assocModelFitter: R interface for TASSEL's association methods
In maize-genetics/bioconductor_test: R front-end for TASSEL

View source: R/AnalysisAssociationFunctions.R

This function acts as a front-end for TASSEL's extensive association analysis methods. Using this function, users can run the following TASSEL association methods:

best linear unbiased estimates (BLUEs)
generalized linear model (GLM)
mixed linear model
Fast association (Shabalin 2012)

1
2
3

assocModelFitter(tasObj, formula, fitMarkers = FALSE, kinship = NULL,
  fastAssociation = FALSE, minClassSize = 0, biallelicOnly = FALSE,
  appendAddDom = FALSE)

`tasObj`	An object of class `TasselGenotypePenotype`.
`formula`	An R-based linear model formula. The general layout of this formula uses the following TASSEL data scheme: `<data> ~ <factor> and/or <covariate>`. If all traits in a Phenotype object should be ran, a simplified formula (`. ~ .`) can be used. This scheme can also be used for running all `<data>` or `<factor>` and/or `<covariate>` data as well. Single variables are separated witha `+` operator. See vignette for further clarification.
`fitMarkers`	Should marker data be fitted? If `TRUE`, GLM analysis will be executed. If `FALSE`, BLUEs will be calculated. Defaults to `FALSE`.
`kinship`	Should kinship data be accounted for in the model? If so, please submit a TASSEL kinship matrix object using the rTASSEL function, `kinshipMatrix`. Defaults to `NULL`
`fastAssociation`	Should TASSEL's Fast Association plugin be used? Consider setting to `TRUE` if you have many phenotypes in your data set.
`minClassSize`	The minimum acceptable genotype class size. Genotypes in a class with a smaller size will be set to missing. Defaults to 0.
`biallelicOnly`	Only test sites that are bi-allelic. The alternative is to test sites with two or more alleles. Defaults to `FALSE`
`appendAddDom`	If true, additive and dominance effect estimates will be added to the stats report for bi-allelic sites only. The effect will only be estimated when the data source is genotype (not a probability). The additive effect will always be non-negative. Defaults to `FALSE`