Commonly used Hardy-Weinberg equilibrium filtering schemes impact population structure inferences using RADseq data

Area of Science:

• Population genetics
• Genomics
• Bioinformatics

Background:

• Reduced representation sequencing (RRS) and restriction site associated DNA sequencing (RADseq) are vital for studying genetic diversity, especially in non-model organisms.
• Bioinformatic filters are essential for RADseq data quality but can influence population genetic inferences.
• Hardy-Weinberg equilibrium (HWE) filtering is a common but often poorly described practice.

Purpose of the Study:

• To investigate the impact of Hardy-Weinberg equilibrium (HWE) filtering on population genetic inference from RADseq data.
• To highlight the lack of detailed reporting for HWE filtering methods in published studies.
• To provide recommendations for best practices in HWE filtering for RADseq datasets.

Main Methods:

• Analysis of in silico and empirical RADseq datasets.
• Evaluation of commonly used Hardy-Weinberg equilibrium (HWE) filtering approaches.
• Assessment of the impact of HWE filtering on population structure inference.

Main Results:

• Many HWE filtering approaches lack sufficient detail for replication.
• Specific HWE filtering methods, particularly pooling across samples, significantly reduce inferred population structure.
• The choice of HWE filtering critically affects population genetic analyses.

Conclusions:

• Standard HWE filtering practices can obscure true population structure in RADseq data.
• Researchers must carefully consider and report HWE filtering methods for reproducible population genetic studies.
• Best practices for HWE filtering are needed to ensure reliable inferences from RADseq data.