TCGA Glycogene Metaanalysis

Scripts and Pipelines used to analyze the TCGA for patterns in Glycogene Expression across Cancer Types

Data Gathering Using TCGABiolinks:

The "TCGAbl_gather_data.R" gathers TCGA data if passed a valid TCGA code. If in an HPC environment all datasets can be downloaded simultaneously using independent processes using a slurm script formatted in a manner similarly to "slurm_TCGAblRun" script. The TCGAbl_gather_data.R script utilizes functions in the "TCGAbl_functions.R" file.

Directories:

glycoBio_data:

This directory contains the GlycoOnto.owl file, as well as R scripts used to create pathway and function lists used for the enrichment analyses. Running the "make_pathLists.sh" script will regenerate the "pathwayList.rda" and "functionList.rda" lists. The script runs the following in order:

1. glycoOnto_Graph.R will do the following to get the ontology into R:

• Converts ontology into a JSON file using the "ROBOT" package
• JSON ontology is read into R, and glycogene and pathway relationships are stored into a nested list structure. This nested list structure is then saved for use in step 2.

2. create_pathLists.R will read the nested list structure to create path lists, where each pathway has a list of glycogenes which fit into each pathway classification.

The output from this pipeline will be two files: "pathwayList.rda" and "functionList.rda", which are used for enrichments after differential expression analyses.

analyses:

This directory contains analysis scripts for each section of the manuscript:

• t-SNE embeddings
• Consensus Clustering
• Differential expression analysiss
• Enrichment analyses
• Breast Cancer Glyco50 classifier code
• Glyco50 Cancer Progression Maps