filterTaxonMatrix: Filter taxa in an abundance matrix
In ConesaLab/MDM: Complete MDM Network Analysis for 16S rRNA Data

Description Usage Arguments Value Examples

function from seqtime package, Discard taxa with less than the given minimum number of occurrences

1 2	filterTaxonMatrix(x,minocc = 0, dependency = FALSE, keepSum = FALSE, return.filtered.indices = FALSE )

`x`	taxon abundance matrix, rows are taxa, columns are samples
`minocc`	minimum occurrence (minimum number of samples w/ non-zero taxon abundance)
`dependency`	if true, remove all taxa with a slope above -0.5 or a non-linear slope in the periodogram in log-scale (samples are supposed to represent equidistant time points)
`keepSum`	if keepSum is true, the discarded rows are summed and the sum is added as a row with name: summed-nonfeat-rows
`return.filtered.indices`	if true, return an object with the filtered abundance matrix in mat and the indices of removed taxa in the original matrix in filtered.indices

filtered abundance matrix

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (x, minocc=0,dependency = FALSE, keepSum = FALSE,
    return.filtered.indices = FALSE)
{
    toFilter = c()
    xcopy = x
    xcopy[xcopy > 0] = 1
    rowsums = apply(xcopy, 1, sum)
    toFilter = which(rowsums < minocc)
    if (dependency == TRUE) {
        nt = identifyNoisetypes(x, epsilon = 0.5)
        toKeep = c(nt$pink, nt$brown, nt$black)
        toFilter = c(toFilter, setdiff(c(1:nrow(x)), toKeep))
    }
    indices.tokeep = setdiff(c(1:nrow(x)), toFilter)
    if (keepSum == TRUE) {
        filtered = x[toFilter, ]
        x = x[indices.tokeep, ]
        rownames = rownames(x)
        sums.filtered = apply(filtered, 2, sum)
        x = rbind(x, sums.filtered)
        rownames = append(rownames, "summed-nonfeat-rows")
        rownames(x) = rownames
    }
    else {
        x = x[indices.tokeep, ]
    }
    if (return.filtered.indices == TRUE) {
        res = list(x, toFilter)
        names(res) = c("mat", "filtered.indices")
        return(res)
    }
    else {
        return(x)
    }
}