GPRankTargets: Ranking possible target genes or regulators

In tigre: Transcription factor Inference through Gaussian process Reconstruction of Expression

Description

GPRankTargets ranks possible target genes by forming optimized models with a fixed transcription factor, a set of known target genes and targets to be tested. The transcription factor and the known targets are always included in the models while the tested targets are tested by including them in the models one at a time. The function determines itself whether to use GPSIM or GPDISIM based on the input arguments.

Usage

GPRankTargets(preprocData, TF = NULL, knownTargets = NULL,
              testTargets = NULL, filterLimit = 1.8,
              returnModels = FALSE, options = NULL,
              scoreSaveFile = NULL,
              datasetName = "", experimentSet = "")
GPRankTFs(preprocData, TFs, targets,
          filterLimit = 1.8, returnModels = FALSE, options = NULL,
          scoreSaveFile = NULL, datasetName = "", experimentSet = "")

Arguments

preprocData	The preprocessed data to be used.
TF	The transcription factor (TF) probe present in all models when TF protein translation model is used. Set to NULL (default) when translation model is not used.
knownTargets	The target genes present in all models.
testTargets	Target genes that are tested by including them in the models one at a time. Can be names of genes, or a set of indices in preprocData.
filterLimit	Genes with an average expression z-score above this figure are accepted after filtering. If this value is 0, all genes will be accepted.
returnModels	A logical value determining whether the function returns the calculated models.
options	A list of additional arguments to pass to GPLearn.
scoreSaveFile	Name of file to save the scores to after processing each gene.
TFs	The transcription factors that are tested by including them in the models one at a time.
targets	The target genes present in all models.
datasetName	For exporting the scores using export.scores: Name of the dataset in the database.
experimentSet	For exporting the scores using export.scores: Name of the experiment set in the database.

Details

The models are formed by calling GPLearn. If there is no value given to the transcription factor, a model without protein translation is used. Without protein translation model, some known targets are needed. If known targets are given, a model is first created with only the transcription factor and the known targets. The parameters extracted from this model are used as initial parameters of the models with test targets.

GPRankTFs is very similar to GPRankTargets, except it loops over candidate regulators, not candidate targets.

Value

The function returns a scoreList containing the genes, parameters and log-likelihoods of the models If returnModels is true, the function returns a list of the calculated models.

Author(s)

Antti Honkela, Jonatan Ropponen, Magnus Rattray, Neil D. Lawrence

See Also

GPLearn, scoreList, generateModels, export.scores.

Examples

## Not run: 
  # Load a mmgmos preprocessed fragment of the Drosophila developmental
  # time series
  data(drosophila_gpsim_fragment)

  # Get the target probe names
  targets <- c('FBgn0003486', 'FBgn0033188', 'FBgn0035257')
  library(annotate)
  aliasMapping <- getAnnMap("ALIAS2PROBE",
                    annotation(drosophila_gpsim_fragment))
  twi <- get('twi', env=aliasMapping)
  fbgnMapping <- getAnnMap("FLYBASE2PROBE",
                   annotation(drosophila_gpsim_fragment))
  targetProbes <- mget(targets, env=fbgnMapping)

  scores <- GPRankTargets(drosophila_gpsim_fragment, TF=twi,
                          testTargets=targetProbes,
                          options=list(quiet=TRUE),
                          filterLimit=1.8)

## End(Not run)

tigre documentation built on Nov. 8, 2020, 5:32 p.m.