Estimation of kinship coefficient in structured and admixed populations using sparse sequencing data.

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Biological relatedness estimation is vital for genetic studies, including human genetic association studies.
• Accurate kinship estimation helps control for population structure and estimate trait heritability.
• Existing kinship estimation methods struggle with sparse sequencing data due to genotype uncertainty and missing data.

Purpose of the Study:

• To develop a computationally efficient method for kinship estimation using sparse sequencing data.
• To address the downward bias in kinship estimation caused by genotype uncertainty in existing methods.
• To improve the accuracy of kinship estimation in human genetic studies with population structure and admixture.

Main Methods:

• Developed SEEKIN, a novel method that models genotype uncertainty and utilizes linkage disequilibrium through imputation.
• Applied SEEKIN to whole exome sequencing (WES) data from Singaporean Chinese and Malay populations.
• Tested SEEKIN on down-sampled (0.15X) and full WES data (0.75X) to assess performance with shallow sequencing data.

Main Results:

• SEEKIN accurately estimates kinship coefficients and classifies genetic relatedness from sparse off-target sequencing data.
• The method demonstrates superior performance compared to existing methods when analyzing shallow off-target data.
• SEEKIN improves the estimation of trait heritability in WES studies using both simulated and real phenotypes.

Conclusions:

• SEEKIN provides a robust solution for kinship estimation from sparse sequencing data, overcoming limitations of existing methods.
• The method's ability to model genotype uncertainty enhances the accuracy of genetic relatedness assessments.
• Accurate kinship estimation using SEEKIN has significant implications for improving heritability studies in large-scale human genetic research.