tempoR (version 1.0.4.4)

tempo.run: The main method.

Description

For each gene set in genesets, train a model on the control samples of Y~X. Generates scores for each gene set, performs permutation testing, and (optionally) writes a table of results and plots

Usage

tempo.run(X, Y, genesets, phen = NULL, ctrl = NULL, test = NULL,
  train = NULL, output = "", numPerms = 500, validation = "LOO",
  minGsSize = 2, nCores = 24, pCutoff = 0.05, fdrCutoff = 0.25,
  pMseCutoff = 0.05)

Arguments

X

a matrix with sample ids as row names and gene ids as column names.

Y

a list indexed by sample ids, containing numerical values.

genesets

a list of lists. Outer list is indexed by gene set name, inner list contains all gene ids in a given gene set

phen

a list indexed by sample ids, containing phenotypes. If ctrl and test are null, the phenotype of the first non-NA entry in the list is assumed to be the control phenotype; all others are test phenotypes

ctrl

a list of sample ids. The list of control samples to use in scoring. If train is null, these are also the training samples. Used only if phen is null.

test

a list of sample ids. The list of test samples to use in scoring. Used only if phen is null.

train

a list of sample ids. The list of control samples to train models on. If null, samples ids in ctrl are used. Used only if phen is null.

output

optional. A prefix to write table of output and plots for each reported gene set

numPerms

number of permutations to do in permutation testing. Defaults to 500

validation

type of validation to do. Defaults to leave one out, which generates deterministic results. "CV" performs 10-fold cross-validation, which is significantly faster but generates non-deterministic results.

minGsSize

minimum acceptable size for a gene set, considering only features which exist in colnames(X)

nCores

number of thread to spawn for permutation testing. This should likely be set to some number less than or equal to the number of cores on your machine. If nCores is less than 0, nCores will be set to the return value of detectCores.

pCutoff

report only gene sets with a p-value below this cutoff

fdrCutoff

report only gene sets with a FDR below this cutoff

pMseCutoff

report only gene sets with the p-value of the control mean squared error below this cutoff

Value

the output from tempo.runInstance annotated with significance for each gene set

Examples

# NOT RUN {
data("dflatExample")
data("gse32472Example")

# }
# NOT RUN {
# This runs a simple TEMPO analysis on the example data set with default settings
# (with the exception of nCores, which will instead be automatically set to a suitable
# value) and saves the output in a two temporary files.
# Note that running this example may take several minutes.
results = tempo.run(phen=gse32472Example$bpd,
    genesets=dflatExample,
    X=gse32472Example$data,
    Y=gse32472Example$age,
    output=tempfile(tmpdir = tempdir()),
    nCores=-1)

# If phen is used, the first item in the list is assumed to the control phenotype
# and all other phenotypes test. Specifiy ctrl and test exactly for more control.
# Note that running this example may take several minutes.
results = tempo.run(ctrl=gse32472Example$ctrl,
    test=gse32472Example$test,
    genesets=dflatExample,
    X=gse32472Example$data,
    Y=gse32472Example$age,
    nCores=-1)


# If training models on a held out set of data is desired, train can be specified seperately
# Note that running this example may take several minutes.
results2 = tempo.run(train=gse32472Example$ctrl[1:10],
    ctrl=gse32472Example$ctrl[11:20],
    test=gse32472Example$test,
    genesets=dflatExample,
    X=gse32472Example$data,
    Y=gse32472Example$age,
    nCores=-1)
# }
# NOT RUN {
# Reporting thresholds, number of permutations, and number of CPU cores used can all be changed.
# This command is suitable for demonstration purposes, but significance values will not be
# meaningful.
results3 = tempo.run(phen=gse32472Example$bpd,
    genesets=dflatExample,X=gse32472Example$data,
    Y=gse32472Example$age,output=tempfile(tmpdir = tempdir()),
    numPerms=2,nCores=2,pCutoff=1,fdrCutoff=2,pMseCutoff = 1)
# }