getMCI: Calculating MCI Scores

In xyang2uchicago/NPS: BioTIP: An R package for characterization of Biological Tipping-Point

Calculating MCI Scores

Description

This function calculates a module critical index (MCI) score for each module per state within a dataset. Each module is a cluster of transcripts generated from the function getCluster_methods. Note that a dataset should contains three or more states (samples in groups).

Usage

getMCI(
  groups,
  countsL,
  adjust.size = FALSE,
  fun = c("cor", "BioTIP"),
  df = NULL
)

Arguments

groups	A list of elements whose length is the member of states. The elements could be either be vectors or communities object of the R package igraph. If a vector, it is the output of the function getCluster_methods. The names of each vector are the pre-selected transcript IDs generated by the function sd_selection. Each vector, whose length is the number of pre-selected transcripts in a state, contains the module IDs. If a communities object, it can be obtained by getCluster_methods using the "rw" method. It is also an output of the function sd_selection.
countsL	A list of x numeric count matrices or x data frame, where x is the number of states.
adjust.size	A Boolean value indicating if MCI score should be adjusted by module size (the number of transcripts in the module) or not. Default FALSE. This parameter is not recommended for fun=BioTIP.
fun	A character chosen between ("cor", "BioTIP"), indicating where an adjusted correlation matrix will be used to calculate the MCI score.
df	NULL or a numeric matrix or data frame, where rows and columns represent unique transcript IDs (geneID) and sample names, respectively. Used only when fun='BioTIP'. By default is NULL, estinating the correlation among selected genes. Otherwise, estinating the correlation among all genes in the df, ensuring cross state comparision.

Value

A list of five elements (members, MCI, Sd, PCC, and PCCo). Each of element is a two-layer nested list whose length is the length of the input object groups. Each internal nested list is structured according to the number of modules identified in that state.

• members: vectors of unique ids

• MCI: the MCI score

• sd: standard deviation

• PCC: Mean of pairwise Pearson Correlation Coefficient calculated among the loci in a module.

• PCCo: Mean of pairwise Pearson Correlation Coefficient calculated between the loci in a module and the loci outside that module but inside the same state.

Author(s)

Zhezhen Wang [e-mail]; Xinan H Yang [e-mail]

Examples

test = list('state1' = matrix(sample(1:10, 6), 4, 3), 'state2' =
matrix(sample(1:10, 6), 4, 3), 'state3' = matrix(sample(1:10, 6), 4, 3))

## Assign colnames and rownames to the matrix
for(i in names(test)){
   colnames(test[[i]]) = 1:3
   row.names(test[[i]]) = c('g1', 'g2', 'g3', 'g4')}

cluster = list(c(1, 2, 2, 1), c(1, 2, 3, 1), c(2, 2, 1, 1))
names(cluster) = names(test)
for(i in names(cluster)){
   names(cluster[[i]]) = c('g1', 'g2', 'g3', 'g4')}

membersL_noweight <- getMCI(cluster, test, fun='cor')
names(membersL_noweight)
## [1] "members" "MCI"     "sd"      "PCC"     "PCCo"  

xyang2uchicago/NPS documentation built on Nov. 7, 2023, 1 a.m.