movingGrid {mvngGrAd}

Moving Grid Adjustment In Plant Breeding Field Trials

Description

The function uses the phenotypic information from plants or plots (entries) in a designed grid to obtain an adjusted (for environmental variation) phenotypic value of the entry in the center of the grid. The adjustment is done by calculating the mean of all the entries included and using it as a covariate.

Usage

movingGrid(rows, columns, obsPhe, shapeCross, layers, excludeCenter = TRUE)

Arguments

rows

An integer vector with the row subscripts of the phenotypic values to be adjusted.

columns

An integer vector with the column subscripts of the phenotypic values to be adjusted.

obsPhe

A numeric vector with the observed phenotypic values of the entries that are to be adjusted. The arguments to rows, columns and obsPhe must, of course, correspond.

shapeCross

"list" of length 4, each element contains the cells that are to be included in the grid that extends from the center in 0, 90, 180 and 270 degree direction.

Elements can be NULL (no extension in this direction) and must contain no '0', duplicated or negative values.

shapeCross[[1]]DOWN (180 degree) from the center cell (same column)

shapeCross[[2]]UP (0 degree) from the center cell (same column)

shapeCross[[3]]LEFT (270 degree) from the center cell (same row)

shapeCross[[4]]RIGHT (90 degree) from the center cell (same row)

layers

"integer", vector giving the layers to be included in the grid that extents from the center cell in all directions other than 0, 90, 180 and 270 degree. Can be NULL ( no extension) and must not contain '0', negative or duplicated values.

excludeCenter

Should the center entry (the one which is adjusted) be excluded from the calculation of the moving mean, logical TRUE (default) or FALSE.

Details

Please see the included vignette for a more detailed description of the rational underlining the adjustment procedure and usage of the function.

The function movingGrid calls the functions extendHorVer and extendCircular to form the grid (i.e. determine the row and column subscripts of the cells included). These two functions are not intended to be called by the user.

Value

An object of class "movG". See movG-class for details.

Rdversion

1.1

Note

The field layout is always assumed to be a rectangular matrix with max(rows) * max(columns) cells! All of the cells that were not mentioned in arguments rows and columns are NA.

If, for example, the last row was only planted with half the number of entries as the other rows, the other half of the row will consist of cells with NA values.

These NA values are ignored by movingGrid and the various extractor functions. They do not influence the results of movingGrid.

The general linear model, that is part of the adjustment procedure, is fit via the function lm.

Author(s)

Frank Technow

See Also

sketchGrid, entryData, movG, mvngGrAd-package

Examples

### data creation ###
#####################


### The field plan consists of 50 rows and 50 columns
### (= 2500 cells = 2500 entries)

## row vector

rows <- rep(1:50,each=50)


## column vector
cols <- rep(1:50,50)

## simulation of growing conditions with a horizontal gradient between
## the rows and some random noise within

set.seed(666)
growingCond <- rep(1:50,each=50) + rnorm(2500)

## simulation of the genotypic effects
gEffects <- rnorm(2500,mean=0,sd=5)

## observed phenotypic values
yield <- growingCond + gEffects

## adjustment with movingGrid

cross <- list(c(1),
              c(1),
              c(1:4),
              c(1:4))

exampleResults <- movingGrid(rows = rows,
                             columns = cols,
                             obsPhe = yield,
                             shapeCross = cross,
                             layers = c(1,2),
                             excludeCenter = TRUE)

## a summary

summary(exampleResults)

## the adjusted phenotypic values of the first 25 entries

fitted(exampleResults)[1:25]

## all information on entry, for the first 25

entryData(exampleResults)[1:25,]

## moving means for the first 25

movingMean(exampleResults)[1:25]