gemini_create_input: gemini_create_input

In sellerslab/gemini: GEMINI: Variational inference approach to infer genetic interactions from pairwise CRISPR screens

gemini_create_input

Description

Creates a gemini.input object from a counts matrix with given annotations.

Usage

gemini_create_input(
  counts.matrix,
  sample.replicate.annotation = NULL,
  guide.annotation = NULL,
  samplesAreColumns = TRUE,
  sample.column.name = "samplename",
  gene.column.names = NULL,
  ETP.column = 1,
  LTP.column = NULL,
  verbose = FALSE
)

Arguments

counts.matrix	a matrix of counts with rownames corresponding to features (e.g. guides) and colnames corresponding to samples.
sample.replicate.annotation	a data.frame of annotations for each sample/replicate pair. Note that at least one column in sample.replicate.annotation must correspond to the colnames of counts.matrix (see Details) (default = NULL)
guide.annotation	a data.frame of annotations for each guide. Note that at least one column in guide.annotation must correspond to the rownames of counts.matrix (default = NULL)
samplesAreColumns	a logical indicating if samples are on the columns or rows of counts.matrix. (default = TRUE)
sample.column.name	a character or integer indicating which column of sample.replicate.annotation describes the samples.
gene.column.names	a character or integer vector of length(2) indicating which columns of guide.annotation describe the genes being targeted.
ETP.column	a character or integer vector indicating which column(s) of counts.matrix contain the early time-point(s) of the screen (i.e. pDNA, early sequencing, etc.). Defaults to the first column.
LTP.column	a character or integer vector indicating which column(s) is the later time-point of the screen (i.e. day21, post-treatment, etc.). Defaults to (1:ncol(counts.matrix))[-ETP.column], or all other columns except for those specified by ETP.column.
verbose	Verbosity (default FALSE)

Details

This function initializes a gemini.input object from a counts matrix. There are a few key assumptions made in the input format.

• The counts matrix is regular.

• The counts matrix structure is in accordance with the samplesAreColumns parameter.

• The first column of sample.replicate.annotation matches with the existing dimension names of the counts matrix.

• The first column of guide.annotations matches with the existing dimension names of the counts matrix.

• sample.column.name must specify a column in sample.replicate.annotation (either by name or index) that describes unique samples.

• gene.column.names must specify two columns in sample.replicate.annotation (either by name or index) that describe genes.

Value

a gemini.input object

Examples

data("counts", package = "gemini")
data("sample.replicate.annotation", package = "gemini")
data("guide.annotation", package = "gemini")
Input <- gemini_create_input(
    counts.matrix = counts,
    sample.replicate.annotation = sample.replicate.annotation,
    guide.annotation = guide.annotation,
    sample.column.name = "samplename",
    gene.column.names = c("U6.gene", "H1.gene")
)

sellerslab/gemini documentation built on Dec. 5, 2022, 8:56 a.m.