isCell: isCell

In plantformatics/Socrates: Analysis of scATAC-seq data

isCell

Description

This function compares each barcode to obj background and cell bulk reference to identify barcodes representing ambient DNA or broken nuclei. Three columns are added to the metadata: background, cellbulk, is_cell, which reflect the correlation of the cell's chromatin profile with the background reference, the correlation with the predicted cellbulk reference, and whether a barcode is predicted to be a cell (1 = cell, 0 = background noise).

Usage

isCell(
  obj,
  num.test = 20000,
  num.tn5 = NULL,
  num.ref = 1000,
  background.cutoff = 100,
  min.pTSS = 0.2,
  min.FRiP = 0.2,
  min.pTSS.z = -2,
  min.FRiP.z = -2,
  verbose = F
)

Arguments

obj	Object output from generateMatrix. Requires counts and meta data (buildMetaData) slots populated.
num.test	Number of barcodes to query (ranked by total # of unique Tn5 insertions). Set to NULL to select barcodes to test using obj minimum total number Tn5 insertions. Defaults to 20,000
num.tn5	Set the minimum number of Tn5 insertions to select test cells. Overridden by num.test. Defaults to NULL.
num.ref	Number of cells to use as the cell bulk reference (top X cells based on # unique Tn5 insertions)
background.cutoff	Maximum unique Tn5 insertions to use for selecting barcodes for the background reference set
min.pTSS	Minimum per cent TSS for good cells. Defaults to 0.2.
min.FRiP	Minimum fraction reads in peaks for good cells. Defaults to 0.2.
min.pTSS.z	Minimum z-score from per cent TSS for good cells. Defaults to -2.
min.FRiP.z	Minimum z-score from fraction reads in peaks for good cells. Defaults to -2.
verbose	Default to False. Set to TRUE to progress messages.

plantformatics/Socrates documentation built on April 3, 2025, 1:02 p.m.