omicsArt User Manual

omicsArt is tool for quality control, statistical analysis, and visualization of omics data. omicsArt is currently under development and test and we will regularly release the documentation and tutorials

Citation:

Rahnavard, A. omicsArt: omics pattern discovery by visualization. Version 1.0.0.0, https://github.com/omicsEye/omicsArt (2021).

For installation and a quick demo, read the omicsArt User Manual.

omicsArt user manual

Contents

• Features
• omicsArt
  • omicsArt approach
  • Requirements
  • Installation
• Ordination plots omicsArt
  • t-sne plot
  • PCoA plot
• Microbial community diversity
  • Overall tests
  • Diversity metadata associations and plots
• Statistical summary
  • Metadata variable correlation tests
  • Statistics summary table\
• Multi-barplot form Tweedieverse results

Installation {#installation}

Install omicsArt in RStudio

1. Install devtools :
   • > install.packages('devtools')
   • > library(devtools)
2. Install omicsArt (and also all dependencies from CRAN):

install.packages(c('dplyr', 'pbapply', 'lme4', 'lmerTest', 
                   'car', 'cplm', 'pscl', 'logging', 'ggrepel', 
                   'gridExtra', 'future', 'cowplot', 'Hmisc', 
                   'TSP', 'htmlTable', 'igraph', 'insight',      
                   'lubridate', 'mgcv', 'mvtnorm', 'optparse', 
                   'parameters', 'pillar', 'pkgload', 'plotly', 'rlang', 
                   'rvest', 'seriation', 'usethis', 'viridis',    
                   'signal', 'tsne', 'openxlsx', 'readxl', 'xfun', 
                   'yulab.utils', "labdsv", "seriation","diffusionMap"), 
                   repos='http://cran.r-project.org')

if (!requireNamespace("BiocManager", quietly = TRUE)) {
    install.packages("BiocManager")
}

BiocManager::install(c("ropls", "MultiDataSet"))

devtools::install_github('omicsEye/omicsArt', force = TRUE)

Input files (omics format)

omicsArt demo

Comming soon

Load the library

library(omicsArt)

R script neteome_process.R to use and call the common functions

# load omics format data
loaded_data <- omicsArt::load_data()
# check the wiki for detailed parameters

# explanatory visualization
pcoa_plots <- omicsArt::ordplots(data = loaded_data$data, metadata = loaded_data$sample_metadata, output = output_path, outputname = 'pcoa', method = 'pcoa')

Functions

omicsArt support statistical analyses with visualizations. Here we dicuss several common functions between all studies and for more details on these functions and other functions check out our Wiki pages.

Load the file of your metabolite profiles

# load the library
library(omicsArt)

# call the load_data function
loaded_data <- omicsArt:::load_data(input=/path-to-file/filename.xlsx, type='known', sheet = 1, name = 'Metabolite')

data <- loaded_data$data

# ensure all data are stored as numeric
data <- omicsArt:::numeric_dataframe(data)

# sample info
sample_info <- loaded_data$sample_metadata

# feature info ( e.g. m/z and RT)
features_info <- loaded_data$feature_metadata

parameters:

• input: is a excel file in format

• type: can be known for characterized metabolites with a name or all for all measured features including known and uncharxterized metabolites

• sheet: defualt is 1, to read first tab of the excel sheet of the excel files but user can use different tab number.

• ID: is a word to use identifier for features (metabolites), the options are Meatbolite, HMDB_ID, and Compound_ID.

Diversity test and visualization

• data: includes taxa profiles n*m where n is number of observations (samples) and m are number features (e.g. microbila species or taxa).

• metadata: includes n*p where n is number of observations (samples) and p are number of metadat or information about samples (e.g. age, sex, and health status).

