Bait-ER

Bait-ER estimates selection coefficients from Evolve-and-Resequence time series using a fully Bayesian framework to infer adaptive evolutionary dynamics in laboratory populations.

Key Features:

• Bayesian inference: Implements a fully Bayesian approach to estimate selection coefficients from allele frequency time series.
• Moran model: Uses the Moran model of allele evolution and explicitly incorporates overlapping generations.
• Data type compatibility: Accepts time series data from Evolve-and-Resequence (E&R) experiments and high-throughput sequencing allele frequency estimates.
• Computational efficiency: Avoids simulating empirical null distributions, reducing computational cost compared with methods that require such simulations.
• Genome-wide scalability: Designed to analyze large-scale, genome-wide datasets typical of E&R experiments.
• Robustness: Validated under various demographic and experimental conditions, showing high accuracy and precision in most scenarios.
• Limitations: Performance is reduced for trajectories dominated by genetic drift or with low starting allele frequencies.
• Comparative accuracy: Demonstrates superior accuracy relative to existing software in scenarios that extend beyond classic sweep models.

Scientific Applications:

• Selection coefficient estimation: Infer selection coefficients from time-resolved allele frequency data in E&R experiments.
• Detection of adaptation: Detect signatures of adaptation across genomes using time-series sequencing.
• Experimental evolution analysis: Study adaptive evolutionary processes in laboratory populations under controlled conditions.
• Method comparison and validation: Benchmark and compare selection inference performance across demographic and experimental scenarios.

Methodology:

Performs fully Bayesian inference using the Moran model of allele evolution with overlapping generations and does not require simulation of empirical null distributions.