Key Learning Deliverables & Competencies

• Identify Spatially Variable Genes (SVGs) utilizing Moran's I autocorrelation and nnSVG.
• Differentiate Highly Variable Genes (HVGs) from SVGs exhibiting spatial spatial dependency.
• Quantify microenvironmental cellular niches (tumor core, invasive stroma, tertiary lymphoid structures).
• Design publication-grade figures (300 DPI, viridis color scales, accessible palettes, physical scale bars).
• Assemble an end-to-end, reproducible computational analysis report in Quarto / R Markdown.

Moran's I Autocorrelation Statistic

$$I = rac{N}{W} rac{\sum_{i=1}^N \sum_{j=1}^N w_{ij}(x_i - ar{x})(x_j - ar{x})}{\sum_{i=1}^N (x_i - ar{x})^2}$$ Where $N$ is the number of spatial bins, $w_{ij}$ is the spatial weight matrix (e.g., inverse Euclidean distance), and $x_i$ is gene expression. An $I$ value near $+1$ indicates strong spatial clustering (focal expression gradients), while $I pprox 0$ indicates random spatial distribution.

# Calculate Spatial Variable Genes via Seurat / Moran's I
seurat_obj <- FindSpatiallyVariableFeatures(
    seurat_obj,
    assay = "SCT",
    selection.method = "moransi",
    features = VariableFeatures(seurat_obj)[1:1000],
    r.metric = 5
)

# Top Spatially Variable Features
top_svgs <- head(SpatiallyVariableFeatures(seurat_obj, selection.method = "moransi"), 6)
print(top_svgs)

# Visualize Key Spatial Gradients
SpatialFeaturePlot(seurat_obj, features = top_svgs[1:4], ncol = 2)

---
title: "Visium HD Spatial Transcriptomics Clinical Capstone"
author: "Bioinformatics Research Fellow"
date: "November 26, 2026"
format:
  html:
    toc: true
    toc-depth: 3
    theme: cosmo
    code-fold: show
    code-tools: true
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE)
set.seed(42)
library(renv)
library(Seurat)
library(ggplot2)
```

## Session Environment Capture
```{r session_info}
sessionInfo()
```