results <- omicsArt::diversity_test(data, metadat)
alpha_diversity_data <- results$alpha_diversity_data
alpha_diversity_test <- results$alpha_diversity_test
alpha_diversity_plots <- results$diversity_test_plots
overall_diversity_barplot <- results$overall_diversity_barplot


pdf(
  paste('analysis', '/alpaha_diversity.pdf', sep = ''),
  width = 2.4,
  height = 2.25,
  onefile = TRUE
)

for (meta in unique(colnames(metadata))){
  tryCatch({
    stdout <-
      capture.output(print(alpha_diversity_plots[[meta]]), type = "message")
  }, error = function(e) {
    print(meta)
    print(paste('error:', e))
  })
}
pdf(
  paste('analysis', '/alpaha_diversity_barplot.pdf', sep = ''),
  width = 2.4,
  height = 2.25,
  onefile = TRUE
)
print(overall_diversity_barplot)

fig1 <- ggdraw() + draw_plot(overall_diversity_barplot, x = 0, y = .5, width = 1, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 5), axis.title.y = element_text(size = 7, face ="bold"), axis.text.y = element_text(size = 5)), x = 0, y = 0, width = .165, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 7), axis.title.y = element_blank(), axis.text.y = element_text(size = 5)), x = .165, y = 0, width = .165, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 7), axis.title.y = element_blank(), axis.text.y = element_text(size = 5)), x = .33, y = 0, width = .165, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 7), axis.title.y = element_blank() , axis.text.y = element_text(size = 5)), x = .495, y = 0, width = .165, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 7), axis.title.y = element_blank() , axis.text.y = element_text(size = 5)), x = .66, y = 0, width = .165, height = .5) + draw_plot(alpha_diversity_plots[[meta]] + theme( axis.title.x = element_text(size = 7), axis.text.x = element_text(size = 7), axis.title.y = element_blank() , axis.text.y = element_text(size = 5)), x = .825, y = 0, width = .165, height = .5) +

draw_plot_label((label = c("a", "b", "c", "d", "e")), size = 7,x = c(0, 0, 0, .17, .33, .5, 0.65, 0.82), y = c(1, .6, 0.26, 0.26, 0.26, 0.26, 0.26, 0.26)) fig1

ggsave(filename = 'Manuscript/Figures/Fig1/fig1.pdf', plot=fig4, width = 183, height = 150, units = "mm", dpi = 350) ggsave(filename = 'Manuscript/Figures/Fig1/fig1.png', plot=fig4, width = 183, height = 150, units = "mm", dpi = 350)

Statistical summary {#statistical-summary}

Metadata variable correlation tests {#metadata-variable-corrlation-tests}

metadataCorr gets a metadata data frame rows as samples (observations) and columns features and returns a list of results (result) including two matrix (data frames) for p-values of correlations and test statistics, and a heatmap plot (ggplot object).

# read metadata file
metadata <- read.delim(
  "metatada.tsv",
  sep = '\t',
  header = T,
  fill = F,
  comment.char = "" ,
  check.names = F,
  row.names = 1
)
result <- omicsArt::metadataCorr(metadata)

Statistics summary table {#statistics-summary-table}

Multi-barplot form Tweedieverse results {#multi-barplot-form-tweedieverse-results}

library(tidyr)
library(dplyr)
library(reshape2)
library(deepath)
#setting the working directory
setwd("~/Projects/")

number_of_sig_to_keep <- 20
sig_threshold <- 0.05

## read metabolites
metabolites_Tweedieverse <- read.delim(
  "analysis/meatbolites_Tweedieverse/all_results.tsv",
  sep = '\t',
  header = T,
  fill = F,
  comment.char = "" ,
  check.names = F,
  #row.names = NA
)
metabolites_score_data_severe <- metabolites_Tweedieverse[metabolites_Tweedieverse$metadata=="Group" & metabolites_Tweedieverse$value=="Severe" ,]
rownames(metabolites_score_data_severe) <- metabolites_score_data_severe$feature

metabolites_score_data_non_severe <- metabolites_Tweedieverse[metabolites_Tweedieverse$metadata=="Group" & metabolites_Tweedieverse$value=="non-Severe" ,]
rownames(metabolites_score_data_non_severe) <- metabolites_score_data_non_severe$feature

metabolites_score_data_non_covid <- metabolites_Tweedieverse[metabolites_Tweedieverse$metadata=="Group" & metabolites_Tweedieverse$value=="non-COVID-19" ,]
rownames(metabolites_score_data_non_covid) <- metabolites_score_data_non_covid$feature



