run_all_analyses_and_plots.sh

#!/bin/sh
set -euo pipefail

# This script runs analysis code and plotting scripts for the publication,
# including both cancer types (both) or just one cancer type (BRCA or GBM).
# The script calls:
#  1. run_machine_learning_experiments.sh
#  2. run_differential_expression_experiments.sh (subtype only)
#  3. plots/scripts/visualize_expression.R (subtype only)
#  4. plotting scripts, as appropriate

# cancer type (must be both, BRCA, or GBM)
cancer_type=$1

if [ $cancer_type != "both" ] && [ $cancer_type != "BRCA" ] && [ $cancer_type != "GBM" ]; then
  echo Cancer type must be both, BRCA or GBM in run_all_analyses_and_plots.sh [cancer_type]
  exit
fi

################################################################################
# BRCA
################################################################################

if [ $cancer_type == "both" ] || [ $cancer_type == "BRCA" ]; then

  # BRCA subtype ---------------------------------------------------------------

  # run machine learning and DEG analysis scripts
  bash run_machine_learning_experiments.sh BRCA subtype 7
  bash run_differential_expression_experiments.sh BRCA Basal Her2,LumA Her2,LumA 7

  # plot array vs. RNA-seq expression levels after normalization
  Rscript plots/scripts/visualize_expression.R --cancer_type BRCA --predictor subtype

  # plot difference in subtype prediction kappa between non-reconstructed and reconstructed data
  Rscript plots/scripts/recon_kappa_difference.R --cancer_type BRCA --output_directory plots/supplementary

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor subtype \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor subtype \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type BRCA \
    --predictor subtype \
    --output_directory plots/main

  # violin + line plots showing kappa values from predictions on reconstructed test data
  Rscript plots/scripts/6-plot_recon_kappa.R \
    --cancer_type BRCA \
    --predictor subtype \
    --output_directory plots/supplementary

  # violin + line plots showing gene-level MASE values from reconstructed test data
  Rscript plots/scripts/6-plot_recon_error.R \
    --cancer_type BRCA \
    --predictor subtype \
    --output_directory plots/supplementary

  # violin plots showing proportion of pathways significant in PLIER analyses
  Rscript plots/scripts/7-plot_plier_pathways.R \
    --cancer_type BRCA \
    --predictor subtype \
    --output_directory plots/main

  # bar plot showing proportion of genes differentially expressed (Basal vs. Others)
  # line plot showing overlap with silver standard DEGs (Basal vs. Others)
  Rscript plots/scripts/1A-plot_DEGs.R \
    --cancer_type BRCA \
    --subtype_vs_others Basal \
    --proportion_output_directory plots/supplementary \
    --overlap_output_directory plots/supplementary \
    --overlap_measure Jaccard,Spearman
  
  # bar plot showing proportion of genes differentially expressed (Her2 vs. LumA)
  # line plot showing overlap with silver standard DEGs (Her2 vs. LumA)
  Rscript plots/scripts/1A-plot_DEGs.R \
    --cancer_type BRCA \
    --subtype_vs_subtype Her2,LumA \
    --proportion_output_directory plots/supplementary \
    --overlap_output_directory plots/main \
    --overlap_measure Jaccard,Spearman

  # line plot showing overlap with silver standard DEGs (Her2 vs. LumA) across small n values
  Rscript plots/scripts/2A-plot_small_n_differential_expression.R \
    --cancer_type BRCA \
    --subtype_vs_subtype Her2,LumA \
    --output_directory plots/main \
    --overlap_measure Jaccard,Spearman

  # ----------------------------------------------------------------------------

  # BRCA TP53 ------------------------------------------------------------------

  # run machine learning analysis scripts
  bash run_machine_learning_experiments.sh BRCA TP53 7

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor TP53 \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor TP53 \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type BRCA \
    --predictor TP53 \
    --null_model \
    --output_directory plots/supplementary

