MPH (MINQUE for Partitioning Heritability)

MPH performs fast restricted maximum likelihood estimation of genetic (co)variance components for large-scale genome partitioning of quantitative genetic variation.

Key Features:

• Computational Efficiency: MPH utilizes stochastic trace estimators and comprehensive parallel computation to enhance convergence, speed, and memory efficiency compared to tools such as GCTA and LDAK.
• REML Estimation: MPH implements restricted maximum likelihood for estimating genetic (co)variance components in genome partitioning analyses.
• Large-Scale Analysis Capability: MPH handles extensive datasets, including those from farm animal populations, that exceed the practical limits of Haseman-Elston regression and linkage disequilibrium score regression.
• Accuracy: MPH achieves comparable accuracy to widely used tools while significantly improving computational performance.

Scientific Applications:

• Genome Partitioning: MPH partitions quantitative genetic variation across the genome to dissect the genetic architecture of complex traits.
• Farm Animal Genetics: MPH facilitates large-scale, comprehensive analyses of complex genetic architectures in farm animal datasets.
• Genetic (Co)variance Estimation: MPH estimates genetic variances and covariances for quantitative traits in structured populations.

Methodology:

MPH employs restricted maximum likelihood with stochastic trace estimators and parallel computation to estimate genetic (co)variance components.