Repository for my BSc Thesis

This repository contains the configuration files and scripts used in my thesis: "In-depth analysis of various shotgun metagenomics bioinformatics methods and gut microbiome characterization in mother and neonatal sepsis patients."

TL;DR: Among the tools evaluated, yacht provided the best results.

Description

The tools I evaluated include:

• Kraken 2 v2.1.1
• MetaPhlAn 3 v3.0.9
• mOTUs2 v2.6.1
• sourmash v4.8.6
• yacht v1.2.2

This repository includes the following directories:

• conda_config: Contains YAML configuration files for creating conda environments for the evaluated tools.
• data: Contains the metadata for the Tourlousse dataset used in the thesis.
• scripts: Contains the scripts used in the thesis, organized into:
  • analysis: Quarto files (*.qmd) written in R v4.3.2 for analyzing and visualizing results from the evaluated tools and clinical data.
  • benchmarking: Written in Python 3.10 and bash, includes:
    • classification: Evaluates the performance of tools in classifying shotgun metagenomics data.
    • process_taxo_results: Processes taxonomic classification results.
    • running: Scripts for preprocessing, quality control, and running the evaluated tools.

Other tools and data used include:

• Data:
  • mOTUs2: db_mOTU_v2.6.1
  • MetaPhlAn 3: mpa_v31_CHOCOPhlAn_201901
  • Kraken 2: NCBI RefSeq Complete V205 100 GB
  • bracken: k2_standard_20210517
  • sourmash and yacht: gtdb-rs207.genomic-reps.dna.k31
• Supporting tools:
  • KrakenTools
  • BBMap

Steps to Follow

Download Tourlousse data and Reference Databases

• Download Tourlousse metadata from here.
• Get Tourlousse sequence data from here.
• Follow the respective tool instructions to obtain reference data.

Tip: I used ENAdatabase-Downloader version 2 to download data from ENA. It took approximately 5 hours to download the Tourlousse data.

Create Conda Environment

• Use the YAML files in the conda_config directory. Two main configurations are used: quantnm_sourmashv4 (for yacht and sourmash) and quantnm_Tourlousse2022 (for other tools).
• Create a conda environment using the command:

conda env create -f <path_to_yaml_file>

Tip: Use mamba instead of conda for creating and solving environment faster (It took me 4 months to find a faster way to solve the environment, trust me bro :) ).

Install supporting tools (if necessary)

• KrakenTools v1.2
• BBMap v38.82

Conduct evaluations

1. Use the scripts in scripts/benchmarking/running to run the evaluated tools.
2. Use the scripts in scripts/benchmarking/process_taxo_results to process the results.
3. Use the scripts in scripts/benchmarking/classification and scripts/analysis to evaluate and visualize the performance of the tools.

How to Cite

@misc{tnmquann_bsc_thesis,
  author = {Minh-Quan Ton-Ngoc},
  title = {tnmquann/{{thesis_bsc_hcmus2019}}: Repository for BSc thesis},
  urldate = {2024-06-28},
  howpublished = {\url{https://github.com/tnmquann/thesis_bsc_hcmus2019}},
}

Author Information

Ton Ngoc Minh Quan
minhquan.tdn.ct1619@gmail.com
Page updated on 28/06/2024