metsHeatmap: Creates a correlation heatmap using ggplot2 syntax

In buddha2490/BERGMets: Metabolomics data and analysis in the BERG

Description

This function will create a heatmap using ggcorrplot(). The function will produce the correlation (or partial correlation) matrix using corMatrix() and then construct the figure. Users are given a limited number of options for adjusting the figures and can supply their own theme().

Usage

metsHeatmap(dat, compid, covariates = NULL, biochem = NULL, method = "pearson", hclust = T, title, subtitle = NULL, plot_colors = NULL, xaxis_size = 7, yaxis_size = 7, title_size = 12, subtitle_size = 8, plot_theme = NULL)

Arguments

dat	data frame including metabolites and (optionally) covariates
compid	Character vector of COMP_IDs included in the heatmap
covariates	Optional vector of covariates. If supplied, heatmap will be based on a partial correlation matrix
biochem	Data frame of biochemical metadata. At a minimum it must have two variables: COMP_ID and BIOCHEMICAL. The BIOCHEMICAL will be used to label the y-axis. If left NULL, COMP_IDs will be used to label the x- and y-axes.
method	Correlation methods: "pearson" (default), "spearman", or "kendall"
hclust	Logical: use heirarchical clustering for arranging the heatmap? Defaults to TRUE
title	Title for your plot, required.
subtitle	Optional subtitle. If left NULL, some summary information about the correlations will be added.
plot_colors	Adjust the color scheme by provided exactly THREE colors. Correlations range from -1.0 to 1.0. The first color wil be -1.0, middle color will be 0, and third color will be 1.0. The function will create gradations between these colors.
xaxis_size	Text size for the x-axis
yaxis_size	Text size for the y-axis
title_size	Text size for the title
subtitle_size	Text size for the subtitle
plot_theme	Optional ggplot2 theme. metsHeatmap() will create a theme based on some of your options. You can create a custom theme using theme() and include the name of this theme object in this argument to apply it.

Details

Building heatmaps from ggcorrplot() is a tedious process, hopefully this makes it a bit easier. Users should be able to edit most of the plot using theme() elements, or they can accept the defaults. The function returns a ggplot2 object, so you can edit the plot after running the function.

Value

ggplot object

Author(s)

Brian Carter

See Also

ggcorrplot, corMatrix, theme

Examples

breast_metabolomics <- getMetabolites("breast_metabolomics")
biochem <- b$biochem
metdata <- dplyr::left_join(survey,breast_metabolomics$metabolites,"ID")
comp.id <- sample(biochem$COMP_ID,50,replace=T)


png(file.path("test mets heatmap.png"),height=10,width=10,units="in",res=400)
metsHeatmap(metdata,
            compid=comp.id,
            covariates=NULL,
            biochem=NULL,
            method="pearson",
            hclust=T,
            title="Title to test my new metsHeatmap() function",
            subtitle="here's a subtitle describing something or another",
            plot_colors=NULL,
            xaxis_size=7,
            yaxis_size=7,
            title_size=12,
            subtitle_size=8,
            plot_theme=NULL)
dev.off()

buddha2490/BERGMets documentation built on Sept. 6, 2020, 5:11 p.m.