Scope

The regeval repository provides R scripts that implement a simulation design for comparing a suite of regression methods for high-dimensional microbiome data. For the complete background, simulation and model specifications as well as evaluation results, please review:

Shankar J, Szpakowski S et al. A systematic evaluation of high-dimensional, ensemble-based regression for exploring large model spaces in microbiome analyses. BMC Bioinformatics. 2015 Feb;16(1):31+. Available from: http://dx.doi.org/10.1186/s12859-015-0467-6.

R code

The files regeval_simulation.R and regeval_analysis.R are the entry points for the running the scripts.

• regeval_simulation.R implements the simulation design and provides evaluation and graphing routines for a systematic comparison of approaches.
• regeval_analysis.R applies the regression approaches on the example data and provides graphing routines for comparing the findings from the approaches.

Please step-through the code and comments within the following R scripts for detailed instructions.

File	Description	Type
example_dataset.rda	An example design matrix	Data
example_response.rda	An example response vector	Data
regeval_packages.R	Installs all packages needed for the evaluation and loads the libraries	Libraries
regeval_algorithms.R	All the regression algorithms used in the evaluation.	Functions
regeval_simulation.R	Implementation of the simulation design and evaluation for a systematic comparison	Simulation
regeval_graphing.R	Graphing routines for data generated from evaluation.	Evaluation + Visualization
regeval_analysis.R	Application of the algorithms on the example data + Comparison of findings	Analysis + Visualization
regeval_colorlegend.R	Corrplot color legend	Visualization
regeval_corrplot.R	Slightly modified corrplot code	Visualization
regeval_colored_dendrogram.R	Slightly modified cluster dendrogram code	Visualization
mit_license.txt	MIT License	License

Additional reading and code:

For an application of the best-performing Bayesian ensemble regression model on experimental mouse microbiome data, please review:

Shankar, J. et al. Using Bayesian modelling to investigate factors governing antibiotic-induced Candida albicans colonization of the GI tract. Scientific Reports. 5, 8131; DOI:10.1038/srep08131 (2015). Available at: http://dx.doi.org/10.1038/srep08131

Citing the regeval repository

Please cite this repository as:

Shankar J, Szpakowski S et al. A systematic evaluation of high-dimensional, ensemble-based regression for exploring large model spaces in microbiome analyses. BMC Bioinformatics. 2015 Feb;16(1):31+. Available from: http://dx.doi.org/10.1186/s12859-015-0467-6. regeval repository: http://github.com/openpencil/regeval.

BibTeX:

@ARTICLE{Shankar2015systematic,
  title    = "A systematic evaluation of high-dimensional, ensemble-based
              regression for exploring large model spaces in microbiome
              analyses",
  author   = "Shankar, Jyoti and Szpakowski, Sebastian and Solis, Norma V and
              Mounaud, Stephanie and Liu, Hong and Losada, Liliana and Nierman,
              William C and Filler, Scott G",
  journal  = "BMC bioinformatics",
  volume   =  16,
  number   =  1,
  pages    = "31",
  month    =  "1~" # feb,
  year     =  2015,
  url      = "http://dx.doi.org/10.1186/s12859-015-0467-6",
  issn     = "1471-2105",
  pmid     = "25638274",
  doi      = "10.1186/s12859-015-0467-6",
  pmc      = "PMC4339743",
  note     =  "regeval repository:\url{http://github.com/openpencil/regeval}"
}