ML-DS: Myeloid Leukemia of Down Syndrome Analysis

This repository contains computational analysis code for single-cell RNA sequencing (scRNA-seq) data processing and analysis of Myeloid Leukemia of Down Syndrome (ML-DS). The code supports data processing, analyses, and figure generation for the associated publication.

Overview

This project focuses on understanding the molecular mechanisms underlying ML-DS progression by analyzing:

• Transient Abnormal Myelopoiesis (TAM) - a pre-leukemic condition in Down syndrome newborns
• ML-DS - the acute leukemia that can develop from TAM
• Fetal liver development - normal hematopoietic development in diploid and trisomy 21 (Down syndrome) contexts
• Megakaryocyte lineage trajectories - developmental pathways from hematopoietic stem cells to mature megakaryocytes

Key Analyses

1. Data Processing (main_analyses/01_fetalLivers_processing/, main_analyses/02_MLDS_processing/, main_analyses/03_otherLeukaemia_processing/)

• Quality control and preprocessing of scRNA-seq data
• Cell type annotation using reference-based label transfer
• Sample metadata integration and batch effect correction
• Processing of fetal liver references, ML-DS samples, and other leukemia datasets

2. Fetal Aneuploidy Analysis (main_analyses/01.2_fetal_aneuploidy_analyses/)

• Analysis of trisomy 21 effects on fetal hematopoietic development
• Comparison of diploid vs. trisomy 21 developmental trajectories
• Early molecular changes in Down syndrome hematopoiesis

3. Transcriptional Module Derivation (main_analyses/04_derive_transcriptional_modules/)

• GATA1s module analysis (x4_GATA1s_module.R) - characterization of truncated GATA1 regulatory network
• ML-DS imprint in T21 fetal liver (x3_MLDS_imprint_in_fLiver_T21_v2.R) - early molecular changes predisposing to leukemia
• TAM vs ML-DS comparison (x5_good_vs_bad_TAM.R, x6_TAM_vs_MLDS.R) - progression markers

4. Module Specificity Testing (main_analyses/05_transcriptional_modules_specificity/)

• Cross-dataset validation of transcriptional modules
• Module scoring across different cell types and conditions
• Bulk RNA-seq validation of single-cell derived signatures

5. ML-DS Relapse/Refactory Analysis (main_analyses/06_MLDS_refractory_relapse/)

• Analysis of treatment response and relapse patterns
• Identification of markers associated with clinical outcomes

6. Trajectory Analysis (eda_scripts/)

• Palantir-based pseudotime analysis (2.1_fLiver_trajectory_2401.R) of fetal liver hematopoietic development
• Construction of megakaryocyte/erythroid/mast cell developmental trajectories (MLDS_trajectory_projection/)
• Projection of TAM/ML-DS cells onto normal developmental trajectories

Repository Structure

├── main_analyses/                              # Primary analysis pipeline
│   ├── 01_fetalLivers_processing/             # Fetal liver reference processing
│   ├── 01.2_fetal_aneuploidy_analyses/        # Trisomy 21 effects analysis
│   ├── 02_MLDS_processing/                    # ML-DS dataset processing
│   ├── 03_otherLeukaemia_processing/          # Other leukemia datasets
│   ├── 04_derive_transcriptional_modules/     # Module identification
│   ├── 05_transcriptional_modules_specificity/ # Module validation
│   ├── 06_MLDS_refractory_relapse/            # Clinical outcomes
│   ├── figures_generation/                    # Publication figures
│   └── utils/                                 # Shared utility functions
├── eda_scripts/                               # Exploratory analyses
│   ├── MLDS_trajectory_projection/            # Trajectory analysis
│   └── *.R                                    # Various exploratory scripts
└── LICENSE                                    # License file

Key Dependencies

R Packages

• Seurat - scRNA-seq analysis framework
• DESeq2 / edgeR - differential expression analysis
• UCell - gene module scoring
• ComplexHeatmap - advanced visualization
• tidyverse - data manipulation and visualization

Python Packages

• Palantir - trajectory inference and pseudotime analysis
• scanpy - single-cell analysis in Python
• pandas / numpy - data manipulation

External Tools

• CellTypist - automated cell type annotation
• SoupX - ambient RNA removal
• Harmony - batch effect correction

Data Types

The analysis incorporates multiple datasets:

• ML-DS patient samples - diagnostic and follow-up timepoints
• TAM patient samples - transient abnormal myelopoiesis
• Fetal liver references - normal hematopoietic development (diploid and T21)
• Fetal adrenal - additional developmental reference
• Infant ALL - pediatric leukemia comparison
• Published atlases - external validation datasets

Usage

Prerequisites

1. Install required R and Python packages
2. Configure file paths in scripts to match your data organization
3. Ensure access to reference genomes and annotation files

Basic Workflow

1. Data Processing: Run scripts in main_analyses/01_fetalLivers_processing/, main_analyses/02_MLDS_processing/, and main_analyses/03_otherLeukaemia_processing/
2. Aneuploidy Analysis: Execute scripts in main_analyses/01.2_fetal_aneuploidy_analyses/
3. Transcriptional Modules: Run module derivation scripts in main_analyses/04_derive_transcriptional_modules/
4. Module Validation: Execute validation scripts in main_analyses/05_transcriptional_modules_specificity/
5. Clinical Analysis: Run outcome analysis scripts in main_analyses/06_MLDS_refractory_relapse/
6. Trajectory Analysis: Run exploratory trajectory analysis in eda_scripts/MLDS_trajectory_projection/
7. Visualization: Generate publication figures using main_analyses/figures_generation/

Citation

If you use this code, please cite the associated publication:

[Publication details to be added upon acceptance]

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

For questions about the analysis or code implementation, please contact the authors through the associated publication or repository issues.