getFeatureCounts: Get counts of annotation within a defined window around each...

In hiAnnotator: Functions for annotating GRanges objects

Description

Given a query object and window size(s), the function finds all the rows in subject which are <= window size/2 distance away. If weights are assigned to each positions in the subject, then tallied counts are multiplied accordingly. For large annotations, use getFeatureCountsBig.

Usage

getFeatureCounts(sites.rd, features.rd, colnam = NULL,
  chromSizes = NULL, widths = c(1000, 10000, 1e+06),
  weightsColname = NULL, doInChunks = FALSE, chunkSize = 10000,
  parallel = FALSE)

Arguments

sites.rd	GRanges object to be used as the query.
features.rd	GRanges object to be used as the subject or the annotation table.
colnam	column name to be added to sites.rd for the newly calculated annotation...serves as a prefix to windows sizes!
chromSizes	named vector of chromosome/seqnames sizes to be used for testing if a position is off the mappable region. DEPRECATED and will be removed in future release.
widths	a named/numeric vector of window sizes to be used for casting a net around each position. Default: c(1000,10000,1000000).
weightsColname	if defined, weigh each row from features.rd when tallying up the counts.
doInChunks	break up sites.rd into small pieces of chunkSize to perform the calculations. Default is FALSE. Useful if you are expecting to find great deal of overlap between sites.rd and features.rd.
chunkSize	number of rows to use per chunk of sites.rd. Default to 10000. Only used if doInChunks=TRUE.
parallel	use parallel backend to perform calculation with foreach. Defaults to FALSE. If no parallel backend is registered, then a serial version of foreach is ran using registerDoSEQ.

Value

a GRanges object with new annotation columns appended at the end of sites.rd. There will be a column for each width defined in widths parameter. If widths was a named vector i.e. c("100bp"=100,"1K"=1000), then the colname parameter will be pasted together with width name else default name will be generated by the function.

Note

• If parallel=TRUE, then be sure to have a parallel backend registered before running the function. One can use any of the following libraries compatible with foreach: doMC, doSMP, doSNOW, doMPI. For example: library(doMC); registerDoMC(2)

See Also

makeGRanges, getNearestFeature, getSitesInFeature, getFeatureCountsBig.

Examples

# Convert a dataframe to GRanges object
data(sites)
alldata.rd <- makeGRanges(sites, soloStart = TRUE)

data(genes)
genes.rd <- makeGRanges(genes)

geneCounts <- getFeatureCounts(alldata.rd, genes.rd, "NumOfGene")
## Not run: 
geneCounts <- getFeatureCounts(alldata.rd, genes.rd, "NumOfGene",
doInChunks = TRUE, chunkSize = 200)
geneCounts
## Parallel version of getFeatureCounts
# geneCounts <- getFeatureCounts(alldata.rd, genes.rd, "NumOfGene",
parallel = TRUE)
# geneCounts

## End(Not run)

hiAnnotator documentation built on Nov. 8, 2020, 7:27 p.m.