A method for analysis of phenotypic change for phenotypes described by high-dimensional data.

Area of Science:

• Evolutionary biology
• Quantitative genetics
• Morphometrics

Background:

• Multivariate phenotypic data are crucial for evolutionary biology research.
• High-dimensional data (more variables than observations) pose analytical challenges for traditional parametric tests.
• Geometric morphometrics generate high-dimensional shape data, often analyzed independently, limiting biological insights.

Purpose of the Study:

• To present a nonparametric method for evaluating effect size in high-dimensional phenotypic data.
• To demonstrate the method's utility with geometric morphometric data for analyzing phenotypic change.
• To address the paradox of parametric tests requiring high observation-to-variable ratios.

Main Methods:

• Developed a nonparametric effect size evaluation method.
• Applied the method to geometric morphometric data of desert fish body shape.
• Compared sexual dimorphism between two populations using different characterizations of body shape.

Main Results:

• The nonparametric method is not constrained by the number of phenotypic variables.
• Utilizing more phenotypic variables can increase effect sizes.
• The approach allows for stronger inferences in comparative analyses of phenotypic change.

Conclusions:

• The proposed nonparametric method effectively handles high-dimensional phenotypic data.
• This method enhances the analysis of evolutionary changes, particularly in morphometrics.
• It facilitates more robust comparisons and deeper understanding of evolutionary processes.