README.md

# Squadrito_LiverTumor2022_scRNAseq

Kerzel T, Giacca G, Beretta S, Scotti MG, et al.
**Targeted Interferon Delivery Through In Vivo Engineered Macrophages Eradicates Liver Metastases.**
2022.
GEO: [GSE221357](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE221357)

---

### Directories and Files ###
- SampleSheet.csv: names of samples and corresponding treatment condisions
- scripts: folder with R scripts used for the analyses
  - `0_PreProcessData.R`: initial analysis with DoubletFinder and Full object creation;
  - `1_Analysis_TNK.R`: subset of T and NK, and analysis;
  - `2_Analysis_APC.R`: subset of APC and analysis;
  - `3_Analysis_UndRem.R`: integration of subset analyses and _undetermined_ removal;
  - `4_Paper_Results.R`: post-analysis and visualization for paper;
  - `5_Analysis_TAMS_harmony.R`: subset of TAMs and analysis with harmony.
- data: results of the analyses
  - Full_final_*: results of Full object analysis;
  - APCs_cells_*: results of APC analysis;
  - T_NK_cells_*: results of T and NK analysis;
  - scRNAseq_*.xlsx: excel files with markers.

---

### Analyses ###

Single-cell (from 10X Genomics) analysis of 8 liver tumoral samples.

Samples were merged into a single Seurat dataset, which was pre-processed to remove low-quality cells, that is, those with a feature count below 1000 and above 6000, as well as cells with a fraction of mitochondrial genes higher than 10%.
Afterwards, cells annotated as doublets with DoubletFinder were excluded as well from the analysis with Seurat.</br>
RNA UMI-counts were normalized using a global-scaling normalization method and the Variance Stabilizing Transformations (SCTransform) was performed to scale based on the percentage of mitochondrial genes, the absolute count of RNAs in each cell, and the difference between S and G2/M cell cycle scores computed for each cell.
A principal component analysis with 50 principal components (PCs) was performed for dimensional reduction, and a UMAP-representation as well as clusterswere computed on those reductions.</br>
Marker genes for each cluster were obtained using the FindAllMarkers Seurat function and, consequently, clusters were manually annotated (removing a small population of undefined cells).</br>
Analysis of the subclusters "T and NK cells" and "APCs" was performed accordingly.</br>
Top upregulated markers of each population were calculated based on the FindAllMarkers function and a heatmap was generated based on the top 20 upregulated genes in each cluster to represent them.
For the calculation of differentially expressed genes within individual clusters comparing the different treatment cohorts, namely control, partial responders and resistant, the FindMarker function was utilized.
For GSEA the gene sets from [MSigDB](https://www.gsea-msigdb.org/gsea/msigdb/genesets.jsp) were used.