Package 'MPAC'

Title: Multi-omic Pathway Analysis of Cell
Description: Multi-omic Pathway Analysis of Cell (MPAC), integrates multi-omic data for understanding cellular mechanisms. It predicts novel patient groups with distinct pathway profiles as well as identifying key pathway proteins with potential clinical associations. From CNA and RNA-seq data, it determines genes’ DNA and RNA states (i.e., repressed, normal, or activated), which serve as the input for PARADIGM to calculate Inferred Pathway Levels (IPLs). It also permutes DNA and RNA states to create a background distribution to filter IPLs as a way to remove events observed by chance. It provides multiple methods for downstream analysis and visualization.
Authors: Peng Liu [aut, cre] , Paul Ahlquist [aut], Irene Ong [aut], Anthony Gitter [aut]
Maintainer: Peng Liu <[email protected]>
License: GPL-3
Version: 1.1.4
Built: 2025-01-31 08:34:57 UTC
Source: https://github.com/bioc/MPAC

Help Index

Cluster samples by pathway over-representation

Description

Cluster samples by pathway over-representation

Usage

clSamp(ovrmat, n_neighbors = 10, n_random_runs = 200, threads = 1)

Arguments

ovrmat

A matrix of gene set over-representation adjusted p-values with rows as gene sets and columns as samples. It is the output from ovrGMT().
n_neighbors

Number of neighbors for clustering. A larger number is recommended if the size of samples is large. Default: 10.
n_random_runs

Number of random runs. Due to randomness introduced to the Louvain algorithm in R igraph 1.3.0 (https://github.com/igraph/rigraph/issues/539), a large number of runs are recommended to evaluate randomness in the clustering results. Default: 200, which shall be safe for sample size < 50. Please increase it accordingly for a larger sample size.
threads

Number of threads to run in parallel. Default: 1

Value

A data table with each row representing one clustering result, and the first column denotes the number of occurrences of a clustering result and the rest of columns indicating each sample's cluster index. Rows are ordered by the number of occurrences from high to low.

Examples

fovr = system.file('extdata/clSamp/ovrmat.rds', package='MPAC')
ovrmat = readRDS(fovr)

clSamp(ovrmat)

Collect Inferred Pathway Levels (IPLs) from PARADIGM runs on permuted data

Description

Collect Inferred Pathway Levels (IPLs) from PARADIGM runs on permuted data

Usage

colPermIPL(indir, n_perms, sampleids = NULL, threads = 1)

Arguments

indir

Input folder that saves PARADIGM results. It should be set as the same as outdir as in runPrd().
n_perms

Number of permutations to collect.
sampleids

Sample IDs for which IPLs to be collected. If not provided, all files with suffix '_ipl.txt' in indir will be collected. Default: NULL.
threads

Number of threads to run in parallel. Default: 1

Value

A data.table object with columns of permutation index, pathway entities and their IPLs.

Examples

indir = system.file('/extdata/runPrd/', package='MPAC')
n_perms = 3

colPermIPL(indir, n_perms)

Collect Inferred Pathway Levels (IPLs) from PARADIGM runs on real data

Description

Collect Inferred Pathway Levels (IPLs) from PARADIGM runs on real data

Usage

colRealIPL(indir, sampleids = NULL, file_tag = NULL)

Arguments

indir

Input folder that saves PARADIGM results. It should be set as the same as outdir as in runPrd().
sampleids

Sample IDs for which IPLs to be collected. If not provided, all files with suffix '_ipl.txt' in indir will be collected. Default: NULL.
file_tag

A string of output file name tag. Default: NULL

Value

A data.table object with columns of pathway entities and their IPLs.

Examples

indir = system.file('/extdata/runPrd/', package='MPAC')

colRealIPL(indir)

Find consensus pathway motifs from a list of pathways

Description

Find consensus pathway motifs from a list of pathways

Usage

conMtf(subntwl, omic_genes = NULL, min_mtf_n_nodes = 5)

Arguments

subntwl

A list of igraph objects representing input pathways from different samples. It is the output from subNtw()
omic_genes

A vector of gene symbols to narrow down over-representation calculation to only those with input genomic data. If not provided, all genes in the GMT file will be considered. Default: NULL.
min_mtf_n_nodes

Number of minimum nodes in a motif. Default: 5

Value

A list of igraph objects representing consensus pathway motifs

Examples

fsubntwl = system.file('extdata/conMtf/subntwl.rds', package='MPAC')
subntwl = readRDS(fsubntwl)

fomic_gns = system.file('extdata/TcgaInp/inp_focal.rds', package='MPAC')
omic_gns = rownames(readRDS(fomic_gns))

conMtf(subntwl, omic_gns, min_mtf_n_nodes=50)

Filter IPLs from real data by distribution from permuted data

Description

Filter IPLs from real data by distribution from permuted data

Usage

fltByPerm(realdt, permdt, threads = 1)

Arguments

realdt

A data.table object containing entities and their IPLs from real data. It is the output from colRealIPL().
permdt

A data.table object containing permutation index, entities and their IPLs from permuted data. It is the output from colPermIPL().
threads

Number of threads to run in parallel. Default: 1

Value

A matrix of filtered IPLs with rows as entities and columns as samples. Entities with IPLs observed by chance are set to NA.

