stepwiseqtlF: Stepwise selection for multiple QTL in function valued trait...

In ikwak2/funqtl: QTL Mapping for Function-Valued Traits

Stepwise selection for multiple QTL in function valued trait data

Description

Extension of the R/qtl function stepwiseqtl. Performs forward/backward selection to identify a multiple QTL model for function valued trait data, with model choice made via a penalized LOD score, with separate penalties on main effects and interactions.

Usage

stepwiseqtlF(
  cross,
  chr,
  pheno.cols,
  qtl,
  usec = c("slod", "mlod"),
  formula,
  max.qtl = 10,
  covar = NULL,
  method = c("hk", "imp"),
  incl.markers = TRUE,
  refine.locations = TRUE,
  additive.only = TRUE,
  penalties,
  keeptrace = FALSE,
  verbose = TRUE
)

Arguments

cross	An object of class "cross". See read.cross for details.
chr	Optional vector indicating the chromosomes to consider in search for QTL. This should be a vector of character strings referring to chromosomes by name; numeric values are converted to strings. Refer to chromosomes with a preceding "-" to have all chromosomes but those considered. A logical (TRUE/FALSE) vector may also be used.
pheno.cols	Columns in the phenotype matrix to be used as the phenotype.
qtl	Optional QTL object (of class "qtl", as created by makeqtl) to use as a starting point.
usec	Which method to use ("slod" or "mlod")
formula	Optional formula to define the QTL model to be used as a starting point.
max.qtl	Maximum number of QTL to which forward selection should proceed.
covar	Data frame of additive covariates.
method	Indicates whether to use multiple imputation or Haley-Knott regression.
incl.markers	If FALSE, do calculations only at points on an evenly spaced grid.
refine.locations	If TRUE, use refineqtlF to refine the QTL locations after each step of forward and backward selection.
additive.only	If TRUE, allow only additive QTL models; if FALSE, consider also pairwise interactions among QTL.
penalties	Vector of three values indicating the penalty on main effects and heavy and light penalties on interactions. See the Details below. If missing, default values are used that are based on simulations of backcrosses and intercrosses with genomes modeled after that of the mouse.
keeptrace	If TRUE, keep information on the sequence of models visited through the course of forward and backward selection as an attribute to the output.
verbose	If TRUE, give feedback about progress. If verbose is an integer > 1, even more information is printed.

Value

The output is a representation of the best model, as measured by the penalized LOD score (see Details), among all models visited. This is QTL object (of class "qtl", as produced by makeqtl), with attributes "formula", indicating the model formula, and "pLOD" indicating the penalized LOD score.

If keeptrace=TRUE, the output will contain an attribute "trace" containing information on the best model at each step of forward and backward elimination. This is a list of objects of class "compactqtl", which is similar to a QTL object (as produced by makeqtl) but containing just a vector of chromosome IDs and positions for the QTL. Each will also have attributes "formula" (containing the model formula) and "pLOD" (containing the penalized LOD score.

Author(s)

Il-Youp Kwak, <email: [e-mail]>

References

Manichaikul, A., Moon, J. Y., Sen, S, Yandell, B. S. and Broman, K. W. (2009) A model selection approach for the identification of quantitative trait loci in experimental crosses, allowing epistasis. _Genetics_, *181*, 1077-1086.

Broman, K. W. and Speed, T. P. (2002) A model selection approach for the identification of quantitative trait loci in experimental crosses (with discussion). _J Roy Stat Soc B_ *64*, 641-656, 731-775.

Haley, C. S. and Knott, S. A. (1992) A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. _Heredity_ *69*, 315-324.

Zeng, Z.-B., Kao, C.-H. and Basten, C. J. (1999) Estimating the genetic architecture of quantitative traits. _Genetical Research_, *74*, 279-289.

See Also

refineqtlF, addqtlF

Examples

data(simspal)

# Genotype probabilities for H-K
simspal <- calc.genoprob(simspal, step=0)
phe <- 1:nphe(simspal)

qtlslod <- stepwiseqtlF(simspal, pheno.cols = phe, max.qtl = 4, usec = "slod",
                        method = "hk", penalties = c(2.36, 2.76, 2) )

ikwak2/funqtl documentation built on April 20, 2022, 3:58 a.m.