BMI206-Statistical-Methods-Group-Project

Project Overview

This repository contains the code, data, and results for our BMI206 Statistical Methods Group Project at UCSF.
Our group is reanalyzing data from the paper:

Tissue-specific enhancer–gene maps from multimodal single-cell data identify causal disease alleles
Saori Sakaue et al., Nature Genetics (2023)

This study integrates multimodal single-cell data to build tissue-specific enhancer–gene maps and link them to causal alleles underlying complex diseases. Our goal is to replicate and extend one of the key analyses from the paper using the statistical and computational tools learned in BMI206.

---

Set-Up Python Environment

python3.11 -m venv BMI206-Env

source BMI206-Env/bin/activate

pip install --upgrade pip

# data analysis, visualization, and notebook tools:
pip install numpy pandas matplotlib seaborn scipy scikit-learn
pip install scanpy statsmodels anndata

pip install jupyter ipykernel

# Add this environment to Jupyter
python -m ipykernel install --user --name=BMI206-Env --display-name "Python (BMI206-Env)"

Push to the repo

git add .


git commit -m "Add project README"


git push origin main

Objectives PLS add objectives from Assignment #2

• Reproduce a subset of analyses from the Sakaue et al. paper.

• Explore enhancer–gene link inference across tissues using statistical models.

• Document challenges, methodological choices, and key lessons learned.

---

Background

Enhancer–gene maps are crucial for understanding how noncoding genetic variation influences gene regulation and disease.
Sakaue et al. combined chromatin accessibility, gene expression, and GWAS summary statistics to identify putative causal enhancer–gene relationships across tissues.

---

Methods

We will use a combination of:

• Single-cell multiomics datasets from the original publication (downloaded from GEO)
• Statistical modeling techniques covered in BMI206:
  • ?
  • ?
  • ?
• Tools: Python (pandas, scanpy, statsmodels) and R (tidyverse, ggplot2, Seurat)

---

Deliverables

• Code notebooks: Data preprocessing, statistical reanalysis, visualization
• Slides: Summarizing approach, key findings, and challenges
• Figures: Replicated or reinterpreted plots based on the Sakaue et al. dataset
• README: Documentation of workflow and analysis plan

---

Presentation

• Due: November 24–26, 2025

• Evaluation: Based on approach, interpretation, and participation during Q&A

---

Team Members

• Juan Carlos Gomez
• Sara
• Izabella
• Joseph

---

Repository Structure