Repository for the paper using omicsGMF for dimensionality reduction, batch effect removal and imputation of missing values for mass-spectrometry based proteomics data. The bioRxiv preprint is available at https://www.biorxiv.org/content/10.1101/2025.03.24.644996v1, and our package is available at https://github.com/statOmics/omicsGMF.
First, all the data are downloaded and preprocessed. The data for the CPTAC and Ecoli data
are already available through the github download of this repository.
Then, one can download the other datasets and continue the preprocessing these
based on the scripts under Data/. Using the Makefile, this can be achieved by
using make CPTAC_preparation, make CPTAC_nolab1_preparation, make Petrosius_preparation,
make Leduc_preparation and make Ecoli_preparation.
Then, one can proceeds to run all the analyses on the different datasets. These scripts
are situated under Analyses/, followed by the directory of the respective dataset. Here,
not only R-scripts are used, but also python scripts for the PIMMS imputation methods.
Using the Makefile, this can be achieved by
using make CPTAC_analysis, make CPTAC_nolab1_analysis, make Petrosius_analysis,
make Leduc_analysis and make Ecoli_analysis.
Then, one can summarize the imputation results of all different seeds and perform DE-analyses
for the CPTAC and \it{E. coli} dataset. These scripts are also situated under Analyses/,
followed by the directory of the respective dataset. Using the Makefile, this can be achieved by
using make CPTAC_results, make CPTAC_nolab1_results, make Petrosius_results,
make Leduc_results and make Ecoli_results.
To replicate all the figures of the manuscript, one can run the script
Figures/Figures_manuscript.Rmd. Alternatively, using make this is possible by
using make figures.
The code can be run by executing the different scripts directly or through the Makefile.
Using make, you can execute the full project using make all.