Anaconda

tArgeted differeNtial and globAl enriChment analysis of taxOnomic raNk by shareD Asvs

DOI: 10.13140/RG.2.2.11117.67048

Official Publication

– Download paper from PLOS ONE

Other publications that used Anaconda for example:

– See here for skin dugong (marine mammal) microbiome publication (2025)

– See here for a multimarker study in a coral reef biotope (2025)

Citation

Primary citation:

Stenger PL., Léopold A., Dinh K., Mournet P., Robert N., Drouin J., Wamejonengo J., 
Russet S., Ibanez T., Maggia L., Carriconde F. (2025). Advancing biomonitoring of eDNA 
studies with the Anaconda R package: Integrating soil and One Health perspectives in the 
face of evolving traditional agriculture practices. PLOS ONE, 20(1), e0311986.
https://doi.org/10.1371/journal.pone.0311986

Package citation:

Stenger P-L. The R Package "Anaconda": Targeted Differential and Global Enrichment 
Analysis of Taxonomic Rank by Shared Asvs. CRAN; 2022. pp. 1-28. 
http://doi.org/10.13140/RG.2.2.11117.67048

Installation

From CRAN

install.packages("Anaconda")

From GitHub (Updated version - better one)

install.packages("devtools")
library(devtools)
# install_github("PLStenger/Anaconda")
install.packages("Anaconda", repos = c('https://plstenger.r-universe.dev', 'https://cloud.r-project.org'))
library("Anaconda")

# If you encounter "Error in fetch(key) : lazy-load database" 
# Run:
.rs.restartR()

What is Anaconda?

Anaconda is an R package designed for targeted differential and global enrichment analysis of taxonomic ranks using shared ASVs (Amplicon Sequence Variants). It supports high-throughput eDNA sequencing analysis of fungi, bacteria, and metazoan.

Analysis Workflow

The package operates in two main steps:

Step I: Targeted Differential Analysis

Uses QIIME2 data with DESeq2 algorithm
Estimates variance-mean dependence in count/abundance ASV data
Tests for differential represented ASVs using negative binomial distribution

Step II: Global Enrichment Analysis

Taxon Mann-Whitney U test analysis from targeted analysis results
Uses continuous significance measures (fold-change or -log(p-value))
Identifies taxonomic ranks significantly enriched with up/down-represented ASVs

Key Features

Multi-kingdom support: Fungi, Bacteria, Metazoan
Statistical robustness: Negative binomial modeling for count data
Taxonomic hierarchy: Analysis across all taxonomic ranks
Rich visualizations: Heatmaps, PCA plots, MA plots, taxonomic trees
Database integration: FunGuild and Bactotraits databases
Automated organization: Creates structured output folders and files

Required External Databases

Note: These files are too large for the R package and must be downloaded separately.

Download Links

Kingdom	File	Download Link
Metazoan	`taxonomy_all_metazoan_QIIME2_and_NCBI_format.txt`	Download
Bacteria	`taxonomy_all_bacteria_QIIME2_and_NCBI_format.txt`	Download
Fungi	`taxonomy_all_fungi_QIIME2_and_NCBI_format.txt`	Download

Optional Database Files

File	Purpose	Download Link
`ncbitaxon_ontology.obo`	Custom database creation	Download

Input Files Required

Before running Anaconda, prepare these 4 essential files from your QIIME2 pipeline:

File	Description
`ASV.tsv`	ASV abundance table for all samples
`taxonomy.tsv`	Taxonomy-ASV mapping (rarefied dataset)
`taxonomy_RepSeq.tsv`	Taxonomy-ASV mapping (representative sequences)
`SampleSheet_comparison.txt`	User-created sample metadata file

SampleSheet_comparison.txt Format

Sample_Name    Input_File         Condition
F1             input_F1.txt       F
F2             input_F2.txt       F
F3             input_F3.txt       F
LF1            input_LF1.txt      LF
LF2            input_LF2.txt      LF
SF1            input_SF1.txt      SF

Quick Start Guide

Basic Setup

library(Anaconda)

# Save original directory
original_dir <- getwd()

# Place your 4 input files in the working directory

Choose Your Kingdom

# Run analysis for specific kingdom
Fungi()          # For fungal analysis
# OR
Bacteria()       # For bacterial analysis

Organize Files

# Move into kingdom directory
setwd("Fungi")  # or "Bacteria"
kingdom <- getwd()

# Move input files to kingdom folder
move_files()

Targeted Analysis (Step I)

# Prepare taxonomy and create input files
taxo <- get_input_files()
setwd("01_Targeted_analysis")
targeted_analysis_dir <- getwd()