  # ----------------------------------------------------------------------------

  # BRCA PIK3CA ----------------------------------------------------------------

  # run machine learning analysis scripts
  bash run_machine_learning_experiments.sh BRCA PIK3CA 7

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor PIK3CA \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type BRCA \
    --predictor PIK3CA \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type BRCA \
    --predictor PIK3CA \
    --null_model \
    --output_directory plots/supplementary

  # ----------------------------------------------------------------------------

fi

################################################################################
# GBM
################################################################################

if [ $cancer_type == "both" ] || [ $cancer_type == "GBM" ]; then

  # GBM subtype ----------------------------------------------------------------

  # run machine learning and DEG analysis scripts
  bash run_machine_learning_experiments.sh GBM subtype 7
  bash run_differential_expression_experiments.sh GBM Proneural Classical,Mesenchymal Classical,Mesenchymal 7

  # plot array vs. RNA-seq expression levels after normalization
  Rscript plots/scripts/visualize_expression.R --cancer_type GBM --predictor subtype

  # plot difference in subtype prediction kappa between non-reconstructed and reconstructed data
  Rscript plots/scripts/recon_kappa_difference.R --cancer_type GBM --output_directory plots/supplementary

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor subtype \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor subtype \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type GBM \
    --predictor subtype \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on reconstructed test data
  Rscript plots/scripts/6-plot_recon_kappa.R \
    --cancer_type GBM \
    --predictor subtype \
    --output_directory plots/supplementary

  # violin + line plots showing gene-level MASE values from reconstructed test data
  Rscript plots/scripts/6-plot_recon_error.R \
    --cancer_type GBM \
    --predictor subtype \
    --output_directory plots/supplementary

  # violin plots showing proportion of pathways significant in PLIER analyses
  Rscript plots/scripts/7-plot_plier_pathways.R \
    --cancer_type GBM \
    --predictor subtype \
    --output_directory plots/main

  # bar plot showing proportion of genes differentially expressed (Proneural vs. Others, Classical vs. Mesenchymal)
  # line plot showing overlap with silver standard DEGs (Proneural vs. Others, Classical vs. Mesenchymal)
  Rscript plots/scripts/1A-plot_DEGs.R \
    --cancer_type GBM \
    --subtype_vs_others Proneural \
    --subtype_vs_subtype Classical,Mesenchymal \
    --proportion_output_directory plots/supplementary \
    --overlap_output_directory plots/supplementary \
    --overlap_measure Jaccard,Spearman

  # line plot showing overlap with silver standard DEGs (Classical vs. Mesenchymal) across small n values
  Rscript plots/scripts/2A-plot_small_n_differential_expression.R \
    --cancer_type GBM \
    --subtype_vs_subtype Classical,Mesenchymal \
    --output_directory plots/supplementary \
    --overlap_measure Jaccard,Spearman

  # ----------------------------------------------------------------------------

  # GBM TP53 -------------------------------------------------------------------

  # run machine learning analysis scripts
  bash run_machine_learning_experiments.sh GBM TP53 7

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor TP53 \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor TP53 \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type GBM \
    --predictor TP53 \
    --null_model \
    --output_directory plots/main

  # ----------------------------------------------------------------------------

  # GBM PIK3CA -----------------------------------------------------------------

  # run machine learning analysis scripts
  bash run_machine_learning_experiments.sh GBM PIK3CA 7

  # stacked bar plot showing distribution of subtypes in train/test sets (one representative example)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor PIK3CA \
    --output_directory plots/supplementary

  # stacked bar plots showing distribution of subtypes in train/test sets (all seeds)
  Rscript plots/scripts/0-plot_predictor_category_distributions.R \
    --cancer_type GBM \
    --predictor PIK3CA \
    --plot_all_seeds \
    --output_directory plots/supplementary

  # violin + line plots showing kappa values from predictions on test data
  Rscript plots/scripts/3-plot_category_kappa.R \
    --cancer_type GBM \
    --predictor PIK3CA \
    --null_model \
    --output_directory plots/supplementary

  # ----------------------------------------------------------------------------

fi

################################################################################
# PLIER pathway analysis of BRCA and/or GBM
################################################################################

Rscript -e "rmarkdown::render('8-PLIER_pathways_analysis.Rmd', clean = TRUE)"