Tools and pipelines from the lab, all open source and versioned. Issues and pull requests welcome.
Cohort-level trajectory alignment for single-cell data. It uses dynamic time-warping barycenter averaging to build a shared reference pseudotime across patients, then clusters gene trends into early, intermediate, and late modules that recur across the cohort. Applied to LUAD-to-SCLC and adeno-to-squamous transformation.
Analysis code for POsiE, a cohort of 11 patients with EGFR-mutant lung adenocarcinoma treated with osimertinib, some of which underwent neuroendocrine transformation. It combines clinical timelines, WES and MSK-IMPACT mutation tracking, single-cell RNA-seq with inferred copy number and RB1/APOBEC analysis, and RB1-knockout RNA-seq and ATAC-seq experiments, validated against TCGA.
scBLender: a set of plasticity scores built on time-course scRNA-seq, using the mixed basal-luminal state as a stand-in for plasticity. It identified JAK-STAT inflammatory signaling as a driver of plasticity in prostate cancer. Includes notebooks for the figures in Chan et al., Science 2022.
git clone https://github.com/joechanlab/Lineage_Plasticity_in_CRPC && conda create --name scblender --file requirements.txtSingle-cell atlas of small cell lung cancer built from scRNA-seq, MIBI, and Vectra imaging, as part of the Human Tumor Atlas Network. Contains code to reproduce the figures in Chan et al., Cancer Cell 2021.