MACH

MACH performs genotype imputation and haplotyping using a Markov Chain framework to infer unobserved genotypes and haplotypes and improve variant resolution for genome-wide association studies (GWAS).

Key Features:

• Markov Chain Framework: Models sampled chromosomes as mosaics of each other to enable efficient inference of unobserved genotypes and haplotypes.
• Genotype Imputation and Haplotyping: Resolves long haplotypes and infers missing genotypes in samples from unrelated individuals.
• Quality Measures: Reports measures of the quality of genotype and haplotype estimates.
• Cross-Study Comparisons and Meta-Analyses: Produces consistent imputation results to facilitate comparison across studies and GWAS meta-analyses.
• Accuracy and Utility Evaluation: Accuracy and utility have been evaluated via simulations and experimental genotypes while considering genotyping panels and reference panel configurations.
• Increased Power in Genetic Association Studies: Imputation enhances the power to detect genetic associations in association studies.
• Applicability Across Populations: Applicable to diverse population datasets.
• Advancements with Larger Reference Panels and Sequencing Technologies: Benefits from larger HapMap reference panels and whole genome shotgun sequencing technologies to improve accuracy for unobserved variants.

Scientific Applications:

• GWAS variant discovery: Imputes genotypes to improve detection of common variant associations in complex disease studies.
• Cross-study meta-analysis: Harmonizes variant representation across datasets to enable meta-analyses of GWAS.
• Evaluation of genotyping and sequencing designs: Assesses the impact of genotyping panels, reference panel configurations, and designs that replace genotyping with shotgun sequencing on imputation accuracy.
• Population genetics analyses: Supports analyses of haplotype structure and variant frequency across diverse populations.

Methodology:

Implements a Markov Chain model that represents sampled chromosomes as mosaics of each other to estimate unobserved genotypes and haplotypes from genotype data and whole-genome shotgun sequence information; accuracy assessed using simulations and experimental genotypes while varying genotyping panels and reference panel configurations.