updog

updog infers genotypes in polyploid organisms from next-generation sequencing (NGS) read counts using empirical Bayes methods that account for allelic bias, overdispersion, and sequencing errors.

Key Features:

• Empirical Bayes Approaches: Implements empirical Bayes estimation to genotype polyploids without requiring prior specification of technical artifacts.
• Flexible Genotype Distributions (flexdog()): Allows specification of a wide range of genotype distributions via the flexdog() function to accommodate diverse genetic structures in polyploids.
• Intermediate Flexibility Priors: Provides proportional normal and unimodal prior genotype distribution classes to balance flexibility and regularization in prior specification.
• Optimization and Characterization: Includes optimization methods and characterization for the class of unimodal prior distributions to improve genotype inference.
• Simulation Functions (rgeno(), rflexdog()): Supplies rgeno() to simulate genotypes and rflexdog() to simulate read counts for validation and testing.
• Error Rate and Correlation Calculations (oracle_mis(), oracle_cor()): Offers oracle_mis() to compute oracle genotyping error rates and oracle_cor() to compute correlation with true genotypes for performance assessment.
• Relatedness Consideration (mupdog()): Includes an experimental mupdog() function to model varying levels of relatedness among samples.

Scientific Applications:

• Genotyping polyploids: Accurate inference of allele copy number and genotypes in polyploid organisms from NGS data.
• Plant and animal genetics: Analysis of genetic variation and inheritance patterns in polyploid species used in plant and animal genetics research.
• Evolutionary biology: Investigation of genetic diversity and population structure in polyploid lineages for evolutionary studies.
• Conservation studies: Assessment of genetic diversity and structure in polyploid populations relevant to conservation genetics.

Methodology:

Uses empirical Bayes estimation with proportional normal and unimodal prior classes, optimization routines for unimodal priors, simulation via rgeno() and rflexdog(), oracle error and correlation calculations via oracle_mis() and oracle_cor(), and an experimental mupdog() to account for relatedness.