# use score_data_severe is reference
order_sig <- rownames(metabolites_score_data_severe)[1:number_of_sig_to_keep]
metabolites_score_data_severe <- metabolites_score_data_severe[order_sig,]
metabolites_score_data_severe<- metabolites_score_data_severe[order(metabolites_score_data_severe$coef),]
order_sig <- rownames(metabolites_score_data_severe)
metabolites_score_data_severe <- within(metabolites_score_data_severe,
                                        feature <- factor(feature,
                                                          levels=order_sig))
metabolites_score_data_non_severe <- metabolites_score_data_non_severe[rownames(metabolites_score_data_severe),]
metabolites_score_data_non_severe <- within(metabolites_score_data_non_severe,
                                            feature <- factor(feature,
                                                              levels=order_sig))


metabolites_score_data_non_covid <- metabolites_score_data_non_covid[rownames(metabolites_score_data_severe),]
metabolites_score_data_non_covid <- within(metabolites_score_data_non_covid,
                                           feature <- factor(feature,
                                                             levels=order_sig))

metabolites_severe_temp_diff_bar <- diff_bar_plot(metabolites_score_data_severe, threshold = sig_threshold, pvalue_col = "pval",  method = "none",
                                                  fdr ="qval", orderby = NA, x_label = 'Coefficient', y_label = '')
metabolites_non_severe_temp_diff_bar <- diff_bar_plot(metabolites_score_data_non_severe, threshold = sig_threshold, pvalue_col = "pval",  method = "none",
                                                      fdr ="qval", orderby = NA, x_label = 'Coefficient', y_label = '')
metabolites_non_covid_temp_diff_bar <- diff_bar_plot(metabolites_score_data_non_covid, threshold = sig_threshold, pvalue_col = "pval",  method = "none",
                                                     fdr ="qval", orderby = NA, x_label = 'Coefficient', y_label = '')



## read association
box_association <- readRDS("/Users/rah/Dropbox/Ali-Docs/Research_docs/Projects/COVID-Omics/analysis/meatbolites_Tweedieverse/figures/Group_gg_associations.RDS")
## do plots

fig2_metabolites <- ggdraw() +
  draw_plot(metabolites_severe_temp_diff_bar,
            x = 0, y = .47, width = .55, height = .53) +
  draw_plot(metabolites_non_severe_temp_diff_bar + theme(axis.title.y = element_blank(),
                                                         axis.text.y = element_blank(),
                                                         axis.ticks.y = element_blank(),
                                                         axis.line.y = element_blank()),
            x = .55, y = .47, width = .225, height = .53) +
  draw_plot(metabolites_non_covid_temp_diff_bar + theme(axis.title.y = element_blank(),
                                                        axis.text.y = element_blank(),
                                                        axis.ticks.y = element_blank(),
                                                        axis.line.y = element_blank()),
            x = .775, y = .47, width = .225, height = .53) +
  draw_plot(box_association[[11]] + theme(
    axis.title.x = element_text(size = 7),
    axis.text.x = element_text(size = 7),
    axis.title.y = element_text(size = 7),
    axis.text.y = element_text(size = 5)), x = 0, y = 0, width = .25, height = .45) +
  draw_plot(box_association[[52]] + theme(
    axis.title.x = element_text(size = 7),
    axis.text.x = element_text(size = 7),
    axis.title.y = element_text(size = 7),
    axis.text.y = element_text(size = 5)), x = .25, y = 0, width = .25, height = .45) +
  draw_plot(box_association[[139]] + theme(
    axis.title.x = element_text(size = 7),
    axis.text.x = element_text(size = 7),
    axis.title.y = element_text(size = 7),
    axis.text.y = element_text(size = 5)), x = .5, y = 0, width = .25, height = .45) +
  draw_plot(box_association[[2]] + theme(
    axis.title.x = element_text(size = 7),
    axis.text.x = element_text(size = 7),
    axis.title.y = element_text(size = 7),
    axis.text.y = element_text(size = 5)), x = .75, y = 0, width = .25, height = .45) +

  draw_plot_label((label = c("a",  "Severe", "non-Severe", "non-COVID", "b", "c", "d", "e")),
                  size = 7,x = c(0, .28, .53, .76, 0, .25, .5, .75), y = c(1, 1, 1, 1, 0.47, 0.47, 0.47, 0.47))
fig3_metabolites

ggsave(filename = 'figures/fig3/fig#_barplot.pdf', plot=fig2_metabolites, width = 183, height = 110, units = "mm", dpi = 350)
ggsave(filename = 'figures/fig3/fig#_barplot.pdf', plot=fig2_metabolites, width = 183, height = 110, units = "mm", dpi = 350)

User Manual! (in the wiki)