ASReml

ASReml fits general linear mixed models and estimates variance components by residual maximum likelihood (REML) for analysis of complex experimental, breeding, and spatial datasets.

Key Features:

• General linear mixed model fitting: Fits models that include both fixed and random effects to represent complex experimental structures.
• Variance component estimation: Estimates variance components using residual maximum likelihood (REML) for unbiased component inference.
• Average information algorithm: Uses an average information matrix strategy to compute parameter updates and improve computational performance.
• Fixed and random effects handling: Supports specification and estimation of multiple fixed and random effect terms within models.
• Spatial analysis capabilities: Enables modelling of spatial variation in field experiments to account for spatial correlation.
• Designed experiment support: Accommodates incomplete block designs and other structured experimental layouts.
• Large multi-environment dataset handling: Applied to datasets spanning multiple experiments, years, and locations for multi-environment analyses.

Scientific Applications:

• Agricultural research: Estimation of variance components for analyses such as wheat variety means across numerous experiments and diverse geographical locations.
• Designed experiments: Analysis of incomplete block designs and other structured experimental designs.
• Spatial analysis of field experiments: Accounting for spatial variability and correlation in field trial data.
• Animal and plant sciences studies: Applied in analyses relevant to animal and plant breeding and experimental research.

Methodology:

Fits general linear mixed models with fixed and random effects, estimates variance components via residual maximum likelihood (REML), and uses the average information matrix for parameter estimation.