Conduct a Scanonevar Using the DGLM Function

Share:

Description

scan.via.dglm should not typically be called by a user. This function is used by both scanonevar and scanonevar.perm. This function is not typically necessary for a typical user.

Usage

1
2
3
4
scan.via.dglm(mean.alt.formula, var.alt.formula, genoprobs, mapping.df,
  chr.by.marker, pos.by.marker, marker.names, return.effects = FALSE,
  return.effect.ses = FALSE, return.effect.ps = FALSE,
  cor.threshold = 0.8, perm = 1:nrow(genoprobs), family = "gaussian")

Arguments

mean.alt.formula

The formula for the trait mean in the alternative model. mean.null.formula and test.mean.effect are inferred from it.

var.alt.formula

The formula for the trait variance in the alternative model. var.null.formula and test.var.effect are inferred from it.

genoprobs

The probability of each genotype for each individual.

mapping.df

The tbl_df with the response, all covariates, and space for the focal genotype.

chr.by.marker

a vector of the chromosome name of each marker

pos.by.marker

a vector of the position of each marker

marker.names

a vector of the name of each marker

return.effects

Logical indicating whether the estimated effects should be returned.

return.effect.ses

Logical indicating whether the standard errors of the estimated effects should be returned.

return.effect.ps

Logical indicating whether the p-value of the estimated effects should be returned.

cor.threshold

Numeric between 0 and 1 indicating how tightly a locus must be correlated with a covariate to be skipped. e.g. if cor.threshold is 0.8 (it's default) any locus with cor(locus, covariate) > 0.8 will be skipped.

perm

The permutation to apply to the genotypes. Defaults to identity permutation.

family

Family of distribution to be modeled. Defaults to 'gaussian'. See dglm and glm documentation for other options. Most notable other options is 'poisson'

Details

none

Value

Returns a scanonevar object.

Author(s)

Robert Corty rcorty@gmail.com

See Also

scanonevar, scanonevar.perm