Detecting signatures of selection on gene expression

Area of Science:

• Evolutionary Biology
• Genomics
• Transcriptomics

Background:

• Phenotypic diversity arises from gene expression changes, making transcriptome evolution a key research area.
• Existing models for sequence evolution lack a parallel in gene expression evolution.
• Comparative transcriptomic approaches are susceptible to biases, potentially yielding false signatures of selection.

Purpose of the Study:

• To review and outline common approaches for studying gene expression evolution and their inherent biases.
• To integrate phylogenetic comparative methods with transcriptomics to identify pitfalls in inferring selection on expression patterns.
• To highlight the multi-dimensional nature of transcriptional variation and identify future research directions.

Main Methods:

• Review of existing methodologies for analyzing gene expression evolution.
• Integration of phylogenetic comparative methods and transcriptomic data analysis.
• Simulation studies to assess the impact of confounding factors like tissue composition shifts on selection inference.

Main Results:

• Comparative approaches for studying gene expression evolution are prone to biases.
• Shifts in tissue composition can significantly skew inferences of selection on gene expression patterns.
• Transcriptional variation is inherently multi-dimensional, complicating selection analysis.

Conclusions:

• There is a need for robust models to study gene expression evolution, analogous to those for sequence evolution.
• Careful consideration of biases, such as tissue composition, is essential when inferring selection from transcriptomic data.
• Further research is required to fully understand how selection operates on the complex, multi-dimensional transcriptome.