in a macroevolutionary context I:
•    Phylogenetic non-independence
•    Phylogenetic comparative methods for multidimensional phenotypes
•    Analysing the relationship between multidimensional phenotype and predictors (phylogenetic independent contrasts and phylogenetic generalised least squares, including using regularization)

Co-variation between multidimensional data:
•    How things co-vary: partial least squares analysis (PLS)
•    Measuring the strength of co-variation: levels of association between multidimensional datasets
•    Association between phenotypes and ecological data
•    Modularity and phenotypic integration within species and in a macroevolutionary context

 Thursday – Classes 1-7 pm Berlin time

Multidimensional phenotypes in a macroevolutionary context II:
•    Fitting macroevolutionary models to multidimensional phenotypes (e.g., Brownian motion, Ornstein–Uhlenbeck)
•    Quantifying and comparing rates of phenotypic evolution across clades and lineages

Quantifying and testing for parallelism and convergence in multidimensional phenotypes:
•    Within species
•    In a macroevolutionary context
 
 Friday– Classes 1-7 pm Berlin time

Notes on integrating multidimensional phenotypes and genetic/genomic data:
•    Broad associations at the population level
•    Elements of association mapping (QTL analysis)

Assessing variability of multidimensional phenotypes:
•    Quantifying disparity and morphospace occupation

Geographic variation in phenotypes:
•    Trend surface analysis of multidimensional phenotypes
 
Review and open discussion:
•    Presentation of participants’ analyses on their own data