/2023_Metamobilome

Primary LanguagePython

Cancer Metamobilome 2023

Code used to replicate data analyses in the manuscript The gut meta-mobilome associates with pancreatic cancer.

Code

Code used to analyze SGVFinder, ICRA, mOTUs and corncob outputs and metadata and generate plots in Fig. 1-4 and Supplementary Fig. 1-7.

  • Fig. 1a: uses SGVFinder output based on data from the Spanish study. (Input files can be found under ‘Data/SGVFinder/’)
  • Fig. 1b: uses SGVFinder and ICRA outputs based on data from all seven studies. (Input files can be found under ‘Data/SGVFinder/’)
  • Fig. 1d: uses manually analyzed data from CAT and SVfinder.
  • Fig. 2: uses ICRA generate coverage data for plasmids and phages based on data from the German and Spanish studies. (Input files can be found under ‘Data/ICRA/’)
  • Fig. 3a-b: uses corncob outputs on data from the Spanish study. (Input files can be found under ‘Data/corncob/results/’)
  • Fig. 3c-d: uses PDAC predictions with LightGBM models trained on different combinations of plasmids, phages and bacteria. (Input files can be found under ‘Data/LightGBM_predictions/’)
  • `Fig. 4a-f: uses ICRA generate coverage data for plasmids and phages as well as mOTUs bacterial relative abundance based on data from the German and Spanish studies. (Input files can be found under ‘Data/ICRA/’ and ‘Data/mOTUs/’)
  • Fig. 4g: uses corncob outputs on data from the Spanish study. (Input files can be found under ‘Data/corncob/’)
  • Fig. 4h: uses corncob outputs on data from the Spanish study. (Input files can be found under ‘Data/LightGBM_predictions/’)
  • Supplementary Fig. 1: uses SGVFinder output based on data from the Spanish study. (Input files can be found under ‘Data/SGVFinder/’)
  • Supplementary Fig. 2: uses mOTUs outputs based on data from all seven studies. (Input files under ‘Data/mOTUs/’)
  • Supplementary Fig. 3: uses mOTUs outputs based on data from the Spanish and German studies. (Input files under ‘Data/mOTUs/’)
  • Supplementary Fig. 4: uses PDAC predictions with LightGBM models trained on different combinations of plasmids, phages and bacteria. (Input files can be found under ‘Data/LightGBM_predictions/’)
  • Supplementary Figs. 5-6: uses ICRA generate coverage data for plasmids and phages as well as mOTUs bacteria relative abundance based on data from the German and Spanish studies. (Input files can be found under ‘Data/ICRA/’ and ‘Data/mOTUs/’)
  • Supplementary Fig. 7: uses PDAC stages I/II predictions with LightGBM models trained on different combinations of plasmids, phages and bacteria. (Input files can be found under ‘Data/LightGBM_predictions/’)