Investigating Multi-Omic Mapping Methods to Uncover Chromatin Variability at the Single-Cell Level

Bachelor’s Thesis – Luis Raphael Reich
Technical University of Munich & LMU Munich

Overview

This project investigates spatial variability in chromatin accessibility at the single-cell level by integrating scRNA-seq, scATAC-seq, and spatial transcriptomics data. The work combines state-of-the-art methods from computational biology and spatial omics to identify putative cis-regulatory interactions and uncover spatially structured gene regulation.

Methods

• Datasets

  • Spatial transcriptomics (MERFISH, mouse primary motor cortex)
  • Joint scRNA-seq + scATAC-seq (SHARE-seq, whole mouse brain)

• Core tools

  • Tangram: Spatial mapping of single-cell data
  • Descart: Detection of spatially variable peaks
  • scverse stack: scanpy, squidpy, muon for data processing

• Analyses

  • Mapping chromatin accessibility onto spatial coordinates
  • Correlation analysis of peak accessibility with spatial gene expression
  • cCRE annotation overlap and novel peak discovery
  • Motif enrichment analysis (JASPAR)

Key Findings

• Tangram preserves known spatial cell-type distributions when mapping SHARE-seq data.
• Peaks correlated with niche-specific marker genes are enriched in those spatial niches.
• Many such peaks overlap with canonical cCREs, and some may represent novel elements.
• SPI1 motif-enriched peaks show strong microglial enrichment, reflecting known biology.

Thesis

A detailed explanation of the methods, analyses, and findings is provided in the thesis manuscript (work in progress):

Thesis PDF

Project Structure

├── README.md                  # Project overview
├── results/                   # Output results and figures (not yet uploaded)
├── src/                       # Source code and notebooks
│   ├── niche_analysis.ipynb   # Spatial niche detection with NichePCA (https://github.com/imsb-uke/nichepca)
│   ├── running_tg.ipynb       # Tangram spatial mapping
│   ├── preprocessing/         # Preprocessing scripts
│   │   ├── SHARE-seq_pp.ipynb
│   │   └── spatial_pp.py
│   └── Descart/               # Descart-based peak-gene correlation
│       ├── Gene-peak_interaction_detection.ipynb
│       ├── correlation_analysis.ipynb
│       ├── descart.py
│       ├── peak_gene_utils.py