Skip to content

r_dependencies.md

Latest commit

 

History

History
53 lines (41 loc) · 1.86 KB

r_dependencies.md

File metadata and controls

53 lines (41 loc) · 1.86 KB
Raw

R Dependencies

GeneLab benchmark preprocessing requires R 4.2+ with Bioconductor packages. Python analysis (run_baselines.py, evaluate_submission.py) does not require R.

Installation

# Install Bioconductor package manager
if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

# Core preprocessing packages
BiocManager::install(c(
    "DESeq2",    # Size-factor normalization (per-mission)
    "limma",     # Linear model for differential expression
    "edgeR",     # Count data normalization utilities
    "GSVA",      # Gene Set Variation Analysis (pathway scores)
    "fgsea"      # Fast gene set enrichment
))

# CRAN packages
install.packages(c(
    "sva",       # ComBat-seq batch correction
    "ggplot2",   # Visualization
    "dplyr",     # Data manipulation
    "readr"      # CSV I/O
))

R Scripts and Their Dependencies

Script R Packages Used Purpose
scripts/normalize_*.R DESeq2, edgeR Per-mission size-factor normalization
scripts/batch_correct.R sva (ComBat-seq) Batch correction (Category J analysis)
scripts/compute_pathway_scores.R GSVA, limma GSVA Hallmark 50-pathway scores
scripts/run_fgsea.R fgsea, limma Fast gene set enrichment analysis
scripts/_install_deps.R — Interactive dependency installer

Notes

  • DESeq2 normalization is applied per-mission (not joint across missions): joint normalization would mix mission-specific library size effects into the spaceflight signal.
  • Python preprocessing (quality_filter.py, generate_tasks.py) is applied after R normalization.
  • R scripts produce per-mission *_log2_norm.csv files in processed/A_detection/{tissue}/.
  • The final {tissue}_all_missions_log2_norm.csv is assembled in Python.
  • R ≥ 4.2 required; Bioconductor 3.16+ recommended.