Metabolomics

In this course we will touch upon some basic ideas and considerations on

  1. What metabolomics is

  2. Data acquisition

  3. What to expect from your data

  4. in silico classifications

  5. Multivariate analysis of metabolomics

Reccomended reading

In silico classification

  1. The concept of mass spectral molecular networking explained for the first time

  2. The Global Natural Products Social Molecular Networking (GNPS) platform

  3. The preprocessing software we are going to use (MZmine)

  4. Feature-based molecular networking in GNPS

  5. Reproducible Molecular Networking Of Untargeted Mass Spectrometry Data Using GNPS

  6. Unsupervised substructure discovery (MS2LDA)

  7. In silico structure annotation <Network annotation propagation, NAP

  8. MolNetEnhancer <a tool which combines output from GNPS, MS2LDA and NAP

  9. MASST: A Web-based Basic Mass Spectrometry Search Tool for Molecules to Search Public Data (analogous to BLAST)

Case study for hands on

  1. Rasmussen et al. 2022 - Investigation of the gut microbiome in rainbow trout, using metabolomics etc.

Some general course information

We will run the analysis on a VM with all dependencies, but if you like to have the tools on your computer please see information below

  1. Mzmine3

  2. SIRIUS:CSI-FingerID

  3. Multivariate analysis, please find tutorial for installing everthing here

Installation of conda environment and dependencies for QuickFixR

Please find more info on: https://github.com/JacobAgerbo/QuickFixR

I base this tutorial on conda and therefore miniconda should be installed prior the tutorial, please see link: https://docs.conda.io/en/latest/miniconda.html

First thing we need to do is, creating a conda environment.

For this you will a config file with all dependencies. This file has already been made and can be downloaded here. It is called Metabolomics.yml.

conda env create -f Metabolomics.yml

This environment has installed R (>4.1) with several packages, but a few more is needed. These packages are not yet to be found on condas channels and therefore we will install them in R

Launch conda environment and subsequently R, by typing:

conda activate Metabolomics #activating the environment
R #starting R

Now install dependencies

dependencies <- c("boral","ggboral", "pbkrtest", "ggiraph", "hilldiv")
installed_packages <- dependencies %in% rownames(installed.packages())
if (any(installed_packages == FALSE)) {
  install.packages(dependencies[!installed_packages])}
#BiocManager
installed_packages <- dependencies %in% rownames(installed.packages())
if (any(installed_packages == FALSE)) {
BiocManager::install(dependencies[!installed_packages])}
#Github
installed_packages <- dependencies %in% rownames(installed.packages())
if (installed_packages[2] == FALSE) {
  remotes::install_github("mbedward/ggboral")}

Now please install my R package QuickFixR

devtools::install_github("JacobAgerbo/QuickFixR")

After this you should be golden! And should be able launch the shiny app simply by typing:

QuickFixR::QuickFix()

Metabolomics Analysis

Dear all

Thank you for attending!

Pre-processing of data, using MZmine3

Please see documentation here

In silico classification using SIRIUS:CSI-FingerID

Please see documentation here

Multivariate analysis with QuickFixR

First things first! Open Terminal

  • Go to Applications

  • Open System Tools > MATE Terminal

Now Terminal should open, and we need to launch our environment. First, we can our possible environments in conda.

conda env list

Here you see a list of environments, including the “Metabolomics” environment. This needs to be activated.

conda activate Metabolomics

Easy! Now we can use R and all the dependencies in the environment. First thing, launch R

R

Now you are in the R program and can launch this code to open my package for multivariate analysis called “QuickFixR”. First we load the package and set our browser options

library(QuickFixR)
options(browser="firefox")

Now launch the software! After the command, a browser window will open with an user-interface for your multivariate analysis.

QuickFix()

If you would like more commandline-based R

Please find a markdown here