NANOTAXI

Offering Real-time 16s DNA Classification of Long Read Sequencing.

Features

• Faster Results from Real-time Workflow for Multiplexed Data
• Easy to use GUI for Researchers with minimal coding background
• Richer Insights from various Downstream analyses and publication-ready plots
• Offline Analysis using different pipelines available for Nanopore Sequencing

Classification Tools and Database Compatibility in NANOTAXI

Tool	Mode	GTDB	REFSEQ	GSR	MIMT	EMU DB
Kraken2	Both	✓	✓	✓	✓	✓
Minimap2	Both	✓	✓	✓	✓	✓
EMU	Both	✓	✓	✓	✓	✓
MMSEQS	Both	✓	✓	✓	✓	✓

Legend

• ✓: Tool supports this database
• EMU DB: Proprietary database format for EMU.
• Both: Supported in both real-time mode and offline mode (Post-run).

Installation

To run the app locally, please install R version >= 4.4.2. Also, please ensure that the MinKNOW app version >= 24.06.10.

You can install R and all the required packages using a single command. If not present, this command will install the conda and create a new environment named nanotaxi-env.

if which conda >/dev/null; then
        
        echo "Conda Exist"

else
        source ~/.bashrc
        
        wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh \
        && chmod +x miniconda.sh && bash miniconda.sh -b -p miniconda

        rm -r miniconda.sh
        
        base_dir=$(echo $PWD)
        
        export PATH=$base_dir/miniconda/bin:$PATH
        
        source ~/.bashrc
        
        echo -e "$base_dir/miniconda/etc/profile.d/conda.sh" >> ~/.profile
        
        conda init bash

fi

conda create -n nanotaxi-env --file Installation/nanotaxi-env.txt -y

If you have installed R from the above command, then to run the app, please activate the conda environment first using the following command:

conda activate nanotaxi-env

Launch the app using the following command:

shiny::runApp("main_app.R")

You can also run the app directly from GitHub using the following command:

# Make Sure that the shiny package is installed in the R.
shiny::runGitHub("NANOTAXI", "Nirmal2310")

The app will download and install the required Conda environments and databases necessary for real-time classification of long reads when it starts up for the first time. So, please ensure that enough free space is available on the system.

Running in Remote Server

If you are running the application on the remote server and want to open the interface in the local browser, please use the given step-by-step guide:

• Assign a port for running a local R Shiny server by using the following command:

echo "options(shiny.port=5316)" >> ~/.Rprofile

Here, I have assigned the port 5316 for running R Shiny Applications. Please ensure that the port is free before assigning it to run Shiny Applications.

• Connect to the Remote Server in a new window by using the following command:

ssh -L 8080:localhost:5316 hostname@ip_address

Change the local host port number, host name and the IP address accordingly.

• Now launch the Shiny App through the server's terminal:

shiny::runApp("main_app.R", launch.browser=FALSE)

• Finally, open NANOTAXI in your browser of choice by entering the following address:

http://127.0.0.1:8080

While setting up the MinKNOW app for sequencing, please make sure that under the barcoding setting, both Trim barcodes and Barcode both ends are disabled as shown below:

This is to minimise the rate of unclassified reads during a live run. NANOTAXI utilises Dorado to trim the barcodes during the analysis of demultiplexed barcoded reads, ensuring the removal of synthetic DNA sequence.

Additionally, in the Output settings, please select Number of reads under the Based on option and enter 500 under the Reads threshold option, as the app will process the barcoded data in batches of 500 reads per barcode. The reference screenshot is shown below:

Lastly, Minknow core 5.x requires a secure channel connection to be made by the Minknow API. In order to do this, please add the following line to the ~/.bashrc:

echo "export MINKNOW_TRUSTED_CA=\"/var/lib/minknow/data/rpc-certs/minknow/ca.crt\"" >> ~/.bashrc
source ~/.bashrc

To set up Offline Analysis, tick the checkbox adjacent to the setup option in Offline Analysis under the INPUT tab. It will first download and install all the required software and databases, and then analyse the data.
However, all pipelines are currently available in both real-time and offline settings. All required software and databases will be downloaded and installed during the initial launch of the application.

Usage/Examples

The Test Dataset is taken from the Bioproject ID PRJEB82315.

We have used Emu to analyse the test dataset comprising 20 samples representing 20 barcodes and classified into three groups based on body fluids.

Barcode	Sample	Group
barcode01	ERR13935186	Crop Digesta
barcode02	ERR13935187	Crop Digesta
barcode03	ERR13935188	Crop Digesta
barcode04	ERR13935189	Crop Digesta
barcode05	ERR13935191	Crop Digesta
barcode06	ERR13935192	Zymobiomics
barcode07	ERR13935193	Zymobiomics
barcode08	ERR13935195	Zymobiomics
barcode09	ERR13935223	Zymobiomics
barcode10	ERR13935225	Zymobiomics
barcode11	ERR13935170	Feces
barcode12	ERR13935171	Feces
barcode13	ERR13935172	Feces
barcode14	ERR13935174	Feces
barcode15	ERR13935176	Feces
barcode16	ERR13935177	Feces
barcode17	ERR13935178	Feces
barcode18	ERR13935181	Feces
barcode19	ERR13935182	Feces
barcode20	ERR13935184	Feces

The user can run the example dataset by selecting Example Data, add the control group name in Select Control Group and clicking Use Example Data under INPUT tab.

Demo

Please see the demo of Real-time classification by using the following link

Documentation

For detailed information about NANOTAXI, please refer to the Documentation.

Roadmap

• Add Differential Abundance Analysis.
• Add support for EMUDB, GSR DB, MIMt DB, REFSEQ and GTDB.
• Add support for PICRUSt2 for functional inference.
• Add support for SQK-MAB114.24 (16S + ITS).

Authors

• Nirmal Singh Mahar
• Ishaan Gupta

Feedback/Help

If you have any feedback/issues, please report the issue via GitHub.

Acknowledgements

Pipelines/Software Used in the App:

• KRAKEN2
• EMU
• MMseqs2
• TaxonKit
• BBTools
• Chopper
• SeqKit
• Minimap2
• GNU Parallel
• PICRUSt2

Databases Used in the App:

• GTDB
• NCBI 16s Database
• GSR DB
• MIMt
• EMU Database

Python Packages Used in the App:

• minknow-api = 6.0.4
• grpcio = 1.60.1
• pandas
• pysam

R Packages Used in the App:

• shiny
• shinyBS
• reticulate
• tidyverse
• shinyjs
• DT
• plotly
• shinyFiles
• markdown
• validate
• ggpubr
• dendextend
• BiocManager
• vegan
• gridExtra
• ggsci
• scales
• viridis
• circlize
• ggrepel
• devtools
• compositions
• forcats
• formattable
• future
• promises
• ggtext
• FactoMineR
• ggforce
• ComplexHeatmap
• pairwiseAdonis
• ANCOM-BC2
• clusterProfiler