Pollen morphology, deep learning, phylogenetics, and the evolution of environmental adaptations in Podocarpus

Area of Science:

• Paleobotany
• Evolutionary Biology
• Plant Morphology

Background:

• Pollen morphology is influenced by both evolutionary history and environmental pressures.
• Understanding these influences is key to reconstructing plant evolutionary trajectories.

Purpose of the Study:

• To analyze the relationship between Podocarpus pollen traits and environmental factors using deep learning.
• To infer the evolutionary adaptations of Podocarpus to climate change.

Main Methods:

• Quantified pollen traits using deep learning algorithms.
• Applied trait-environment regression models and phylogenetic comparative methods.
• Incorporated fossil data into ancestral state reconstructions.

Main Results:

• Mean annual temperature and solar radiation significantly impact pollen morphology.
• Thermal stress leads to increased pollen size; UV-B radiation selects for thicker corpus walls.
• Fossil temperature tolerances align with paleoclimate data, suggesting early ancestors preferred warm climates.

Conclusions:

• Deep learning enhances the quantification of pollen morphological traits for evolutionary studies.
• Cool-temperature tolerance evolved independently in different Podocarpus lineages.
• This study provides insights into plant adaptation mechanisms over evolutionary timescales.