Coupling of MAPK output and lineage identity defines archetypal modes of TKI resistance in EGFR-mutant lung adenocarcinoma

This repository contains code related to the manuscript:

Zatzman, M., Quintanal-Villalonga, A. et al, Coupling of MAPK output and lineage identity defines archetypal modes of TKI resistance in EGFR-mutant lung adenocarcinoma.

Overview

The repository is structured as an installable R package. Everything was written and rendered under R 4.3.3 with Bioconductor 3.18; renv.lock pins every package version.

The notebooks split into two kinds. Figure notebooks run start to finish against the deposited data and reproduce the published panels; they are rendered as a website at https://shahcompbio.github.io/egfr_nucseq. Processing notebooks document how the deposited objects were produced, from FASTQs onward, and are not runnable outside the original compute environment.

Where to look

code/README.md	every notebook, in pipeline order, and where each derivation lives
code/xenium/README.md	the 389-plex panel, the two ingestion branches, what the Xenium deposit holds
R/	processing, plotting and helper functions
rendered notebooks	the figure notebooks, executed

Quick start

git clone https://github.com/shahcompbio/egfr_nucseq
cd egfr_nucseq

1. renv::restore() — rebuilds the pinned library.
2. Render code/00_download_convert.qmd — lists what to download from the Zenodo deposit and builds paper_data/ from it. paper_data/ is not in version control.
3. quarto render — runs the nine figure notebooks. Panels are written to plots/, the site to docs/.

Each figure notebook loads only the objects its own panels need and is independent of the others, so any one can be rendered on its own:

quarto render code/91_figure3.qmd

Figure notebooks

In paper order:

notebook	figures
01_cohort_metadata.qmd	1 (swimmer plot and cohort timeline)
31_impact_analysis.qmd	S2 (MSK-IMPACT panel sequencing)
90_figure2.qmd	2, S3–S5
91_figure3.qmd	3, S6–S8
92_figure4.qmd	4, S9
33_luca_analysis.qmd	4C–D, S7C, S9C (Lung Cancer Cell Atlas)
34_bulk_transformation.qmd	S9D, S9G (published transformation cohorts)
93_figure5.qmd	5, S10
xenium/99_xenium_plots.qmd	6A–G, S11–S13

Data

deposit	holds
GEO GSE292469 — single-nucleus RNA-seq	FASTQs and raw per-library count matrices for all 74 sequencing libraries, plus Series-level GSE292469_counts.h5 and GSE292469_cell_metadata.tsv.gz
GEO GSE342824 — Xenium in situ	per-section XeniumRanger output for all ten sections — counts, cell and nucleus boundaries, decoded transcripts, morphology images — plus xenium_cell_metadata.tsv.gz and the 389-target gene_panel.json
Zenodo — DOI at publication	the figure-reproduction bundle this repository runs off: the three single-nucleus Seurat objects, the Xenium spatial object, and the derived result tables the notebooks read
CELLxGENE — link at publication	the cohort, epithelial and tumor histotime objects, for interactive browsing and download