ESTIMATING THE FORM OF NATURAL SELECTION ON A QUANTITATIVE TRAIT

Area of Science:

• Evolutionary Biology
• Quantitative Genetics
• Ecology

Background:

• The fitness function (f) links individual fitness to quantitative traits under natural selection.
• Understanding fitness functions is crucial for predicting fitness variations and identifying selection optima.
• Current methods like quadratic regression may fail to detect key features such as stabilizing or disruptive selection.

Purpose of the Study:

• To develop a nonparametric method for estimating fitness functions without requiring a priori models.
• To provide a reliable way to identify modes (stabilizing selection) and dips (disruptive selection) in fitness landscapes.
• To offer insights into selection intensity and suggest appropriate parametric models for fitness functions.

Main Methods:

• A novel, model-agnostic procedure for estimating the fitness function (f).
• Nonparametric estimation of f directly from phenotypic and fitness data.
• Analysis of six diverse datasets to evaluate the method's performance.

Main Results:

• The nonparametric method successfully estimates fitness functions and reveals diverse selection patterns.
• The approach can identify stabilizing and disruptive selection, which are often missed by traditional methods.
• Analysis of datasets provided new insights into familiar evolutionary cases.

Conclusions:

• The developed method offers a robust alternative for fitness function estimation in evolutionary studies.
• Limitations include low variation and sparse data, suggesting experimental manipulation to broaden trait distributions.
• The approach enhances understanding of natural selection and quantitative trait evolution.