Retip - Retention Time prediction for metabolomics

Paolo Bonini, Tobias Kind, Hiroshi Tsugawa, Dinesh Barupal, Oliver Fiehn

Published 10 May 2020 in Analytical Chemistry

Please cite:

Retip: Retention Time Prediction for Compound Annotation in Untargeted Metabolomics Paolo Bonini, Tobias Kind, Hiroshi Tsugawa, Dinesh Kumar Barupal, and Oliver Fiehn Analytical Chemistry 2020 92 (11), 7515-7522 DOI: 10.1021/acs.analchem.9b05765

Introduction

Retip is a tool for predicting Retention Time (RT) for small molecules in a high pressure liquid chromatography (HPLC) Mass Spectrometry analysis, as both an R package and a Python package. Retention time calculation can be useful in identifying unknowns and removing false positive annotations. It uses five different machine learning algorithms to built a stable, accurate and fast RT prediction model:

• Random Forest: a decision tree algorithms.
• BRNN: Bayesian Regularized Neural Network.
• XGBoost: an extreme Gradient Boosting for tree algorithms.
• lightGBM: a gradient boosting framework that uses tree based learning algorithms.
• Keras: a high-level neural networks API for Tensorflow.
• H2O autoML: an automatic machine learning tool.

Retip also includes useful biochemical databases like: HMDB, KNApSAcK, ChEBI, DrugBank, SMPDB, YMDB, T3DB, FooDB, NANPDB, STOFF, BMDB, LipidMAPS, Urine, Saliva, Feces, ECMDB, CSF, Serum, PubChem.1, PlantCyc, UNPD, BLEXP, NPA and COCONUT.

Retip installation

Retip 2.0 requires R 4.4.0 and it is recommended to use RStudio IDE to run it.

1. Download and install R from the CRAN (64 bit version recommended)
2. Download and install RStudio
3. Download and install Java JDK

Run the following command lines to install Java in Ubuntu.

sudo apt update
sudo apt install default-jre
sudo apt install default-jdk
sudo R CMD javareconf

It is also possible that r-cran-rjava needs to be installed.

4. Download and install Rtools (Windows)
5. Download and install python3 (required to use Keras model)

Run the following command line to install python with R.

reticulate::install_python(version = 3.10)

6. Run the following command block to install all required packages, as well as the Retip packages and Retip library.

install.packages('rJava', repos='http://cran.rstudio.com/')

install.packages('devtools', version='2.4.5', repos='http://cran.rstudio.com/')
install.packages('caret', version='6.0-94', repos='http://cran.rstudio.com/')
install.packages('ggplot2', version='3.5.1', repos='http://cran.rstudio.com/')
install.packages('rcdk', version='3.8.1', repos='http://cran.rstudio.com/')
install.packages('rcdklibs', version='2.9', repos='http://cran.rstudio.com/')
install.packages('doParallel', version='1.0.17', repos='http://cran.rstudio.com/')
install.packages('stringi', version='1.8.4', repos='http://cran.rstudio.com/')
install.packages('lattice', version='0.22-5', repos='http://cran.rstudio.com/')
install.packages('randomForest', version='4.7-1.1', repos='http://cran.rstudio.com/')
install.packages('xgboost', version='1.7.7.1', repos='http://cran.rstudio.com/')
install.packages('brnn', version='0.9.3', repos='http://cran.rstudio.com/')
install.packages('lightgbm', version='4.3.0', repos='http://cran.rstudio.com/')
install.packages('h2o', version='3.44.0.3')
install.packages('gtable', version='0.3.5', repos='http://cran.rstudio.com/')
install.packages('grid', version='4.4.0', repos='http://cran.rstudio.com/')
install.packages('gridExtra', version='2.3', repos='http://cran.rstudio.com/')
install.packages('reticulate', version='1.37', repos='http://cran.rstudio.com/')

devtools::install_github('olobion/Retiplib')
devtools::install_github('olobion/Retip')

⚠️ It is not possible to install Retip in conda enviroment because rJava requires NVIDIA drivers.

Usage

You will find a tutorial in the Retip app as well as in the examples folder of the GitHub repository.