TRAP.PARAMS: Get parameters to run scAPAtrap

In BMILAB/scAPAtrap: Identification and Quantification of Alternative Polyadenylation Sites from Single-cell RNA-seq Data

Get parameters to run scAPAtrap

Description

Get parameters to run scAPAtrap

Usage

TRAP.PARAMS()

Details

TRAP.PARAMS returns the global variable of list type to store parameters for running the wrapper function scAPAtrap.

• TenX: Whether the inputBam is 10X-format (e.g., BAM file from CellRanger or STARsolo). For non-10X protocols like CEL-seq, when STARsolo is used, the BAM file will also contain cell barcode tag (CB) and UMI barcode tag (UB), then can set TenX=TRUE.

• barcode: string vector to define barcodes, default is NULL (not provided).

• chrs: a string vector of chromosome names (e.g., chr1, chr2, chr3...), default is NULL. This parameter is used in findPeaksByStrand. If chrs=NULL and run one-step scAPAtrap, then will try to get all chrs from inputBAM. However, when running scAPAtrap, if chrs=NULL and there are more than 100 chrs from inputBAM, users should provide trap.params$chrs explicitly to run scAPAtrap just in case there are too many unuseful chrs to run.

• maxwidth: max width (nt) to define a valid peak, default is 1000. Peaks wider than 'maxwidth' will be automatelly split until meet requirement. This parameter is used in findPeaksByStrand.

• readlength: R2 read length, default is 49. 'readlength' determines the min peak width; peaks wider than 'readlength' will be retained. This parameter is used in findPeaksByStrand.

• cov.cutoff: min read coverage for peak calling, default is 10. Peak regions with coverage greater than 'cov.cutoff'will be considered as candidate peaks. This parameter is used in findPeaksByStrand.

• min.cells: number of cells that a peak is expressed (i.e., with read count over 'min.count'), default is 10. 'min.cells' is used to filter peaks present in at least 'min.cells' cells. This parameter is used in reducePeaks and generatescExpMa.

• min.count: number of read counts a peak has, default is 10. ‘min.count' is used to filter peaks with at least 'min.count’ reads. This parameter is used in reducePeaks and generatescExpMa.

• tails.search: the strategy to search A tails, can be set 'peaks', 'genome', 'no' (default). This parameter is used in the wrapper function scAPAtrap.

  no

  Not search tails, which means the final peaks are not linked with any A-tails.

  genome

  Search tails genome-wide, which may be suitable for small genome. Otherwise it will take very long time or MEM.

  peaks

  Search tails within the peak regions by 'd' nt. This way may be suitable for big genome but with relatively small number of peaks.

• d: min distance from a peak to any tail to call a peak as real PA, default is 100. Only applicable when tails.search is peaks or genome. This parameter is used in findTailsByPeaks and generatescExpMa. If provided, then the final scAPAtrapData (the PA list) will remove cells not in 'barcode.'

• findUniqueMap: default is TRUE, means to run findUniqueMap to get unique mappings, and/or sort/index the inputBam.

• findUniqueMap.sort: default is TRUE, to sort inputBam. Only applicable when findUniqueMap=TRUE.

• findUniqueMap.index: default is TRUE, to index inputBam. Only applicable when findUniqueMap=TRUE.

• thread: Number of CPU threads, default is 24.

Value

A list of parameters with default values for running scAPAtrap.

Examples

## see TRAP.PARAMS
TRAP.PARAMS()
## set up new trap.params
trap.params=TRAP.PARAMS()
trap.params$TenX=FALSE
## set up from scratch
trap.params=list(thread=24,
                 TenX=TRUE,
                 chrs=c('2L','2R','3L','3R','4','X','Y'),
                 maxwidth=1000,
                 readlength=49,
                 cov.cutoff=10,
                 min.cells=10,
                 min.count=10,
                 tails.search='peaks',
                 d=100,
                 barcode=NULL,
                 findUniqueMap=TRUE,
                 findUniqueMap.sort=TRUE,
                 findUniqueMap.index=TRUE )

BMILAB/scAPAtrap documentation built on Oct. 13, 2023, 2:36 a.m.