Insights into rapid adaptation patterns in chironomus riparius through advanced bioinformatics

Abstract

Understanding the genetic mechanisms underlying rapid adaptation remains a significant challenge in evolutionary biology. While populations can adapt to environmental changes within just a few generations, the genetic architecture behind these rapid responses is complex. Adaptive traits are often influenced by complex networks of interacting genes, each contributing small effects to the overall phenotype. This polygenic nature of adaptation creates substantial challenges for detecting and analyzing evolutionary change, as selection can act simultaneously on many genomic regions with subtle individual effects. Traditional methods struggle to capture these distributed patterns of selection, particularly during ongoing adaptation. This thesis presents a multi-faceted investigation combining methodological development, experimental evolu- tion, and genomic analysis to examine rapid adaptation. First, I developed a novel computational approach combining unsupervised machine learning with a classic statistical test (OCSVM-FET) to detect adaptation patterns in sequencing data. Using simulated datasets, this method demonstrated supe- rior performance in detecting selection signatures across a wide range of evolutionary scenarios, particularly for highly polygenic traits under ongoing selection. The method was then applied to analyze a selection experiment on develop- ment time in the non-model organism Chironomus riparius. This experimental system revealed substantial phenotypic adaptation across multiple fitness- related traits over seven generations. More importantly, it provided an ideal test case for investigating the temporal dynamics of rapid adaptation in real populations. Application of the OCSVM-FET approach to the experimental data revealed a novel two-phase adaptation process. The initial phase showed rapid phe- notypic changes corresponding with selection on broadly shared metabolic pathways, while the subsequent phase demonstrated replicate-specific spe- cialization in signaling pathways. Notably, despite minimal overlap in selected genomic positions between replicates, all populations converged on similar regulatory pathways, particularly in key cellular signaling networks. This work provides novel insights into the temporal dynamics of rapid adaptation and demonstrates how populations can achieve similar phenotypic outcomes through distinct genetic trajectories while maintaining pathway-level conver- gence.