Notebooks and helper functions accompanying the paper: LIPOGRID: A HIGH-THROUGHPUT MULTI-OMICS PERTURBATION SCREEN DISSECTS THE GENETIC ARCHITECTURE OF LIPID METABOLISM. Analyze co-registered (see FOCUS) MALDI-MSI and spatial transcriptomic data, impact of gene KOs on lipidome and transcriptome
The notebooks reference this folder as ../data/... (a sibling of ../../../Desktop/LipoGrid/notebooks and
../../../Desktop/LipoGrid/scripts at the repo root). None of the data itself is tracked in this repo — this
file documents what each notebook expects to find here and what it writes back, so
you can populate the folder before running anything.
Paths below are relative to ../../../Desktop/LipoGrid/data.
Everything specific to this project's LipoGrid screen.
| Path | What it is | Produced by |
|---|---|---|
gRNA_oligos/gRNA_oligos.csv |
gRNA oligo sequences for the library | external (input) |
gRNA_oligos/oligo_pool_plasmid.fa, .gtf, _structure.gtf |
Custom reference annotation for the oligo pool | CROP_seq_analyses.ipynb (write_annotation(), run once, not automatic) |
mapping/.../outs/filtered_feature_bc_matrix.h5 |
CellRanger outputs, RNA and gRNA libraries, 2 sequencing runs each | external (input, from CellRanger) |
analyses/count_matrix_filtered.h5ad |
QC-filtered CROP-seq count matrix (all cells) | CROP_seq_analyses.ipynb |
analyses/count_matrix_filtered_g.h5ad |
Same, restricted to cells with a single confidently-assigned gRNA | CROP_seq_analyses.ipynb |
| Path | What it is | Produced by |
|---|---|---|
output-XETG*_count_matrix_*.h5ad (4 slides) |
Raw per-cell Xenium count matrices, including gRNA probe counts, watershed/segmentation-expanded | external (input, from Xenium pipeline) |
xenium_david_{1..4}_with_grna_*.h5ad |
Same 4 slides with a called gRNA added to .obs['gRNA'] |
Xenium_gRNA_assignment.ipynb |
The xenium_david_*_with_grna_*.h5ad files are then fed into an external MALDI-MSI/Xenium
coregistration pipeline (not part of this repo) that produces the two files under
MALDI_MSI/d_lipogrid/merged/ below.
| Path | What it is | Produced by |
|---|---|---|
d_lipogrid/merged/preprocessing/MSI_merged_processed.h5ad |
Merged, preprocessed MALDI-MSI data, all 4 runs | external coregistration pipeline (input) |
d_lipogrid/merged/alignment/Xenium_merged_processed_aligned.h5ad |
Xenium cells with coordinates aligned to the MSI raster | external coregistration pipeline (input) |
The bulk of the pipeline's outputs and figures live here.
| Path | What it is | Produced by |
|---|---|---|
final_4_runs/msi_int_cells.h5ad |
Per-cell MSI lipid intensity profiles (inverse-distance-weighted from nearby MSI spots) | alignment_MALDI_Xenium_percell.ipynb |
final_4_runs/filtered_143target_genes.txt |
The 143 gene targets shared between the CROP-seq and MSI screens | external / curated (input) |
final_4_runs/all_lipid_{pos,neg}_log2FC_pvalues_per_gene.csv |
Per-gene-KO lipid log2FC + Mann-Whitney p, computed separately per ion mode | LipoGrid_core_analysis_lipidomics_geneKO.ipynb |
final_4_runs/all_lipid_log2FC_pvalues_per_gene.csv |
Combined pos+neg lipid statistics (one row per lipid, best-detected mode) — the main MSI results table | LipoGrid_core_analysis_lipidomics_geneKO.ipynb |
CROP_seq_log2FC_manw_pergeneKO_final.csv |
Per-gene-KO RNA expression log2FC + p-value from CROP-seq | CROP_seq_analyses.ipynb |
CROP_seq_log2FC_manw_pergeneKO_sameKOs.csv |
Same, restricted to the 143 shared KO genes | CROP_seq_analyses.ipynb |
CROP_seq/target_genes_CROPseq-clusters.tsv, genes_CROPseq_10cols_clusters.tsv |
Gene <-> cluster assignments from the CROP-seq log2FC heatmap | CROP_seq_analyses.ipynb |
CROP_seq/enrichments/* |
GSEA result tables and figures | GSEA_CROP_seq.ipynb |
internalnorm_finalfigs/, other *.pdf |
Final figures | all notebooks |
| Path | What it is | Produced by |
|---|---|---|
BulkLipidomics_quant_species_nmol_mgDNA.tsv |
Orthogonal bulk lipidomics validation (shRNA knockdowns) | external (input) |
Public gene-set collections in .gmt format, used by GSEA_CROP_seq.ipynb
(download from MSigDB, Reactome, KEGG, or
PathBank — not redistributed here):
hallmark.gmt, reactome.gmt, kegg.gmt, go_bp.gmt, c2.cp.symbols.gmt,
c2.cp.biocarta.symbols.gmt, c2.cp.pid.symbols.gmt, c2.cp.wp.symbols.gmt,
c5.go.v2024.1.Hs.symbols.gmt, pathbank.Homo_sapiens.symbols.gmt
lipid_terms.pathways.gmt — a lipid-keyword-filtered, cross-database merge of the
above, generated by GSEA_CROP_seq.ipynb itself.
MS1_database_{POS,NEG}_JI_JD_XS_NR_Washed.csv — external m/z-to-lipid annotation
databases (positive/negative ionization mode), used by alignment_MALDI_Xenium_percell.ipynb
to independently cross-check the lipid annotations coming from the MALDI-MSI pipeline.
Optional. Used for a consistent publication font across figures; every notebook falls back to Helvetica/Arial/DejaVu Sans automatically if this file isn't present.
Xenium_gRNA_assignment.ipynb- (external coregistration pipeline — not in this repo)
alignment_MALDI_Xenium_percell.ipynbLipoGrid_core_analysis_lipidomics_geneKO.ipynbCROP_seq_analyses.ipynbGSEA_CROP_seq.ipynb(needs outputs of both 4 and 5)comparison_bulkvalidations.ipynb(needs output of 4)