Effects of single nucleotide polymorphism ascertainment on population structure inferences

Area of Science:

• Population Genetics
• Genomics
• Bioinformatics

Background:

• Single nucleotide polymorphism (SNP) data are crucial for studying natural populations.
• SNP genotyping is susceptible to ascertainment bias, stemming from variant selection and discovery panels.
• This bias can lead to inaccurate population genetic inferences, particularly in complex natural populations.

Purpose of the Study:

• To investigate the impact of SNP ascertainment bias on population structure inference.
• To assess how different demographic models (island, stepping-stone, population split) interact with ascertainment bias.
• To evaluate the effects of SNP discovery population selection and marker selection on bias.

Main Methods:

• Simulated SNP data under three demographic models: island, stepping-stone, and population split.
• Examined the influence of ascertainment bias introduced during SNP discovery and marker selection.
• Analyzed site frequency spectra and summary statistics to quantify bias.

Main Results:

• Ascertainment bias significantly affects site frequency spectra and summary statistics.
• The impact of bias is dependent on the interplay between population structure and ascertainment schemes.
• Population structure inferences derived from SNP data are demonstrably influenced by ascertainment bias.

Conclusions:

• Ascertainment bias poses a significant challenge for accurate population genetics research.
• The effects of ascertainment bias are context-dependent, varying with population structure and demographic history.
• Evaluating ascertainment strategies before data collection is crucial to mitigate bias and ensure reliable genetic inferences.