First release prepared for Bioconductor submission. The package was rewritten from the ground up; nothing of the 0.1.0 prototype other than the name and the intent remains.
OncoSeqExperiment, an S4 class extending SummarizedExperiment, holds
expression, somatic mutations, gene-level copy number, clinical
follow-up, an analysis registry and a provenance log for one cohort.
Validity checking keeps sample identifiers concordant across layers and
subsetting propagates to every layer.importGDC() retrieves a TCGA project through TCGAbiolinks. It
requests the current "STAR - Counts" workflow; the prototype's
"HTSeq - Counts" request was retired by the GDC at Data Release 32 and
no longer returns data.importCBioPortal(), cbioStudies() and cbioProfiles() query the
cBioPortal REST API for curated studies.importCounts(), importTximport(), importMAF() and asOncoSeq()
cover local files and existing objects.qcMetrics(), filterLowExpression(),
detectOutliers().normalizeCounts() with CPM, TMM, upper-quartile, RLE,
VST, rlog and log2-TPM.diffExpress() over DESeq2, edgeR and
limma-voom behind one contrast interface, plus compareEngines().runGSEA() (fgsea) and runORA() (hypergeometric), with
readGMT() and msigdbrSets().mutationSummary(), mutationFrequency(),
tumourMutationBurden(), oncoplotData(), cnvSummary(),
cnvFrequency(), cnvExpressionCorrelation().survivalFit() with median, tertile, quartile and
search-corrected optimal cut-points, coxModel() and
survivalScreen().mutationStatus(), integrateMutationExpression(),
alignLayers().plot*() functions, all returning ggplot objects.runOncoSeq() chains the central steps in one call.blcaExample, a 420-gene by 400-sample TCGA bladder cancer cohort with
all four layers, and exampleOncoSeq(), a simulated count-scale object
for examples and tests.