Statistics for (Micro/Immuno)Biologists
Created Sep 2020, Last updated: 04 October, 2021
This companion to the Workshop sessions outlines the basics of statistical analyses in R, and how knowing these can help in the design of our experiments. As laboratory-based experimental scientists, we are all used to planning experiments by carefully thinking of protocols, controls, reagents and so on. The one thing you should take away from the Workshop is to plan the analysis along with the methodological details or your experiment, rather than after gathering the data you want to analyse.
Also check out the
grafify package from GitHub (latest release v1.4.1, 10 May 2021). It makes plotting graphs, performing ANOVAs and post-hoc comparisons easier with fewer lines of code, and contains practice data sets! Get usage instructions on the vignettes website. Further details in Getting Started.
Shenoy, A. R. (2021) grafify: an R package for easy graphs, ANOVAs and post-hoc comparisons (Version v1.4.1). Zenodo. http://doi.org/10.5281/zenodo.5136508
GraphPad Prism is easy to use and sufficient for basic tests and plotting many kinds of graphs. However, several other tests are not available in Prism, especially for randomised block designs and the full range of linear mixed effects. The documentation of Prism and their Statistics Guide, are good. Therefore, analysis in Prism from the Workshop is not included here.
R and its graphical interface RStudio Desktop are free and designed for statistical analyses. Most statistical and graphing packages are available on CRAN or GitHub for free. Learning R is also useful for big-data analysis, including biological data through Bioconductor packages. Please see Getting Started and Appendix for help with R.
This document includes code for statistical tests and graphs. My aim here is to cover the basics and point towards more detailed online resources on the mathematics of statistical methods, advanced analyses and plotting in R/RStudio.