transcriptomics

Transcriptomics pipeline

R pipeline that builds the mechanosensitive gene panel and fits its fibrosis-stage expression trajectories from three human liver RNA-seq cohorts (GSE130970, GSE135251, GSE162694). It runs from the uploaded gene list through the per-gene model fits used to validate the virtual-cell model.

Scripts (run in order)

Script	Purpose
0_update_gene_list.R	Add TEAD1–4 / SRF, fix symbols, annotate Ensembl IDs → data/0_genes_nucleo_all_updated.tsv
1_Get_collapsse_data.R	Download the three GEO datasets, collapse to common genes, build the count and metadata masters
1_1_dds.R	Per-dataset DESeq2 normalization over F0–F4, sex effect removed
1_2_Stage_mean_genes.R	Mean panel expression per fibrosis stage × dataset
1_3_Fit_filter.R	Fit linear / power-law / sigmoid per gene (best by AIC), keep direction-concordant genes

Each script resolves its paths relative to its own location, so run them from anywhere (Rscript transcriptomics/0_update_gene_list.R) without editing paths.

Layout

transcriptomics/
├── data/                 gene panel and sample metadata
├── geo/
│   ├── raw_downloads/     GEO downloads (not tracked)
│   └── R_objects/         per-dataset metadata + count objects
└── results/
    ├── normalization/     per-dataset DESeq2 output (meta_filtered_*.tsv)
    ├── stage_means/       mean_expr_per_stage.tsv
    └── model_fits/        model_fits.tsv, conserved_genes.tsv

Small summary tables are tracked. Heavy, regenerable artifacts (raw downloads, count matrices, dds_*.rds, norm_counts_*.tsv, .RData) are git-ignored and rebuilt by rerunning the scripts.

Sample selection

Only fibrosis-graded NAFLD/NASH biopsies enter the fit. Histologically normal controls (Control_Normal) are excluded, and F0 here means NAFLD with steatosis but no fibrosis — not a healthy baseline. Samples with missing sex are dropped before DESeq2, so not every F0 biopsy is used (e.g. GSE130970 keeps 18 of 21).

data/samples_used_metadata.tsv lists the 391 samples that survive filtering and feed the DESeq2 fits:

Dataset	F0	F1	F2	F3	F4	Total
GSE130970	18	28	9	14	2	71
GSE135251	38	48	54	54	14	208
GSE162694	35	30	27	8	12	112
All	91	106	90	76	28	391

Columns: sample_id, dataset, stage, sex, age, steatosis_grade, nas_score, true_control, disease, histology_group.

Cohort background is documented in ../data/RANseq_datasets_info.md.