SEconstructFromCPM: Function to accept count data matrix and produce Z-scores,...
In ATpoint/misterplotR: Functions for plotting and clustering of count data

Description Usage Arguments Details Author(s) Examples

View source: R/src_SEconstructFromCPM.R

The function accepts a standard numeric count matrix or data frame. Will then optionally filter these counts either for overlap with a provided GRanges object (Type = "ranged") or for overlap with a list of gene or region names (Type = "unranged"). Function can optionally filter for certain samples or average samples by groups depending on the GroupIndicator which could e.g. be "_rep", so WT_rep1, WT_rep2, KO_rep1, KO_rep2 => WT and KO were the groups here. Eventually will return the averaged counts (or optionally Z-scores) as individual assays in a (R)SE object.

SEconstructFromCPM(
  InputData,
  Type,
  Reference.Regions = NULL,
  Reference.Genes = NULL,
  GroupIndicator = NULL,
  GroupFilter = NULL,
  SampleFilter = NULL,
  transform.log2 = FALSE,
  AverageByGroup = FALSE,
  Zscore = TRUE
)

`InputData`	a matrix or dataframe with the count data to be plotted
`Type`	"ranged" for GRanges input or "unranged" when a "Gene" column is present to be used as rowData
`Reference.Regions`	GRanges object to use for overlap filtering
`Reference.Genes`	Gene (or any name) list to filter for overlaps
`GroupIndicator`	the suffix that indicates replicates, e.g. "_rep"
`GroupFilter`	a list of group names to be retained
`SampleFilter`	a list of samples to be retained
`transform.log2`	logical, whether to transform counts to log2+1
`AverageByGroup`	logical, whether to average samples by group
`Zscore`	logical, whether to calculate Z score and return as assays(SE)$Z

(tba...)

Alexander Toenges

## Unranged:
cts <- sapply(seq(1,10), function(x) rnorm(1000,100))
colnames(cts) <- paste0("sample", "_rep", seq(1,10))
cts <- data.frame(Gene=paste0("gene", seq(1,1000)), cts)
SE <- SEconstructFromCPM(InputData = cts, Type = "unranged", AverageByGroup = TRUE, GroupIndicator = "_rep")
## Ranged:
cts1 <- sapply(seq(1,10), function(x) rnorm(1000,100))
Gr <- GRanges(seqnames = rep("chr1", 1000), ranges = IRanges(start = seq(1,250), end = rep(1001, 1000)))
elementMetadata(Gr) <- cts1
RSE <- SEconstructFromCPM(InputData = Gr, Type = "ranged", AverageByGroup = TRUE, GroupIndicator = "_rep")