Marker-based estimation of genetic parameters in genomics

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Linear mixed models (LMMs) are widely used for estimating additive genetic variances and heritability in genomic studies.
• LMM analysis, while less computationally intensive than Bayesian methods, faces limitations with large-scale genomic datasets.

Purpose of the Study:

• To introduce and evaluate the symmetric differences squared (SDS) procedure as a computationally simple alternative to LMMs for large genomic datasets.
• To compare the performance of the SDS method against established LMM-based procedures using simulations and empirical analyses.

Main Methods:

• The study proposes the symmetric differences squared (SDS) procedure, a method based on least squares regression analysis.
• Computer simulations and empirical analyses were conducted to compare SDS with two common LMM-based procedures.

Main Results:

• The SDS method demonstrates lower precision than LMMs for small datasets.
• For large sample sizes, the SDS method's precision improves and becomes comparable to LMMs.
• A key advantage of SDS is its continued scalability and precision with increasing sample sizes, unlike LMMs which become computationally infeasible.

Conclusions:

• The SDS method is a viable alternative to LMMs, particularly for analyzing large-scale genomic data where LMMs face computational challenges.
• SDS offers a computationally efficient and scalable approach for estimating genetic parameters in 'big data' genomics.