Estimating Seven Coefficients of Pairwise Relatedness Using Population-Genomic Data

Area of Science:

• Population genetics
• Quantitative genetics
• Genomics

Background:

• Population structure is often assessed using genotypic-correlation coefficients, particularly pairwise relatedness coefficients between individuals.
• While nine pairwise relatedness coefficients exist in general models, they reduce to seven for biallelic loci.
• Estimating six of these seven coefficients empirically has been a significant challenge.

Purpose of the Study:

• To develop and validate a numerical optimization procedure for estimating all seven pairwise relatedness coefficients from population-genomic data.
• To assess the accuracy and potential biases of the new estimation method using simulations.
• To apply the method to real-world data to uncover population structure and relationships in Daphnia pulex.

Main Methods:

• Developed a numerical optimization procedure to estimate seven pairwise relatedness coefficients from population-genomic data.
• Conducted simulations to evaluate the method's unbiasedness and error correlations at varying data coverages (e.g., 3×).
• Applied the developed method to genomic data from four populations of Daphnia pulex.

Main Results:

• The numerical optimization procedure provides nearly unbiased estimates of the seven reduced pairwise relatedness coefficients, even at low genomic coverage.
• Errors in five of the seven estimated coefficients were statistically uncorrelated.
• Application to Daphnia pulex revealed the presence of half-siblings and likely asexual clone mates, with significant negative estimates for relatedness and inbreeding coefficients.

Conclusions:

• The new method successfully estimates pairwise relatedness coefficients from population-genomic data, overcoming previous empirical limitations.
• The findings in Daphnia pulex demonstrate the utility of the method in identifying specific relationships like half-siblings and clones.
• Interpreting genotypic correlations as direct estimates of identity by descent requires caution due to potential complexities and negative coefficient estimates.