Import from more single cell packages in bioconductor

[markziemann]

Bioconductor

• SCDE ✔ colname="Z" is the Z-score. Scde only compares between cell types. This is not what I'm after, but could include it for completeness.
• MAST ✔ colname = "Pr(>Chisq)" is the p value and "coef" is a log fold change. Note that data.table object is converted to normal dataframe before import
• scDD ✖ looks like a useful pkg but because it classifies genes as DE, DM, DP, etc it is hard to summarise this into a rnking value
• Monocle (v2) ✖ none of these vignettes actually work and there are several different ways to apply it such as trajectory and cluster analysis and the significance values are variable between version and in some cases there is no fold change given so hard to actually figure out the direction of genes. I'll wait to see V3 in bioconductor before adding.
• DEsingle ✔ colname = "pvalue" for p-value and "foldChange" for linear fold change. easy.

CRAN

**GitHub** * D3E * SigEMD **Other** * SINCERA

[markziemann]

Biocview for differential expression gives the following extra ones

• ballgown ✔ colname = "pval" for p-value and "fc" for linear fold change. easy.
• DEXseq ✖ vignette does not build
• NOIseq ✔ colname = "ranking" is perfect for enrichment analysis
• EBSeq ✖ classifies genes based on FDR 0.05 category so it's not really suited to mitch/GSEA
• DEGseq ✖ as far as I can tell, this function doesn't return results as an R object, just to an output file.
• TCC ✔ colname = "p.value" for p-value and "m.value" for log2 fold change.
• DEDS (expression array) ✔ using the "t" test stat as ranking statistic
• fishpond ✔ implemented as swish

Also included cuffdiff ✔ with the "test_stat" ranking stat

[markziemann]

Differential methylation

• missMethyl "t" statistic
• DMRcate "Stouffer" significance and "meanbetafc"

[markziemann]

Proteomics

• DEP ✔ 'p.val' and 'ratio'
• msmsTests ✔ "p.value" and "LogFC"
• DAPAR ✖ p-values and logFC are in different locations in the object. EG: head(signif@featureData$logFC) and head(signif@featureData$P_Value)
• plgem ✔ “PLGEM.STN” and “p.value”
• SDAMS ✔ “beta” and “pv_2part”
• DEqMS ✔ "t"

[markziemann]

Epigenomics

• DiffBind ✔ “Fold” and “p.value”
• Rcade ✖ this is an integration tool- not for differential analysis
• MEDIP - tried. too hard