Pointers to statistical analysis of Microbiome Data (in R)

Susan Holmes

2017-07-27

Goals for these slides: only pointers

Phyloseq: Data integration
Transformations, filtering
Testing tools: networks, hierarchical testing, DESeq2, ..
Data integration, CCA, DPCOA,..
Reproducible Research: available Rmd files

Phyloseq

Phyloseq API

Heterogeneous Data Objects: S4 classes

(McMurdie and Holmes, 2013, Plos ONE).

phyloseq-measures-ggplot2

phyloseq-ordinations All standard ordinations.

Variance stabilizing using DESeq2

VST Wolfgang Huber

The way the authors proceed is to follow the rarefying path, this is not recommended and a much better procedure is to use DESeq with phyloseq, see the wiki here:

DESeq with phyloseq wiki

phyloseq-testing-deseq

Waste Not want not supplement

Testing links between communities and factor covariates

phyloseq-networks

Friedman Rafsky Tests

Package: phyloseqGraphTest

Hierarchical testing (Kris Sankaran)

Using the tree

In the Workflow

Taking account of the phylogenetic tree when testing:

CRAN package: structSSI
Journ. Stat. Software paper JSS link

Multiple table analyses

Sparse Canonical Correlation Analysis:

Triplot

In the Workflow

Modulated DPCoA

DPCoA start with taxa

CRAN package (with shiny adaptation): adaptiveGPCA

Multiple choices: example of the Enterotypes data:

Nature Paper Reproducible Research

Longitudinal Data Analyses

Perturbation Design

Vaginal Microbiome and Pregnancy

Markov Model

All Rmd and htmls for PNAS

Looking at longitudinal data in the presence of a tree using treelapse.

Example Sankey

Howto Sankey

Example Treebox (Kris Sankaran)

Tutorial to follow (Kris Sankaran)

References

Workflow paper

Workflow html

Reproducible Research links