R: Check and preprocess an allele dataset

preprocessData {PolyPatEx}

R Documentation

Check and preprocess an allele dataset

Description

Check and preprocess the input allele data frame prior to subsequent analysis.

Usage

preprocessData(adata, numLoci, ploidy, dataType, dioecious,
  selfCompatible = NULL, mothersOnly = NULL, lociMin = 1,
  matMismatches = 0)

Arguments

`adata`	data frame: an allele dataset.
`numLoci`	integer: the number of loci in the allele dataset.
`ploidy`	integer: the species' ploidy, one of `2`, `4`, `6`, or `8`.
`dataType`	character: either `"genotype"` or `"phenotype"`.
`dioecious`	logical: is the species dioecious or monoecious?
`selfCompatible`	logical: In monoecious species (`dioecious=FALSE`), can individuals self-fertilise? When `dioecious=FALSE`, this argument may be left at its default value of NULL - it will be set to `FALSE` by `preprocessData`.
`mothersOnly`	logical: in dioecious species, should females without progeny present be removed from the dataset? If `dioecious=TRUE`, then `mothersOnly` must be set to either `TRUE` or `FALSE`. If `dioecious=FALSE`, argument `mothersOnly` should be left at its default value of `NULL`.
`lociMin`	integer: the minimum number of loci in a individual that must have alleles present for the individual (and its progeny, if any) to be retained in the dataset (default 1).
`matMismatches`	an integer between 0 and `numLoci`-1, being the maximum number of mismatching alleles between mother and offspring that are allowed before the offspring is removed from the dataset. The default value is 0. If an offspring has fewer than `matMismatches` loci that mismatch with its mother, the offending loci are set to contain no alleles.

Details

If inputData is used to load the allele data set into R, then preprocessData will be called automatically on the data frame before it is returned by inputData. However, if the user loads their data into R by some means other than inputData, then preprocessData MUST be called on the data frame prior to using any other PolyPatEx functions to analyse the allele data—preprocessData performs a series of critical checks and preprocessing steps on the data frame, without which other analysis functions in PolyPatEx will fail.

Note that inputData strips leading or trailing spaces (whitespace) from each entry in the allele dataset as it is read in. If you load your data by a means other than inputData, you should ensure that you perform this step yourself, as preprocessData will not carry out this necessary step.

Note also that you should not use spaces in any of your allele codes - PolyPatEx functions use spaces to separate allele codes as they process the data - if allele codes already contains spaces, errors will occur in this processing. If you need a separator, I recommend using either ‘code.’ (a period) or ‘code_’ (an underscore) rather than a space.

preprocessData first performs a number of simple checks on the format and validity of the data set. These checks look for the presence of certain required columns and correct naming and content of these columns. preprocessData will usually stop with an error message should the data fail these basic checks. Correct the indicated problem in the CSV file or R allele data frame, then call inputData or preprocessData again as appropriate. If you use a spreadsheet to edit the CSV file, don't forget that you may also need to call fixCSV on the CSV file, prior to calling inputData again.

If the data is ‘genotypic’ data PolyPatEx requires that all p alleles must be present in each allele set, where p is the species' ploidy. If an allele set contains fewer than p alleles, then it is reset to contain no alleles and is subsequently ignored by other PolyPatEx functions. ID and locus information is printed to the R terminal, to help the user locate these cases in their original dataset.

Further checks look for mismatches between progeny and their mothers' allele sets at each locus—these are situations where a progeny's allele set could not have arisen from any gamete that the mother can provide. When no more than matMismatches mismatching loci occurs in a mother-progeny pair, the offending allele sets in the progeny are reset to contain no alleles (we term these ‘missing’ allele sets). When mismatches occur in more than matMismatches loci, the progeny is removed entirely from the dataset. Information is printed to the R terminal to assist the user in identifying the affected individuals and loci—in particular, note that removal of several (or all) of a single mother's progeny may indicate an error in the mother's allele data, rather than in her progeny.

After the mother/progeny mismatch check above, a subsequent check removes individuals from the dataset that have fewer than lociMin non-missing allele sets remaining. The default value for lociMin is 1—an individual must have at least one non-missing locus to remain in the dataset. If any mothers are removed from the dataset at this stage, all of her progeny are removed also. Again, information about these removals is printed to the R terminal.

Note that in the data frame that is returned by preprocessData, the alleles in each allele set (i.e, corresponding to each locus) will be sorted into alphanumeric order—this sort order is necessary for the correct operation of other PolyPatEx routines, and should not be altered.

PolyPatEx needs to know the characteristics of the dataset being analysed. These are specified in the inputData or preprocessData calls and are invisibly attached to the allele data frame that is returned, for use by other PolyPatEx functions. The required characteristics are:

numLoci: the number of loci in the dataset
ploidy: the ploidy (p) of the species (currently allowed to be 4, 6, or 8. ploidy can also be 2, provided dataType="genotype")
dataType: whether the data is genotypic (all p alleles at each locus are observed) or phenotypic (only the distinct allele states at a locus are observed - alleles that appear more than once in the genotype of a locus only appear once in the phenotype)
dioecious: whether the species is dioecious or monoecious
selfCompatible: whether a monoecious species is self compatible (i.e., whether an individual can fertilise itself)
mothersOnly: whether a dioecious dataset should retain only adult females that are mothers of progeny in the dataset.

Value

A data frame, containing the checked and pre-processed allele data, ready for further analysis by other PolyPatEx functions. All columns in the returned data frame will be of mode character.

Author(s)

Alexander Zwart (alec.zwart at csiro.au)

Examples

## If using inputData to input the allele dataset from CSV file,
## preprocessData() is applied automatically before the dataset is
## returned by inputData().

## Otherwise, if the allele dataset is created or loaded into R
## by other means, such preprocessData() must be applied before
## other PolyPatEx analysis routines are applied:

## Using the example dataset 'GF_Phenotype':
data(GF_Phenotype)

## Since we did not load this dataset using inputData(), we must
## first process it with preprocessData() before doing anything
## else:
pData <- preprocessData(GF_Phenotype,
                        numLoci=7,
                        ploidy=6,
                        dataType="phenotype",
                        dioecious=FALSE,
                        selfCompatible=FALSE)

head(pData)  ## Checked and Cleaned version of GF_Phenotype

## pData is now ready to be passed to other PolyPatEx analysis
## functions...

[Package PolyPatEx version 0.9.2 Index]