Remove any R package installation from notebooks or scripts, especially with BiocManager

[jaclyn-taroni]

Context: #689 - which is adding a script to specifically install Bioconductor packages on the project Docker container.

Having logic that checks whether or not a package is installed and if not installs it is reasonable if someone were to run an analysis outside the Docker container. But if we were to miss something in the Docker image and then run an analysis it could end up unpinning versions (as discovered on the journey of #684). So, we should remove any package installation from notebooks or scripts, especially any installation via BiocManager.

Here's a list from @sjspielman:

focal-cn-file-preparation/04-prepare-cn-file.R:  install.packages("BiocManager")
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("GenomicRanges", update = FALSE)
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("IRanges", update = FALSE)
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("annotatr", update = FALSE)
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("TxDb.Hsapiens.UCSC.hg38.knownGene", update = FALSE)
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("org.Hs.eg.db", update = FALSE)
focal-cn-file-preparation/04-prepare-cn-file.R:  BiocManager::install("AnnotationDbi", update = FALSE)
gene-set-enrichment-analysis/01-conduct-gsea-analysis.R:## Load and/or install libraries ##
gene-set-enrichment-analysis/01-conduct-gsea-analysis.R:  install.packages("msigdbr")
gene-set-enrichment-analysis/01-conduct-gsea-analysis.R:  install.packages("BiocManager")
gene-set-enrichment-analysis/01-conduct-gsea-analysis.R:  BiocManager::install("GSVA")
molecular-subtyping-ATRT/02-ATRT-molecular-subtyping-plotting.R:  install.packages("ComplexHeatmap")
molecular-subtyping-ATRT/02-ATRT-molecular-subtyping-plotting.R:  install.packages("matrixStats")
molecular-subtyping-ATRT/02-ATRT-molecular-subtyping-plotting.R:  install.packages("ggfortify")
oncoprint-landscape/01-plot-oncoprint.R:  install.packages("BiocManager")
oncoprint-landscape/01-plot-oncoprint.R:  BiocManager::install("maftools")
cnv-chrom-plot/gistic_plot.Rmd:  install.packages("BiocManager")
cnv-chrom-plot/gistic_plot.Rmd:  BiocManager::install("ggbio")
cnv-comparison/01-cnv-comparison-plotting.Rmd:  install.packages("BiocManager")
cnv-comparison/01-cnv-comparison-plotting.Rmd:  BiocManager::install("GenVisR")
cnv-comparison/01-cnv-comparison-plotting.Rmd:  install.packages("cowplot")
molecular-subtyping-ATRT/01-ATRT-molecular-subtyping-data-prep.Rmd:  install.packages("DT")
molecular-subtyping-ATRT/01-ATRT-molecular-subtyping-data-prep.Rmd:  install.packages("htmlwidgets")
mutational-signatures/mutational_signatures.Rmd:  BiocManager::install("BSgenome.Hsapiens.UCSC.hg38")
mutational-signatures/mutational_signatures.Rmd:  install.packages("deconstructSigs")
mutect2-vs-strelka2/01-set-up.Rmd:    install.packages("BiocManager")
mutect2-vs-strelka2/01-set-up.Rmd:  BiocManager::install("maftools")
mutect2-vs-strelka2/01-set-up.Rmd:  install.packages("hexbin")
mutect2-vs-strelka2/01-set-up.Rmd:  install.packages("R.utils")
mutect2-vs-strelka2/02-analyze-concordance.Rmd:  install.packages("hexbin")
mutect2-vs-strelka2/02-analyze-concordance.Rmd:  devtools::install_github("clauswilke/colorblindr")
mutect2-vs-strelka2/02-analyze-concordance.Rmd:  install.packages("R.utils")
mutect2-vs-strelka2/02-analyze-concordance.Rmd:  install.packages("VennDiagrams")
snv-callers/compare_snv_callers_plots-tcga.Rmd:  install.packages("ggupset", repos = "http://cran.us.r-project.org")
snv-callers/compare_snv_callers_plots-tcga.Rmd:  install.packages("UpSetR", repos = "http://cran.us.r-project.org")
snv-callers/compare_snv_callers_plots-tcga.Rmd:  install.packages("GGally", repos = "http://cran.us.r-project.org")
snv-callers/compare_snv_callers_plots.Rmd:  install.packages("ggupset", repos = "http://cran.us.r-project.org")
snv-callers/compare_snv_callers_plots.Rmd:  install.packages("UpSetR", repos = "http://cran.us.r-project.org")
snv-callers/compare_snv_callers_plots.Rmd:  install.packages("GGally", repos = "http://cran.us.r-project.org")
snv-callers/vaf_cutoff_experiment.Rmd:  install.packages("GGally")
snv-callers/vaf_cutoff_experiment.Rmd:  install.packages("ggupset")

[cansavvy]

mutect2-vs-strelka2 is long retired and isn't run in CircleCI tests, do we still want to update it?

