Data-driven learning how oncogenic gene expression locally alters heterocellular networks

This repository supplies the code developed in the study of D.J. Klinke, A. Fernandez, W. Deng, H. Latifizadeh, and A.C. Pirkey "Data-driven learning how oncogenic gene expression locally alters heterocellular networks". The corresponding publication can be found in Nature Communications ( Apr 13;13(1):1986. doi: 10.1038/s41467-022-29636-3; PMID: 35418177; PMCID: PMC9007999). It can be used to reproduce the results of the study and investigate the methodology to be used for other datasets.

System Requirements

Hardware Recommendations:

The code contained in this repository requires only a standard computer to run, however, the amount of RAM required will be defined by the desires of the use. For low level performance, 2 GB of RAM should suffice. For optimal performance, a minimum of 16 GB of RAM and 8 Cores with 2.9 GHz/core is recommended. These specifications match the laboratory computer specifications that the code was tested on.

The runtimes below were generated with a computer with the optimal specifications (16 GB RAM and 8 Cores @ 2.9 GHz/core) and an internet download speed of 791 Mbps.

Software Recommendations:

• Computer Operating System: Tested on Windows 10, version 21H2
• The code was originally written using R version 3.6.3. along with a corresponding version of RStudio. Most recent testing done using R version 4.1.0 and RStudio version 1.4.1717
• To use the provided code, you will need to install the following R libraries: BiocManager (v. 1.30.16), SummarizedExperiment (v. 1.22.0), TCGAbiolinks (v. 2.20.1), colorspace (v. 2.0.2), MASS (v. 7.3.54), bnlearn (v. 4.7), Rgraphviz (v. 2.36.0), parallel (v. 4.1.0), stats (v. 4.1.0), lattice (v. 0.20.44)

Data

All necessary data is provided via links to public data provided in the scripts.

For more information about data used in this study please refer to [https://portal.gdc.cancer.gov] for Files: the breast cancer (BRCA) and Skin Cutaneous Melanoma (SKCM) arms of the Cancer Genome Atlas downloaded via `TCGAbiolinks' (V package in R accessed 6/27/2017; and [https://www.ncbi.nlm.nih.gov] entry GSE98394.

Instructions

Installation Guide

• The latest version of R can be downloaded at https://www.r-project.org/ and should download in under one minute
• The latest version of RStudio can be downloaded at https://www.rstudio.com/products/rstudio/download/ and should install in under one minute
• Download the files from the repository and extract the files from the zip folder to your desired working folder. The extraction process should be near instantaneous.
• Open the six provided R scripts and change the address in the setwd() command on each file to reflect the folder you saved the files to. The scripts should open instanteously.
• Ensure all R Libraries listed in the Software Requirements are installed. Code for the installation of all packages is provided below.

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
    
    BiocManager::install("SummarizedExperiment")
    BiocManager::install("TCGAbiolinks")
    BiocManager::install("Rgraphviz")
    
    install.packages(c('colorspace','MASS', 'lattice'))
    install.packages("https://www.bnlearn.com/releases/bnlearn_latest.tar.gz", repos = NULL, type = "source")

Demo/Instructions for Use

Running the scripts in the order provided will generate all relevant figures and data tables for the given portion of the study. Below you'll find a list of each script, it's run time, the files added to the working directory, and the number of figures generated.

• 1_BRCA_PreProcess.R

  • Run time: 14.5 minutes
  • Files created:
    • GDCdata file folder
    • BRCA_RNAseq_HK.txt
    • BRCA_TPM_HK.Rda
    • brca-RNAseq-Counts.rda
  • Figures generated: 1

• 2_BRCA-EMT-SM.R

  • Run time: < 10 seconds
  • Files created:
    • BRCA_EMT_SM.csv
  • Figures generated: 2

• 3_BRCA-Proliferation-SM.R

  • Run time: < 10 seconds
  • Files created:
    • BRCA_proliferation_Zscore.csv
  • Figures generated: 1

• 4_Merge_Datasets.R

  • Run time: < 10 seconds
  • Files created:
    • BRCA_digitalcytometry_tot_noise.csv
  • Figures generated: 7

• 5_BRCA_BN_SeedEdges_Interval.R

  • Run time: 23 minutes
  • Files created:
    • BRCA_HYBRID_CIBERSORT Bootstrap 10000_[insert algorithm].csv
      • algorithms: gs, rsmax2, mmhc, tabu, hc, pc_stable, iamb_fdr, si_hiton_pc, mmpc, iamb
    • BRCA-Hybrid-CorrelationStructure.pdf
    • BRCA-HybridVariableDist.pdf
    • BRCA_hybrid_CIBERSORT_Seed_Summary_Int.csv
  • Figures generated: 1

• 6_BRCA_BN_EnsembleDAG.R

  • Run time: 2 minutes
  • Files created:
    • BRCA_CIBERSORT_NKsubset_L1_mmhc.csv
    • BRCA_CIBERSORT_NKsubset_mmhc_Hybrid_Merge.csv
    • BRCA_ModelvsData-hybrid
  • Figures generated: 1

General notes

The code provided in this repository reproduces the main results of the study of D.J. Klinke, A. Fernandez, W. Deng, H. Latifizadeh, and A.C. Pirkey "Data-driven learning how oncogenic gene expression locally alters heterocellular networks" but it is not meant as a self-contained module for use in further analysis.

Citation

Klinke, D.J., Fernandez, A., Deng, W., Latifizadeh, H. & Pirkey, A.C., "Data-driven learning how oncogenic gene expression locally alters heterocellular networks." Nature Communications (2022). Apr 13;13(1):1986. doi: 10.1038/s41467-022-29636-3. PMID: 35418177; PMCID: PMC9007999.