In idot/deseq2analysis: DESeq2 analysis

filterThreshold <- getFilterThreshold(deseq.result)
comp <- toSortedTibble(deseq.result, ensembl, filterThreshold) 
comparisonFoldchange <- getComparisonFoldChange(deseq.result)
tit <- getComparisonTitle(deseq.result)
if(SAVESIGNIF){
  saveTables(comp, getComparisonString(deseq.result))
}

DESeq2
r tit

p.adj < r pcut |lFC| > r lfc

• r getPosExplanation(getComparison(deseq.result))
• r getNegExplanation(getComparison(deseq.result))

r tit just the table

  # knitr::kable( as.data.frame(extractMultiLFCSignif(deseq.result)))
   knitr::kable(data.frame(a=c("A","B"),b=c("A","B")))

r tit MA plot

p.adj < r pcut |lFC| > r lfc

   plotMAPlot(comp, pcut, lfc, comparisonFoldchange)

r tit Vulcano plot

p.adj < r pcut |lFC| > r lfc

   plotVulcano(comp, pcut, lfc, comparisonFoldchange)

r tit p-value distribution

p.adj < r pcut |lFC| > r lfc

   plotPvalDist(comp, pcut)

r tit independent filtering

p.adj < r pcut |lFC| > r lfc

   plotIndependentFiltering(comp, pcut, filterThreshold)

r tit length bias

p.adj < r pcut |lFC| > r lfc

   #ggplot(comp %>% dplyr::filter(padj < pcut), aes(x=padj,y=log2FoldChange, color=length_bin)) + geom_point() + facet_wrap(. ~ gene_biotype) + geom_hline(yintercept = 0, color="red")

   #ggplot(comp %>% dplyr::filter(padj < pcut, gene_biotype == "lincRNA"), aes(x=length_bin,y=log2FoldChange)) + geom_boxplot() + ggforce::geom_sina() + geom_hline(yintercept = 0, color="red")

 #  byBioType <- ggplot(comp %>% dplyr::filter(padj < pcut), aes(x=gene_biotype,y=log2FoldChange)) + geom_boxplot() + geom_hline(yintercept = 0, color="red") + idoplots::xrot() + ylab(comparisonFoldchange)
   ggplot(comp %>% dplyr::filter(padj < pcut , abs(log2FoldChange) > lfc), aes(x=length_bin,y=log2FoldChange))  + ggforce::geom_sina() + geom_boxplot(alpha=0,color="green") + geom_hline(yintercept = 0, color="red") + ylab(comparisonFoldchange)

r tit GC bias

p.adj < r pcut |lFC| > r lfc

   ggplot(comp %>% dplyr::filter(padj < pcut,  abs(log2FoldChange) > lfc), aes(x=gc_bin,y=log2FoldChange)) + ggforce::geom_sina()  + geom_boxplot(alpha=0,color="green") + geom_hline(yintercept = 0, color="red") + ylab(comparisonFoldchange)

r tit GC + length

p.adj < r pcut |lFC| > r lfc

   ggplot(comp %>% dplyr::filter(padj < pcut, abs(log2FoldChange) > lfc), aes(x=gc_bin,y=log2FoldChange,color=length_bin)) + geom_boxplot() + geom_hline(yintercept = 0, color="red") + ylab(comparisonFoldchange)

r tit length + GC

p.adj < r pcut |lFC| > r lfc

   ggplot(comp %>% dplyr::filter(padj < pcut, abs(log2FoldChange) > lfc), aes(x=length_bin,y=log2FoldChange,color=gc_bin)) + geom_boxplot() + geom_hline(yintercept = 0, color="red") + ylab(comparisonFoldchange) + idoplots::xrot() 

r tit Counts of Significant Genes

p.adj < r pcut |lFC| > r lfc

  compCounts <- comp %>% dplyr::group_by(length_bin, gc_bin, significant=ifelse(is.na(padj), FALSE, padj < pcut & abs(log2FoldChange) > lfc)) %>% dplyr::summarise(count=dplyr::n())
  ggplot(compCounts, aes(x=length_bin,y=gc_bin,fill=count)) + 
          geom_tile() + facet_grid(. ~ significant, labeller = label_both ) + viridis::scale_fill_viridis() + idoplots::xrot() 

r tit Ratios of Significant Genes

p.adj < r pcut |lFC| > r lfc

   compCountsRel <- comp %>% dplyr::group_by(length_bin, gc_bin) %>% dplyr::summarise(total=dplyr::n(), significant=sum(padj < pcut & abs(log2FoldChange) > lfc,na.rm = TRUE), ratio=significant/total)
   ggplot(compCountsRel, aes(x=length_bin,y=gc_bin,color=ratio,size=total)) + geom_point(shape=15) + 
                 viridis::scale_color_viridis(guide = guide_colourbar(title="ratio\nsignificant")) +         
                 scale_size_continuous(range = c(4,18), breaks = c(50,100,500,1000,2000,3000)) + idoplots::xrot() 

r tit Biotype Significant Genes

p.adj < r pcut |lFC| > r lfc

    ggplot(comp %>% dplyr::filter(padj < pcut, abs(log2FoldChange) > lfc), aes(x=gene_biotype,y=log2FoldChange)) + geom_boxplot() + geom_hline(yintercept = 0, color="red") + idoplots::xrot() + ylab(comparisonFoldchange)

r tit Biotype Significant Genes

p.adj < r pcut |lFC| > r lfc

    biotypes <- comp %>% dplyr::group_by(gene_biotype) %>% dplyr::summarise(total=dplyr::n(), significant = sum(padj < pcut & abs(log2FoldChange) > lfc, na.rm = TRUE), ratio=significant/total)
   # ggplot(biotypes, aes(x=gene_biotype,y=) + geom_bar(stat="identity") + geom_hline(yintercept = 0, color="red") + idoplots::xrot() + ylab("ratio significant")
    btab <- knitr::kable(biotypes %>% dplyr::filter(significant > 0) %>% dplyr::arrange(desc(total)))  
    print(btab)

idot/deseq2analysis documentation built on Aug. 14, 2021, 5:12 a.m.