NEWS.md
cTRAP(), raise error if commonPath does not existcTRAP(): new global interface with all cTRAP functionality in one place
qs instead of RDS:
convertGeneIdentifiers() replaces convertENSEMBLtoGeneSymbols():
loadENCODEsamples():
analyseDrugSetEnrichment():
launchDrugSetEnrichmentAnalysis() function to analyse drug set enrichment and visualize respective resultslaunchCMapDataLoader():
launchResultPlotter():
launchMetadataViewer() now correctly parses values from Input attributes as numericprepareCMapPerturbations(): directly set perturbation type, cell line, timepoint and dosage conditions as argumentsrankSimilarPerturbations() and predictTargetingDrugs():
threads argument allows to set number of parallel threads (not supported on Windows)chunkGiB argument allows to set size of data chunks when reading from supported HDF5 files (decreases peak RAM usage)verbose argument allows to increase details printed in the consoleprepareDrugSets(): allow greater control on the creation of bins based on numeric columns, including the setting of maximum number of bins per column and minimum bin sizeanalyseDrugSetEnrichment() and plotDrugSetEnrichment(): allow to select columns to use when comparing compound identifiers between datasetsfilterCMapMetadata(): allow filtering CMap metadata based on multiple perturbation typesprepareDrugSets(): fix issues with 3D descriptors containing missing valuesplot():
targetingDrugs objectsplotTargetingDrugsVSsimilarPerturbations():
perturbationChanges or an expressionDrugSensitivityAssociation object, passing only one argument extracts its columns as in previous versions of cTRAP (similarly to when subsetting a data.frame)analyseDrugSetEnrichment(): for the resulting table, the name of the first column was renamed from pathway to descriptor
launchDiffExprLoader(): load differential expression datalaunchCMapDataLoader(): load CMap datalaunchResultPlotter(): view and plot data resultslaunchMetadataViewer(): check metadata of a given objectdownloadENCODEknockdownMetadata(): metadata is automatically saved to a file in order to avoid downloading metadata every time this function is runplotTargetingDrugsVSsimilarPerturbations():
prepareDrugSets(): drug sets based on numeric molecular descriptors are now prepared using evenly-distributed intervalslistExpressionDrugSensitivityAssociation() lists available gene expression and drug sensitivity associationsrankSimilarPerturbations() and predictTargetingDrugs() changed name from diffExprGenes to input and now accepts:
Named numeric vector containing differential gene expression values with gene symbols as names, as before;Character vector containing a custom gene set to test for enrichment (only to use with GSEA).rankSimilarPerturbations() and predictTargetingDrugs(), when performing gsea method, allow to set different gene set size for top up- and down-regulated genes with geneSize argument:
geneSize=c(100, 200) creates gene sets from the top 100 up- and top 200 down-regulated genesgeneSize=c(150, 150) or geneSize=150 is equivalentplot() now supports plotting predictTargetingDrugs() results for a given drug, e.g. plot(targetingDrugs, "1425")
plot() now allows to set plot title with argument title
plot() now plots results based on available methods instead of trying to plot based on results from spearman method onlyplotDrugSetEnrichment() now returns a list whose names are drug set namesas.table() improvements:
predictTargetingDrugs() resultsdownloadENCODEknockdownMetadata() now correctly retrieves metadata following a change in the metadata content from ENCODEggplot2
predictTargetingDrugs()):
loadExpressionDrugSensitivityAssociation())plotTargetingDrugsVSsimilarPerturbations(), highlighting compounds that selectively select against cells with a similar differential gene expression profileanalyseDrugSetEnrichment()):
prepareDrugSets()
rankSimilarPerturbations() (when ranking against compound perturbations) and predictTargetingDrugs()
convertENSEMBLtoGeneSymbols()
L1000 instances, including in function names:
getL1000perturbationTypes() -> getCMapPerturbationTypes()
getL1000conditions() -> getCMapConditions()
downloadL1000data() -> loadCMapData()
filterL1000metadata() -> filterCMapMetadata()
loadL1000perturbations() -> prepareCMapPerturbations()
compareAgainstL1000() -> rankSimilarPerturbations()
plotL1000comparison() -> plot()
loadENCODEsamples()):
downloadENCODEsamples() to loadENCODEsamples()
loadENCODEsamples()
getCMapPerturbationTypes() (unless if using argument control = TRUE)parseCMapID()
loadCMapData()
prepareCMapPerturbations()):
prepareCMapPerturbations()
prepareCMapPerturbations() is run with argument loadZscores = TRUE
prepareCMapPerturbations()
rankSimilarPerturbations()):
cellLineMean = FALSE
rankIndividualCellLinePerturbations) when the mean is calculatedmethod argument)similarPerturbations object, obtained after running rankSimilarPerturbations():
print() with a similarPerturbations object and a specific perturbation identifieras.table() with a similarPerturbations objectplot()):
plot() with the results obtained after running rankSimilarPerturbations() or predictTargetingDrugs(); non-ranked compared data can also be plotted with argument plotNonRankedPerturbations = TRUE
plot() with a perturbationChanges object (if an identifier regarding the summary of multiple perturbations scores across cell lines is given, the plots are coloured by cell line)getCMapConditions(), including sorting of dose and time points-666 in CMap metadata as missing values and fix specific issues with metadata (such as doses displayed as 300 ng|300 ng)perturbationChanges object with only one rowperturbationChanges objectsrankSimilarPerturbations():
cellLine argument (please filter conditions with upstream functions such as filterCMapMetadata())plot():
cmapPerturbationsCompounds and cmapPerturbationsKD datasets according to new internal changes and fix their respective code in the documentationcmapR codegetL1000conditions() now shows CMap perturbation types except for controlscompareAgainstL1000()):