In facilebio/FacileAnalysis: Modularized and interactive analyses over a FacileDataStore

knitr::opts_chunk$set(
  # code or die
  echo = TRUE,
  # minimize verbosity
  warning = FALSE, message = FALSE,
  # dpi = 150, # for hires images
  comment = "#>")
set.seed(0xFEED)

Note: this document is a scratchpad right now. Much of the README document will replace these contents and be further expanded upon.

Scratch

The FacileAnalysis package defines a set of analysis tasks over genomic data housed in a container that implements the FacileData API (aka a FacileDataStore).

We also attempt to provide a generalized grammar (defined as S3 generics) over their results such that they can be re-used in various contexts, which include:

1. interactive exploration of results (via shiny modules and gadgets)
2. conveniant reporting of results in serialized (shiny-detached) modalities, like Rmarkdown reports
3. further (potentially interactive) interrogation on a line of inquiry that was the impetus for the analysis task itself (ie. feeding a differential expression result into a gene set enrichment analysis)

Analyses implemented within this package include:

1. Principal Components Analysis (fpca)
2. Differential Expression Analysis (fdge)
3. Feature (Gene) Set Enrichment Analysis (ffsea)

Additional analyses can be added within this package, or as new packages which import and implement the FacileAnalysis S3 API from here.

FacileAnalysis API

• S3 methods

  • result.FacileAnalysisResult() and result.ReactiveFacileAnalysisResult()
  • viz()
  • report()
  • shine()
  • ranks() and signature()

• Interactive Invocation

As much as possible, a "complete" analysis should be broken down into intermediate steps. For instance, a complete differential expresion analysis from a FacileDataStore is defined as.

1. filtering down to the samples used for the test
2. defining the model over the samples, ie. the covariate to test and which covariates to used for fixed/batch effects
3. defining the parameters to run the statistical test, ie. i) the assay from the samples to run the analysis on ii) the differential expression pipeline to use (limma, limma-trend, voom ,etc) iii) advanced options, like the threshold to test against (limma/treat), or whether to incorporate sample level weights

Required Shiny

<fanalysis> is a place holder for one othe core functions that are implemented as a facile analysis module, ie."fdge", "ffsea", "fpca", etc.

• <fanalysis>Analysis: a wrapper module that incorporates all the internal modules for each step of the analysis pipeline. The fdgeAnalysis module, for instance, is a wrapper to modules that provide an interface to the steps of a differential expression analysis.
• <fanalysis>Gadget: includes some wrapper boilerplat that enables the <fanalysis>Analysis to be invoked into a gadget from a FacileDataStore, or a facile_sample_frame

facilebio/FacileAnalysis documentation built on March 15, 2024, 7:37 a.m.