Building Bridges from Genome to Physiology Using Machine Learning and

Area of Science:

• Evolutionary Biology
• Genomics
• Physiology

Background:

• Drosophila experimental evolution provides valuable genetic material for functional physiology analysis.
• Identifying gene-to-phenotype relationships is challenging in the genomic era due to polygenic traits.
• Experimental evolution in Drosophila shows multiple phenotypes change due to genome-wide loci evolution.

Purpose of the Study:

• To develop and apply a fused lasso additive model to infer causal loci influencing physiological phenotypes.
• To identify specific genomic regions causally linked to physiological trait differentiation in experimentally evolved Drosophila populations.

Main Methods:

• Utilized 50 Drosophila populations selected for diverse life histories and stress resistance.
• Assayed differentiation in cardiac robustness, starvation, desiccation resistance, lipid, glycogen, water content, and body mass.
• Combined physiological data with pooled-seq genomic data using a fused lasso additive model.

Main Results:

• Identified approximately 2,176 significantly differentiated 50-kb genomic windows across 50 populations.
• Pinpointed 142 genomic regions highly likely to have a causal effect on specific physiological characters.
• Successfully linked genomic sites to physiological trait differentiation.

Conclusions:

• The fused lasso additive model effectively identifies potentially causal genomic regions for physiological traits.
• This approach advances understanding of genotype-phenotype relationships in complex evolutionary contexts.
• The study provides a framework for dissecting the genetic architecture of adaptive physiological changes.