chemodiv

chemodiv is an R package for analysing chemodiversity of phytochemical data

The package can be used to calculate various types of diversities and dissimilarities for phytochemical datasets. This includes the use of diversity indices and dissimilarity indices that incorporate the biosynthetic and/or structural properties of the phytochemical compounds for the calculations, resulting in comprehensive measures of phytochemical diversity. A complete description of the package is available in Petrén et al. 2023.

Installation

The current version of the package can be installed from CRAN. Alternatively, the developmental version of the package can be installed from GitHub. The chemodiv package partly depends on packages from Bioconductor. Therefore, it is recommended to install the package via the install() function in the BiocManager package, rather than using the default install.packages("chemodiv"). This will ensure all dependencies are correctly installed as well.

Install current version from CRAN

install.packages("BiocManager") # Install BiocManager if not already installed
library("BiocManager")
BiocManager::install("chemodiv")

Install the developmental version from GitHub

install.packages("devtools") # Install devtools if not already installed
library("devtools")
install_github("hpetren/chemodiv")

Usage

Detailed usage notes can be found with help(chemodiv) and in the documentation for each function. See the vignette for a worked example. In short, two datasets are required. First, a dataset on the relative relative abundance/concentration (i.e. proportion) of different phytochemical compounds in different samples. Second, a dataset with the compound name, SMILES and InChIKey for all the compounds in the first dataset. The following functions can then be used:

Data formatting checks

Function chemoDivCheck() checks so that the datasets used by functions in the package are correctly formatted.

Compound classification and dissimilarities

Function NPCTable() uses the deep-learning tool NPClassifier to classify chemical compounds into groups largely corresponding to biosynthetic pathways. The function compDis() calculates and outputs a list of dissimilarity matrices with pairwise dissimilarities between chemical compounds, based on their biosynthetic and/or structural properties.

Diversity calculations

Functions calcDiv(), calcBetaDiv() and calcDivProf() are used to calculate phytochemical diversity in different ways, using both traditional indices and Hill numbers. calcDiv() calculates alpha diversity and evenness. calcBetaDiv() calculates beta diversity. calcDivProf() generates diversity profiles. Calculations of functional Hill numbers utilize a dissimilarity matrix generated by the compDis() function.

Sample dissimilarities

Function sampDis() is used to calculate Generalized UniFrac dissimilarities or Bray-Curtis dissimilarities between samples. Calculations of Generalized UniFrac dissimilarities utilizes dissimilarity matrix generated by the compDis() function.

Molecular network and properties

Function molNet() creates a molecular network of the chemical compounds and calculates some network properties using matrices generated by the compDis() function.

Chemodiversity and network plots

molNetPlot() and chemoDivPlot() are used to conveniently create basic plots of the molecular network and different types phytochemical diversity and dissimilarity calculated by the other functions in the package.

Shortcut function

Function quickChemoDiv() uses many of the other functions in the package to in one simple step calculate and visualize chemodiversity for users wanting to quickly explore their data using standard parameters.

References

Petrén H., T.G. Köllner and R.R. Junker. 2023. Quantifying chemodiversity considering biochemical and structural properties of compounds with the R package chemodiv. New Phytologist 237: 2478-2492.