[cansavvy]

I started working on this with some of the notebooks I originally created, I can go ahead and remove the R package installations for all of these listed (and if I find any others).

[cbethell]

I have removed the package installation steps from oncoprint-landscape/01-plot-oncoprint.R in the open PR #666 (I can revert this change if we would prefer that these steps be removed all in one PR however).

[cansavvy]

@cbethell No need to change it. I will do multiple PRs for these anyway. I'll keep track of which modules are revised here as I open PRs for them.

[cansavvy]

Should I be ignoring deprecated/retired analyses, I assume?

[jaclyn-taroni]

What is the downside to fixing those? Lots of time to rerun I assume?

[cansavvy]

What is the downside to fixing those? Lots of time to rerun I assume?

Yes but mainly I’m encountering other errors not related to install.packages due to data release changes that haven’t been fixed since they’ve been deprecated. So it will involved me fixing errors that are outside this issue.

[jaclyn-taroni]

Ah, makes sense - skip 'em!

[cansavvy]

Keeping track of the changes I make here as I do them. (TBD means I haven't gotten to it yet.)

Module	Status	PR ref
chromosomal-instability	No changes needed	NA
cnv-chrom-plot	In gistic_plot.Rmd got rid of: install.packages("BiocManager") and BiocManager::install("ggbio") and refreshed gistic_plot.Rmd only.	#698
cnv-comparison	Skipped because deprecated	NA
collapse-rnaseq	No changes needed	NA
comparative-RNASeq-analysis	No changes needed	NA
copy_number_consensus_call	No changes needed	NA
create-subset-files	No changes needed	NA
focal-cn-file-preparation	Got rid of series of installs in 04-prepare-cn-file.R, re-ran run-prepare-cn.sh	#698
fusion_filtering	No changes needed	NA
fusion-summary	No changes needed	NA
gene-set-enrichment-analysis	Got rid of installations and re-in 01-conduct-gsea-analysis.R and re-ran run-gsea.sh	#703
immune-deconv	No changes needed	NA
independent-samples	No changes needed	NA
interaction-plots	No changes needed	NA
molecular-subtyping-ATRT	Removed installs from 02-ATRT-molecular-subtyping-plotting.R and 01-ATRT-molecular-subtyping-data-prep.Rmd and re ran bash run-molecular-subtyping-ATRT.sh	#703
molecular-subtyping-EPN	No changes needed	NA
molecular-subtyping-EWS	No changes needed	NA
molecular-subtyping-HGG	No changes needed	NA
molecular-subtyping-LGAT	No changes needed	NA
molecular-subtyping-SHH-tp53	Skipped because deprecated	NA
molecular-subtyping-chordoma	No changes needed	NA
molecular-subtyping-embryonal	No changes needed	NA
molecular-subtyping-pathology	No changes needed	NA
mutational-signatures	Got rid of installs in mutation_signatures.Rmd and reran	TBD
mutect2-vs-strelka2	Skipped because deprecated	NA
oncoprint-landscape	Got rid of maftools install in 01-plot-oncoprint.R and reran run-oncoprint.sh	#704
rna-seq-composition	No changes needed	NA
run-gistic	No changes needed	NA
sample-distribution-analysis	No changes needed	NA
selection-strategy-comparison	Skipped because deprecated	NA
sex-prediction-from-RNASeq	No changes needed	NA
snv-callers	Got rid of installs in the compare_snv_callers_plots.Rmd and compare_snv_callers_plots-tcga.Rmd notebooks and reran on AWS	#701
ssgsea-hallmark	Skipped because deprecated	NA
sv-analysis	No changes needed	NA
survival-analysis	Got rid of install in util/survival_models.R and reran survival-analysis_template.Rmd	#704
telomerase-activity-prediction	No changes needed	NA
tmb-compare-tcga	No changes needed	NA
tp53_nf1_score	No changes needed	NA
transcriptomic-dimension-reduction	Get rid of package installations in run-dimension-reducion.R and reran ./dimension-reduction-plots.sh	#704
tcga-capture-kit-investigation	No changes needed	NA