Statistical Methods Implemented in AltAnalyze

AltAnalyze implements various previously described algorithms for gene expression quantification, differential gene expression, alternative exon expression and feature/pathway over-representation. These algorithms are implemented in the context of a rigorous exon/transcript annotation framework, based on data from Ensembl and UCSC mRNAs. Details on all of these methods is provided here.

In general, implemented methods include:

Gene Expression Analyses

  • Robust Multichip Analysis (Affymetrix)
  • Detection Above Background p-values (Affymetrix)
  • Quantile normalization (RNASeq)
  • RPKM (gene, exon, junction) (RNASeq)
  • Geometric fold calculation
  • Moderated and conventional two-sample tests and f-test
  • Benjamini-Hochberg adjusted FDR test

Alternative Exon Analyses

  • Reciprocal-junction analyses (ASPIRE, Linear regression)
  • Alternative exon (splicing-index, FIRMA, MiDAS, adjusted f-tests)

Over-representation

  • Fisher Exact Test
  • Z score calculation
  • Permutation-based p-values
  • Benjamini-Hochberg adjusted FDR test

LineageProfiler analysis

  • Pearson/Spearman's correlation
  • Z score calculation