makeStepList: Construction of a stepwise normalization list
In stepNorm: Stepwise normalization functions for cDNA microarrays

Description Usage Arguments Details Value Author(s) See Also Examples

This function provides a user friendly way to construct a list for input to the function stepWithinNorm. The list indicates intended biases for correction and models for stepwise normalization.

1
2
3

  makeStepList(A = c("median", "rlm", "loess"), PT = c("median", "rlm",
"loess"), PL = c("median", "rlm", "loess"), Spatial2D = c("rlm2D",
"loess2D", "aov2D", "spatialMedian"))

`A`	A character string specifying the normalization models for the adjustment of intensity or A bias: median: global median location normalization rlm: global intensity or A-dependent robust linear normalization using the `rlm` function loess: global intensity or A-dependent robust nonlinear normalization using the `loess` function The user can specify any of these three choices and the selected model will be compared based the goodness fit and model parsimony; If the correction of the $A$ bias is not desired, the user can set `A = NULL`.
`PT`	A character string specifying the normalization models for the adjustment of print-tip or PT bias: median: within-print-tip-group median normalization rlm: within-print-tip-group robust linear normalization using the `rlm` function loess: within-print-tip-group robust nonlinear normalization using the `loess` function none: no normalization for the PT bias If the correction of the $PT$ bias is not desired, the user can set `PT = NULL`.
`PL`	A character string specifying the normalization models for the adjustment of well-plate or PL bias: median: within-well-plate median normalization rlm: within-well-plate robust linear normalization using the `rlm` function loess: within-well-plate robust nonlinear normalization using the `loess` function none: no normalization for the PL bias If the correction of the $PL$ bias is not desired, the user can set `PL = NULL`.
`Spatial2D`	A character string specifying the normalization models for the adjustment of spatial 2D bias: none: no normalization for the spatial 2D bias aov2D: spatial bivariate location normalization using ANOVA rlm2D: spatial bivariate location normalization using the `rlm` function loess2D: spatial bivariate location normalization using the `loess` function spatialMedian: spatial location normalization using a spatial median approach (see Wilson et al. (2003) in reference) If the correction of the $PL$ bias is not desired, the user can set `Spatial2D = NULL`.

This function provides a user friendly way to specify the parameter wf.loc for the main stepwise normalization function stepWithinNorm; see examples for details.

An object of class "list" for input to the stepWithinNorm function.

Yuanyuan Xiao, yxiao@itsa.ucsf.edu,
Jean Yee Hwa Yang, jean@biostat.ucsf.edu

stepWithinNorm.

# Examples use swirl dataset, for description type ? swirl
data(swirl)

# To use the default parameters, which adjusts A, PT, PL and Spatial2D
# biases sequentially and compares all models available, simple type

wf.loc <- makeStepList()

# To apply loess for the A bias, and to omit the Spatial2D step
wf.loc <- makeStepList(A=("loess"), Spatial2D=NULL)

# To compare only rlm and loess in the A bias step, and other biases as default
wf.loc <- makeStepList(A=c("rlm","loess"))

# input to the stepWithinNorm function
## Not run: 
step.swirl1 <- stepWithinNorm(swirl[,1],wf.loc=wf.loc)
## End(Not run)