Towards robust evolutionary inference with integral projection models.

Area of Science:

• Ecology
• Evolutionary Biology
• Quantitative Genetics

Background:

• Integral projection models (IPMs) are widely used for ecological and evolutionary inference.
• IPMs track population phenotype distributions using development, inheritance, survival, and fecundity functions.
• Accurate evolutionary inference requires models that characterize within-individual trait relationships and parent-offspring similarity.

Purpose of the Study:

• To analytically demonstrate how standard IPM constructions underestimate trait covariances and parent-offspring similarity.
• To show that current IPMs predict minimal phenotypic evolution due to construction artifacts.
• To present and validate alternative quantitative genetic approaches for IPM development and inheritance functions.

Main Methods:

• Analytical derivations of covariance underestimation in standard IPM development and inheritance functions.
• Comparison of standard IPM functions with a quantitative genetic framework.
• Empirical application using bighorn sheep population data.

Main Results:

• Standard IPM development functions underestimate age- and trait-value covariances due to regression to the mean.
• Standard IPM inheritance functions are inconsistent with modern genetics and underestimate familial similarity.
• Constant biometric inheritance functions in IPMs are incompatible with evolutionary processes, predicting no evolution.

Conclusions:

• Current IPM implementations contain construction artifacts that limit their utility for evolutionary inference.
• Alternative quantitative genetic approaches for development and inheritance functions can recover critical evolutionary patterns.
• Revised IPMs offer more accurate insights into phenotypic evolution and trait relationships.