README.md

In BenaroyaResearch/DGETools: Differential Gene Expression Tools

DGETools

Differential Gene Expression Analysis Tools

Currenty this includes only the Volcano function for rendering Volcano plots, and writing out dge lists from limma based model fits, including contrast based models.

Here is a more complete example of the code leading up to a plot with the Volcano function.

d <- DGEList(counts=filteredCounts, genes=filteredGenes)
d <- calcNormFactors(d)
keepRows <- rowSums(round(cpm(d$counts)) >= 1) >= cut.filter*ncol(filteredCounts)
table(keepRows)
subsetMono <- grepl("DP|SP", filteredDesign$cellTypeShort)
testDGE <- d[keepRows,subsetMono]
testDGE <- calcNormFactors(testDGE)
testDesign <- filteredDesign[subsetMono,]

combo <- factor(paste(testDesign$cellTypeShort, testDesign$timepoint, sep="."))

dm <- model.matrix(~0+combo)
colnames(dm) <- levels(combo)

cm <- makeContrasts(
  SPTwovsOne   = SP.day7 - SP.day3,
  SPThreevsOne = SP.day14 - SP.day3,
  SPThreevsTwo  = SP.day14 - SP.day7,

  DPTwovsOne   = DP.day7 - DP.day3,
  DPThreevsOne = DP.day14 - DP.day3,
  DPThreevsTwo  = DP.day14 - DP.day7,

  DPandSPTwovsOne   = (DP.day7 + SP.day7) - (DP.day3 + SP.day3),
  DPandSPThreevsOne = (DP.day14 + SP.day14) - (DP.day3 + SP.day3),
  DPandSPThreevsTwo = (DP.day14 + SP.day14) - (DP.day7 + SP.day7),
  levels=dm
)

vv <- voomWithQualityWeights(testDGE, design=dm)
corfit <- duplicateCorrelation(vv, dm, block=testDesign$donorId)
#corfit$consensus.correlation # [1] 0.1501686
vv <- voomWithQualityWeights(testDGE, design=dm, block=testDesign$donorId, correlation=corfit$consensus.correlation)
fit <- lmFit(vv, dm, block=testDesign$donorId, correlation=corfit$consensus.correlation)
fit2 <- contrasts.fit(fit, cm)
fit2 <- eBayes(fit2)

Volcano(fit=fit2, target="DPandSPTwovsOne",
        title="Human Airway Cells in BPD:\nDP and SP from First to Second Timepoints",
        labelList=c("IL1A", "IL1B", "IL1RN"),
        pointSize=2, labelSize=6)

BenaroyaResearch/DGETools documentation built on July 2, 2022, 12:19 p.m.