iCells: iCells

In zhongmicai/bigSCale2_singleCell: A framework for the analysis of single cell data (RNA-seq and ATAC-seq) including clustering,phenotyping,pseudotime and inferring gene regulatory networks.

Description

Creates an iCell dataset starting from the typical outputs of CellRanger (.mex or .h5)

Usage

iCells(file.dir, target.cells, sample.conditions = NA,
  neighbours = 500, verbose = FALSE, pooling = "relative",
  q.cutoffs = 0.05, preproc.cells = 10000, icells.chuncks = 1e+05,
  preproc.chuncks = 2e+05, min_ODscore = 3)

Arguments

file.dir	input dataset, file name
target.cells	Number of iCells you would like to obtain
sample.conditions	optional, a factor indicating your sample conditions (for example, stages, treatments). This will prevent cells from different conditions to be pooled into the same iCell.
neighbours	How many neighbours are used to search for a mate. Increasing it will increase fidelity of iCells, but only slighlty. In fact, multiple, iterative searches for neighbours are perfomed anyway.
verbose	Whether to print on screen all the processing information
pooling	Advanced use only. A technical parameter, you should not touch this.
q.cutoffs	Advanced use only. The cutoff for pooling cells, default 0.05, meaning only cells closer than 0.05 percentile are pooled. Deacresing the values yields better quality iCells but longer times, and viceversa.
preproc.cells	Advanced use only. how many cells to use for the initial creation of the model. Reduce it if you have memory issues. Useless to increase it.
icells.chuncks	Advanced use only. Size of the chunks of the original cells. Reduce it if you have memory issues. Probably useless to increase it.
preproc.chuncks	Advanced use only. Reduce it if you have memory issues. Increase to fasten the initial step where the model is calculated.
min_ODscore	Advanced use only. Increasing the value will result in using less highly variable genes for the creation of iCells.

Value

A list with two elements: icells.data (for you) and debugging (for me, if there are problems).In addition, it writes to current working directory the icells matrix automatically under the name icells.mtx.gz.

i

• icell.mat icell Expression counts in the Matricx format

• iCells indices of the original cells

• output.conditions If you forced a pooling by condition (parameter sample.conditions) this vector contains the condition of each icell.

• model The numerical model used to compute cell to cell distances. Its meaning is explained in the first part (bigSCale core, advanced use) of the GitHub tutorial

• driving.genes The highly variable genes used to caculate cell to cell distances

Examples

check the online tutorial at Github

zhongmicai/bigSCale2_singleCell documentation built on Nov. 5, 2019, 1:26 p.m.