Integrative Multi-Omic Data Analysis Practicals

This repository contains the practical materials for the course Integrative Multi-Omic Data Analysis for Biological and Biomedical Data.

The course introduces practical workflows for preprocessing, quality control, exploratory analysis, and integration of multi-omic datasets. Days 1 and 2 focus on bulk multi-omics integration using a Chronic Lymphocytic Leukemia (CLL) cohort and Multi-Omics Factor Analysis (MOFA). Day 3 focuses on single-cell multi-omics, and Day 4 introduces MOFAcell / knowledge-based interpretation of multi-cellular data.

---

Repository overview

.
├── day1/
│   ├── Day1_single_omics_practical.qmd
│   ├── RNAseq_count.tsv
│   ├── Methylation_Mvalues.tsv
│   ├── Genetics_mutationStatus.tsv
│   ├── drugScreen.tsv
│   └── metadata.tsv
│
├── day2/
│   ├── Day2_MOFA_practical.qmd
│   └── reactomeGS.rda
│
├── day3/
│   ├── 00_prep_python_uv/
│   ├── 01_single_cell_processing/
│   ├── 02_batch_integration/
│   ├── 03_cite_seq/
│   ├── 04_cite_seq_application/
│   ├── 05_sc_atac_seq/
│   ├── conda_envs/
│   ├── Dockerfile
│   └── XX_exam_data/
│
└── final_assignment/
    ├── AML_MOFA_exam_student_assignment.html
    ├── rna_count.tsv
    ├── gene_mutation.tsv
    ├── drug_auc.tsv
    └── metadata.tsv

---

Quick start

For Days 1 and 2, please install the following before the course:

1. R version 4.3 or newer
2. RStudio Desktop
3. Quarto
4. Required R and Bioconductor packages listed below

The Day 1 and Day 2 practicals are written as Quarto (.qmd) documents and are intended to be run in RStudio.

---

Installing software

R

Download and install R from:

https://cran.r-project.org/

Please use R 4.3 or newer.

RStudio Desktop

Download and install RStudio Desktop from:

https://posit.co/download/rstudio-desktop/

If you are not familiar with Quarto, we recommend that you read: https://quarto.org/. We will also give a brief introduction during the first practical session.

---

Installing R packages for Days 1 and 2

Open RStudio and run the following commands in the R console.

# Install BiocManager if needed
if (!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}

# CRAN packages
install.packages(c(
  "tidyverse",
  "data.table",
  "matrixStats",
  "knitr",
  "ggpubr",
  "UpSetR",
  "psych",
  "survival",
  "survminer",
  "pheatmap",
  "RColorBrewer",
  "patchwork"
))

# Bioconductor packages
BiocManager::install(c(
  "DESeq2",
  "SummarizedExperiment",
  "MOFA2"
))

MOFA2 backend

MOFA2 requires a Python backend called mofapy2. In many cases, this can be automatically installed when the MOFA2 R package is installed.

If this fails, please check the official MOFA2 installation guide:

https://biofam.github.io/MOFA2/installation.html

---

Test your setup before the course

Please run the following in RStudio:

library(tidyverse)
library(DESeq2)
library(MOFA2)
library(UpSetR)

sessionInfo()

If all packages load without errors, your R environment is ready for Days 1 and 2.

You can also open day1/Day1_single_omics_practical.qmd in RStudio and render it with the command:

quarto::quarto_render("day1/Day1_single_omics_practical.qmd")

Alternatively, click Render in RStudio.

---

Day 1: single-omics preprocessing and exploratory analysis

Folder: day1/

Main file:

day1/Day1_single_omics_practical.qmd

Day 1 covers:

• loading multi-omic data tables and sample metadata
• checking sample overlap across omics layers
• RNA-seq preprocessing with DESeq2
• methylation data quality control
• mutation matrix exploration
• drug response data preprocessing
• exploratory PCA for individual omics layers
• comparison of single-omics structure before integration

---

Day 2: MOFA integration

Folder: day2/

Main file:

day2/Day2_MOFA_practical.qmd

Day 2 builds on Day 1 and covers:

• preparing integration-ready matrices
• creating and training a MOFA model
• inspecting variance explained by latent factors
• associating factors with metadata
• interpreting feature weights
• pathway-level interpretation of MOFA factors
• linking latent factors to biological and clinical hypotheses

---

Day 3: single-cell multi-omics

Folder: day3/

Day 3 uses Python-based single-cell analysis workflows. Because the Python environment is more complex, we recommend using the provided Docker image (will be provided prior to the day 3 practical session) for this part of the course.

If you prefer a local setup, the Day 3 modules use uv for reproducible Python environments. Each module contains its own pyproject.toml and uv.lock file.

To set up one module locally:

cd day3/01_single_cell_processing
uv sync
uv run jupyter lab

or run Python scripts with:

uv run python main.py

Some Day 3 modules require additional data. If you are not using Docker, download the data from:

https://heibox.uni-heidelberg.de/d/dcb5d295521b47108715/ (data will be uploaded later)

For the scATAC-seq module using snapATAC2, a Rust compiler may be required if no precompiled binary is available for your system:

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

---

Day 4: MOFAcell / knowledge-based interpretation

Day 4 focuses on MOFAcell and multi-cellular factor interpretation.

We will mainly use Google Colab to keep the setup lightweight. As a backup, a Conda environment will be provided. If you do not already have Conda, we recommend installing Miniconda:

https://www.anaconda.com/docs/getting-started/miniconda/install/overview

You can find everything we need for the day here: https://drive.google.com/drive/folders/1EI9_uoCQ7AIdq8X-c3r9RPDSApxY0NRY?usp=sharing

---

Final assignment

Folder: final_assignment/

The student-facing assignment is provided as:

final_assignment/AML_MOFA_exam_student_assignment.html

The main assignment uses a synthetic AML multi-omic dataset and asks you to apply concepts from the practicals, including preprocessing, exploratory analysis, and MOFA-based integration.

Please follow the submission instructions given during the course.

An optional/bonus assignment of analysing a single-cell COVID-19 CITE-seq dataset is also provided in the day3/XX_exam_data folder and detailed will be introduced in the Day 3 practical session.

---

Recommended preparation before the course

Before the first practical, please try to:

1. Install R, RStudio, and Quarto.
2. Install the required R packages.
3. Clone or download this repository.
4. Open day1/Day1_single_omics_practical.qmd in RStudio.
5. Check that the setup test above runs without errors.

Do not worry if everything is not perfect. We will reserve some time during the practical sessions for setup and troubleshooting.