Correcting for Bias in Estimates of θ w and Tajima's D From Missing Data in Next-Generation Sequencing

Area of Science:

• Population genetics
• Genomics
• Evolutionary biology

Background:

• Population genetic analyses rely on the site frequency spectrum to understand evolutionary processes.
• Watterson's estimator (θ) and Tajima's D are key statistics for genetic diversity and detecting non-neutral evolution.
• Missing data in genomic datasets, especially in Variant Call Format (VCF) files, can introduce biases into these estimates.

Purpose of the Study:

• To assess the impact of missing genomic data on population genetic summary statistics.
• To evaluate biases in Watterson's estimator and Tajima's D across different software.
• To develop and implement methods for correcting these biases.

Main Methods:

• Simulated neutral genomic data with controlled levels of missing genotypes and sites.
• Analyzed simulated data using multiple population genetics software packages (VCFtools, PopGenome, pegas, scikit-allel).
• Developed and integrated bias correction functions into the pixy software.

Main Results:

• Consistent underestimation of Watterson's estimator (θ) was observed across software due to missing data.
• Biases in Tajima's D estimates were detected, with directions varying by software.
• Implemented correction methods in pixy significantly reduced the observed biases.

Conclusions:

• Missing data in population genomics can lead to erroneous evolutionary inferences.
• Accurate handling and correction of missing data are crucial for reliable population genetic studies.
• The updated pixy software provides a valuable tool for mitigating biases in population genetic analyses.