TSC2024

Code and data accessibility for the manuscript "Single-nucleus profiling of cortical tubers in tuberous sclerosis complex shows molecular structure preservation and massive reorganization of metabolism." authored by Sørensen, FNF. et al.

Compiled notebooks and analysis from the paper

HTML and markdown files with complete vignettes will be available upon publication.

The analysis flow is organized sequentially from 1 through 6. If you want to skip preprocessing, you can start at 3.all_cells and follow the scripts from there.

• 1.preprocessing
• 2.preprocessing_cb
• 3.all_cells
• 4.glia
• 5.neurons
• 6.smFISH

The neuron analysis workflow includes several sub-workflows:

• GWAS integration with the CELLEX/CELLECT pipeline (GWAS_CELLEX_CELLECT)
• Transcription factor inference with decoupleR/collecTRI (TF_inference)
• CellChat ligand-receptor interactions (Cellchat)
• Single-cell genotyping of probe based hybridization capture of TSC2 cDNA transcripts from same cells as used in 10x snRNA-seq (scgenotype)
• Candidate Caudal Late Inhibioty Progenitor (CLIP) cell identification (CLIP_cell_identification)
• Assesment of tissue origin signatures on expression (postmortem and surgery signatures from control and TSC patients respectively) (Tissue_origin_signatures)

Before starting these make sure to complete the previous workflows 3.all_cells, 4.glia and 5. neurons (TSC2_Neurons.rmd notebook)-

Count matrices and other data files

Data files have been encrypted in one file TSC2024_data.tgz and can be downloaded here or with wget in terminal:

"https://kkh.bric.ku.dk/Frederik/TSC2/TSC2024_data_files.tgz"

The data file folder contains folders analogous to the format as above. Each folder contains the necessary data files for the corresponding folder in the github repository.

• 0.10x_count_matrices
• 0.cellbender_counts
• 1.preprocessing
• 2.preprocessing_cb
• 3.all_cells
• 4.glia
• 5.neurons
• 6.smFISH

10x_count_matrices contains 10x Genomics filtered count matrices after cell calling.

cellbender_counts holds CellBender matrices and additional sample information from the pipeline.

Post-preprocessing and QC-filtered count matrices (cms_final_filter.rds) will serve as the starting object for 3.all_cells, 4.glia, and 5.neurons.

Annotations are supplied at four resolutions

• l1 = low resolution
• l2 = mid/low resolution
• l3 = mid/high resolution
• l4 = high resolution

Annotations are stored as factors in a list (e.g., annotation$l2).

All sample names are already pseudo-anonymized from the primary HPC/research institutions.

Please see the file identifier_list.csv.

---

--- April 2026 ---