Evaluating the concordance between sequencing, imputation and microarray genotype calls in the GAW18 data.

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Genotype errors are known to affect the accuracy of genotype-phenotype association studies, potentially increasing false positives or reducing statistical power.
• The Genetic Analysis Workshop 18 (GAW18) dataset lacks a gold standard for genotype calls, necessitating an assessment of data quality.

Purpose of the Study:

• To evaluate the concordance between different genotype calling platforms (sequencing, imputation, microarray) within the GAW18 dataset.
• To identify potential sources and patterns of genotype errors, including missing data and discordance rates.
• To understand the implications of these errors for downstream genetic analyses.

Main Methods:

• Comparative analysis of genotype calls from sequencing, imputation, and microarray data.
• Calculation of missing data rates and genotype concordance rates between platforms.
• Examination of discordance patterns based on minor allele frequency (MAF) and phenotype association.

Main Results:

• High missing data rates were observed for sequenced individuals.
• A modest level of genotype discordance was found between sequencing and imputation platforms, most prevalent in low MAF single-nucleotide polymorphisms (SNPs).
• Some evidence suggested phenotype-specific differences in discordance rates, with instances of conflicting base calls across technologies.

Conclusions:

• Missing genotypes and errors in called genotypes pose a risk of increased Type I errors and reduced power in the downstream analysis of GAW18 data.
• The observed discordance highlights the importance of careful genotype quality control in genetic association studies.
• Findings underscore the need for robust methods to handle genotype uncertainty in large-scale genetic datasets.