Top-level:
config.sh is to get the relevant file locations
load_fusioncalls.R loads the processed, filtered fusions (called by R scripts in fusions/)
load_metadata.R loads the metadata (called by multiple R scripts)

In fusions/ is the analysis of gene fusions in 10x:
fusions/compute_signature_scores.py computes expression data for the signatures analysis
fusions/process_fusions_*.R are scripts to load initial fusions calls, filter them, aggregate, and create tables for downstream analyses
fusions/stats_fusions_*.R are scripts for statistical analysis of the fusions data, most important are:
- 01 (regression across conditions)
- 01b (statistics)
- 01c (gene set overlaps)
- 04 (gene set signatures)
- 05, 06 (regressions evaluated by different splits of the data, robustness analyses)

metadata/ contains internal scripts to align the 10x metadata with the batch numbers
preprocess/ contains scripts for re-aligning, running STAR-fusion, assigning barcodes, and aggregating outputs.