# Import sample information
target_file <- target_file()
samplesInfo <- samplesInfo()

# Set minimum ASV threshold
threshold <- 1

# Create differential ASV abundance object
dasva <- get_dasva(fitType="parametric")

# Quality control plots
plotDispASVs(dasva)      # Dispersion plot
plotSparsityASV(dasva)   # Sparsity plot

# PCA analysis
data <- PCA_data_dasva()
# Use 'data' object with ggplot2 for custom PCA plots

# Heatmap of top 75 most abundant ASVs
log2.norm.counts <- heatmap_data_dasva()
colnames(log2.norm.counts) <- NULL
heatmap_condition_df <- heatmap_condition()
pheatmap(log2.norm.counts, annotation_col = heatmap_condition_df)

Differential Analysis

# Compare conditions (example: Forest vs Long Fallow)
res_forest_vs_long_fallow <- results(
  dasva, 
  contrast = c("condition", "F", "LF"), 
  pAdjustMethod = "BH", 
  alpha = 0.01
)

# Export results with taxonomy
write.table(
  merge(data.frame(res_forest_vs_long_fallow), taxo, 
        by.x="row.names", by.y="Feature.ID"),
  "Table_res_forest_vs_long_fallow_taxonomy.txt", 
  sep="\t"
)

# Add functional annotations (Fungi only)
res_forest_vs_long_fallow_guilds <- funguild_input_targeted(res_forest_vs_long_fallow)
get_funguilds_targeted(res_forest_vs_long_fallow_guilds)

# Add Bactotraits (Bacteria only)
get_bactotraits_targeted(res_forest_vs_long_fallow)

# Create MA plot
plotMA.dasva(res_forest_vs_long_fallow, alpha=0.01)

Global Analysis (Step II)

# Return to kingdom directory
setwd(kingdom)

# Create annotation database
database_fungi_creation()     # For fungi
# OR
database_bacteria_creation()  # For bacteria

# Set database path
taxon_Annotations <- "database_fungi_package_all.tab"  # Adjust for bacteria

# Move to global analysis directory
setwd("02_Global_analysis")

# Prepare input for MWU test
input_res_forest_vs_long_fallow <- input_global_analysis(res_forest_vs_long_fallow)
write.table(
  input_res_forest_vs_long_fallow, 
  "input_res_forest_vs_long_fallow.txt", 
  sep=",", quote = FALSE, row.names=FALSE
)

# Set paths
input <- "input_res_forest_vs_long_fallow.txt"
taxon_Database <- file.path(original_dir, "Working_scripts/ncbitaxon_ontology.obo")

# Run Mann-Whitney U test
taxon_mwuStats_res <- taxon_mwuStats(
  input, taxon_Database, taxon_Annotations, 
  TR, perlPath="perl", 
  largest=0.1, smallest=1, clusterCutHeight=0.1
)

# Generate taxonomic tree plot
taxon_mwuPlot(
  input, taxon_Annotations, taxon_Division, 
  absValue= -log(0.05,10), 
  level1=0.1, level2=0.05, level3=0.01, 
  txtsize=1.2, treeHeight=0.5
)

Enhanced Fungi Analysis (Optional)

# Add FunGuild information to taxonomic analysis
taxon_list <- taxon_mwu_list(input, taxon_Annotations, taxon_Division)
taxon_list_drawer <- get_taxon_list_drawer(taxon_list)
funguilds <- get_funguilds(taxon_list_drawer)
link_guilds <- get_link_guilds(taxon_list, funguilds)

# Plot with guild information
taxon_mwuPlot_guilds(
  input, taxon_Annotations, taxon_Division, 
  absValue= -log(0.05,10), 
  level1=0.1, level2=0.05, level3=0.01, 
  txtsize=1.2, treeHeight=0.5
)

System Requirements

R version: 3.5.0
Python: 2.7 (for FunGuild functionality)
Perl: Required for global analysis (recommend Strawberry Perl)

Getting Started

Download the main script Anaconda.R
Follow the step-by-step workflow in the script
Download required taxonomy databases before analysis
Ensure all system requirements are installed

Scientific Background

The Anaconda package adapts differential expression analysis methods (originally from Wright et al., 2015) for taxonomic analysis of high-throughput sequencing data. Instead of analyzing gene expression differences, Anaconda focuses on:

ASV abundance differences between experimental conditions
Taxonomic enrichment across hierarchical levels
Functional annotations through established databases

This approach enables researchers to identify not just individual species differences, but broader taxonomic patterns that may indicate ecological shifts or functional changes in microbial communities.

Contributing

Found a bug or have suggestions? Please visit our GitHub repository to:

Report issues
Suggest improvements
Contribute code
Access development protocols

License

This package is distributed under standard CRAN licensing terms. See the DESCRIPTION file for details.

Happy analyzing! ðŸ§¬ðŸ”¬