Examples

freal = system.file('extdata/fltByPerm/real.rds', package='MPAC')
fperm = system.file('extdata/fltByPerm/perm.rds', package='MPAC')
realdt = readRDS(freal)
permdt = readRDS(fperm)

fltByPerm(realdt, permdt)

Get significantly over-represented gene sets for clustered samples

Description

Get significantly over-represented gene sets for clustered samples

Usage

getSignifOvrOnCl(ovrmat, cldt, min_frc = 0.8)

Arguments

ovrmat

A matrix containing over-representation adjusted P with rows as gene set names and columns as sample IDs. It is the output of the ovrGMT() function.
cldt

A data table with each row representing one clustering result, and the first column denotes the number of occurrences of a clustering result and the rest of columns indicating each sample's cluster index. It is the output of the clSamp() function. Only the most frequent clustering result will be used to plot.
min_frc

A minimum fraction of samples in a cluster that have a gene set significantly over-represented (adjusted P < 0.05). This is used to select gene sets to plot. Default: 0.8

Value

A list of a matrix and a data.table object. The matrix has rows as over-represented gene sets, columns as samples, and each cell stores an adjusted P for over-representation. The data.table has the clustering informations with samples in the same order as the matrix's column.

Examples

ovrmat <- system.file('extdata/pltOvrHm/ovr.rds',package='MPAC') |> readRDS()
cldt   <- system.file('extdata/pltOvrHm/cl.rds', package='MPAC') |> readRDS()

getSignifOvrOnCl(ovrmat, cldt)

Calculate over-representation of gene sets in each sample by genes from sample's largest sub-pathway

Description

Calculate over-representation of gene sets in each sample by genes from sample's largest sub-pathway

Usage

ovrGMT(subntwlist, fgmt, omic_genes = NULL, threads = 1)

Arguments

subntwlist

A list of igraph objects represented the largest sub-pathway for each sample. It is the output of subNtw().
fgmt

A gene set GMT file. This will be the same file used for the gene set over-representation calculation in the next step. It is used here to ensure output sub-pathway contains a minimum number of genes from to-be-used gene sets.
omic_genes

A vector of gene symbols to narrow down over-representation calculation to only those with input genomic data. If not provided, all genes in the GMT file will be considered. Default: NULL.
threads

Number of threads to run in parallel. Default: 1

Value

A matrix containing over-representation adjusted P with rows as gene set names and columns as sample IDs.

Examples

fsubntwl  = system.file('extdata/subNtw/subntwl.rds',    package='MPAC')
fgmt      = system.file('extdata/ovrGMT/fake.gmt',       package='MPAC')
fomic_gns = system.file('extdata/TcgaInp/inp_focal.rds', package='MPAC')
subntwl  = readRDS(fsubntwl)
omic_gns = rownames(readRDS(fomic_gns))

ovrGMT(subntwl, fgmt, omic_gns)

Plot a heatmap of pathway and omic states of a protein and its pathway neighbors

Description

Plot a heatmap of pathway and omic states of a protein and its pathway neighbors

Usage

pltNeiStt(real_se, fltmat, fpth, protein = "")

Arguments

real_se

A SummarizedExperiment object of PARADIGM CNA and RNA states. It is the output fromm ppRealInp() and must contain the omic states for the one defined in the protein argument.
fltmat

A matrix contains filterd IPL with rows as entity and column as samples. This is the output from fltByPerm(). Entity with NA value will be set to 0 and plotted as in 'normal' state.
fpth

Name of a pathway file for PARADIGM.
protein

Name of the protein to plot. It requires to have CN and RNA state data, as well as pathway data from the input. Default: ”

Value

A heatmap of pathway and omic states of a protein and its pathway neighbors

Examples

fpth = system.file('extdata/Pth/tiny_pth.txt', package='MPAC')

freal = system.file('extdata/pltNeiStt/inp_real.rds', package='MPAC')
fflt  = system.file('extdata/pltNeiStt/fltmat.rds',   package='MPAC')

real_se = readRDS(freal)
fltmat = readRDS(fflt)
protein = 'CD86'

pltNeiStt(real_se, fltmat, fpth, protein)

Plot a heatmap of over-represented gene sets for clustered samples

Description

Plot a heatmap of over-represented gene sets for clustered samples

Usage

pltOvrHm(ovrmat, cldt, min_frc = 0.8)

Arguments

ovrmat

A matrix containing over-representation adjusted P with rows as gene set names and columns as sample IDs. It is the output of the ovrGMT() function.
cldt

A data table with each row representing one clustering result, and the first column denotes the number of occurrences of a clustering result and the rest of columns indicating each sample's cluster index. It is the output of the clSamp() function. Only the most frequent clustering result will be used to plot.
min_frc

A minimum fraction of samples in a cluster that have a gene set significantly over-represented (adjusted P < 0.05). This is used to select gene sets to plot. Default: 0.8

Value

A heatmap with rows as over-represented gene sets and columns as samples splited by clusters.

Examples

ovrmat <- system.file('extdata/pltOvrHm/ovr.rds',package='MPAC') |> readRDS()
cldt   <- system.file('extdata/pltOvrHm/cl.rds', package='MPAC') |> readRDS()

pltOvrHm(ovrmat, cldt)

Prepare input copy-number (CN) alteration data to run PARADIGM

Description

Prepare input copy-number (CN) alteration data to run PARADIGM

Usage

ppCnInp(cn_tumor_mat)

Arguments

cn_tumor_mat

A matrix of tumor CN focal data with rows as genes and columns as samples. A value of 0 means normal CN, > 0 means amplification, and < 0 means deletion.

Value

A SummarizedExperiment object of CN state for PARADIGM

Examples

fcn = system.file('extdata/TcgaInp/focal_tumor.rds', package='MPAC')
cn_tumor_mat = readRDS(fcn)

ppCnInp(cn_tumor_mat)

Permute input genomic state data between genes in the same sample

Description

Permute input genomic state data between genes in the same sample

Usage

ppPermInp(real_se, n_perms=100, threads=1)

Arguments

real_se

A SummarizedExperiment object of CN and RNA states from real samples with rows as genes and columns as samples. It is the output from ppRealInp().
n_perms

Number of permutations. Default: 100
threads

Number of threads to run in parallel. Default: 1

Value

A list of SummarizedExperiment objects of permuted CN and RNA states. The metadata i in each obbect denotes its permutation index.

Examples

freal = system.file('extdata/TcgaInp/inp_real.rds', package='MPAC')
real_se = readRDS(freal)

ppPermInp(real_se, n_perms=3)

Prepare input copy-number (CN) alteration and RNA data to run PARADIGM

Description

Prepare input copy-number (CN) alteration and RNA data to run PARADIGM

Usage

ppRealInp(
  cn_tumor_mat,
  rna_tumor_mat,
  rna_normal_mat,
  rna_n_sd = 2,
  threads = 1
)

Arguments

cn_tumor_mat

A matrix of tumor CN focal data with rows as genes and columns as samples. A value of 0 means normal CN, > 0 means amplification, and < 0 means deletion.
rna_tumor_mat

A matrix of RNA data from tumor samples with rows as genes and columns as samples
rna_normal_mat

A matrix of RNA data from normal samples with rows as genes and columns as samples
rna_n_sd

Standard deviation range from fitted normal samples to define RNA state. Default: 2, i.e. 2*sd
threads

Number of threads to run in parallel. Default: 1

Value

A SummarizedExperiment object of CN and RNA state for PARADIGM

Examples

fcn = system.file('extdata/TcgaInp/focal_tumor.rds', package='MPAC')
ftumor = system.file('extdata/TcgaInp/log10fpkmP1_tumor.rds', package='MPAC')
fnorm = system.file('extdata/TcgaInp/log10fpkmP1_normal.rds', package='MPAC')

cn_tumor_mat = readRDS(fcn)
rna_tumor_mat = readRDS(ftumor)
rna_norm_mat  = readRDS(fnorm)

ppRealInp(cn_tumor_mat, rna_tumor_mat, rna_norm_mat)

Prepare input RNA data to run PARADIGM

Description

Prepare input RNA data to run PARADIGM

Usage

ppRnaInp(rna_tumor_mat, rna_normal_mat, rna_n_sd = 2, threads = 1)

Arguments

rna_tumor_mat

A matrix of RNA data from tumor samples with rows as genes and columns as samples
rna_normal_mat

A matrix of RNA data from normal samples with rows as genes and columns as samples
rna_n_sd

Standard deviation range from fitted normal samples to define RNA state. Default: 2, i.e. 2*sd
threads

Number of threads to run in parallel. Default: 1

Value

A SummarizedExperiment of RNA state for PARADIGM

Examples

ftumor = system.file('extdata/TcgaInp/log10fpkmP1_tumor.rds', package='MPAC')
fnorm = system.file('extdata/TcgaInp/log10fpkmP1_normal.rds', package='MPAC')
rna_tumor_mat = readRDS(ftumor)
rna_norm_mat  = readRDS(fnorm)

ppRnaInp(rna_tumor_mat, rna_norm_mat, threads=2)

Prepare required files to run PARADIGM

Description

Prepare required files to run PARADIGM

Usage

ppRunPrd(pat, cnmat, rnamat, outdir, file_tag=NULL)

Arguments

pat

Sample ID
cnmat

CN matrix
rnamat

RNA matrix
outdir

Output folder to save all results.
file_tag

A string of output file name tag. Default: NULL

Value

None

Run PARADIGM on permuted data

Description

Run PARADIGM on permuted data

Usage

runPermPrd(perml, fpth, outdir,
    PARADIGM_bin=NULL, nohup_bin=NULL, sampleids=NULL, threads=1)

Arguments

perml

A list of SummarizedExperiment objects of permuted CNA and RNA states generated by ppPermInp().
fpth

Name of a pathway file for PARADIGM.
outdir

Output folder to save all results.
PARADIGM_bin

PARADIGM binary, which can be downloaded from https://github.com/sng87/paradigm-scripts/tree/master/public/exe. Note that the binary is only available for Linux or MacOS. Default: NULL
nohup_bin

nohup binary, which is used for long running PARADIGM jobs. Default: NULL
sampleids

A vector of sample IDs to run PARADIGM on. If not provided, all the samples that exist in both copy-number alteration and RNA files will be ran. Default: NULL
threads

Number of threads to run in parallel. Default: 1

Value

None

Examples

fperm = system.file('extdata/TcgaInp/inp_perm.rds', package='MPAC')
perml = readRDS(fperm)
fpth = system.file('extdata/Pth/tiny_pth.txt', package='MPAC')
outdir = tempdir()
paradigm_bin = '/path/to/PARADIGM'  ## change to binary location
pat = 'TCGA-CV-7100'

# depends on external PARADIGM binary, do not run
runPermPrd(perml, fpth, outdir, paradigm_bin, sampleids=c(pat))

Run PARADIGM on multi-omic data

Description

Run PARADIGM on multi-omic data

Usage

runPrd(real_se, fpth, outdir, PARADIGM_bin=NULL, nohup_bin=NULL,
    sampleids=NULL, file_tag=NULL, threads=1)

Arguments

real_se

A SummarizedExperiment object of PARADIGM CNA and RNA states. It is the same matrix as the output from ppRealInp().
fpth

Name of a pathway file for PARADIGM.
outdir

Output folder to save all results.
PARADIGM_bin

PARADIGM binary, which can be downloaded from https://github.com/sng87/paradigm-scripts/tree/master/public/exe. Note that the binary is only available for Linux or MacOS. Default: NULL
nohup_bin

nohup binary, which is used for long running PARADIGM jobs. Default: NULL
sampleids

A vector of sample IDs to run PARADIGM on. If not provided, all the samples that exist in both copy-number alteration and RNA files will be ran. Default: NULL
file_tag

A string of output file name tag. Default: NULL
threads

Number of threads to run in parallel. Default: 1

Value

None

Examples

freal = system.file('extdata/TcgaInp/inp_real.rds', package='MPAC')
real_se  = readRDS(freal)

fpth = system.file('extdata/Pth/tiny_pth.txt', package='MPAC')
outdir = tempdir()
paradigm_bin = '/path/to/PARADIGM'  ## change to binary location

# depends on external PARADIGM binary
runPrd(real_se, fpth, outdir, paradigm_bin, sampleids=c('TCGA-CV-7100'))

Subset pathways by IPL results

Description

Subset pathways by IPL results

Usage

subNtw(fltmat, fpth, fgmt, min_n_gmt_gns = 2, threads = 1)

Arguments

fltmat

A matrix contains filterd IPL with rows as 'entity' and column as samples. This is the output from fltByPerm(). Entity with NA in all columns will be ignored.
fpth

Name of a pathway file for PARADIGM.
fgmt

A gene set GMT file. This will be the same file used for the gene set over-representation calculation in the next step. It is used here to ensure output sub-pathway contains a minimum number of genes from to-be-used gene sets.
min_n_gmt_gns

Minimum number of genes from the GMT file in the output sub-pathway. Default: 2.
threads

Number of threads to run in parallel. Default: 1

Value

A list of igraph objects representing the largest sub-pathway for each sample.

Examples

fflt = system.file('extdata/fltByPerm/flt_real.rds', package='MPAC')
fltmat = readRDS(fflt)
fpth = system.file('extdata/Pth/tiny_pth.txt',       package='MPAC')
fgmt = system.file('extdata/ovrGMT/fake.gmt',        package='MPAC')

subNtw(fltmat, fpth, fgmt, min_n_gmt_gns=1)