Recent Releases of heartmap

heartmap -

Added PCA computation and neighborhood graph calculation to DataProcessor
Fixed KeyError: 'No neighbors in .uns' by ensuring sc.pp.neighbors() runs before clustering
Updated version to 1.1.2 for patch release
All dependencies (leidenalg, python-igraph) properly configured
CLI now works correctly with real heart data
Created standalone api_server.py for FastAPI deployment
Added requirements-api.txt with FastAPI dependencies
Updated Dockerfile for containerized API deployment
Enhanced docker-compose.yml with health checks and proper volumes
Added comprehensive API_DEPLOYMENT.md guide
API server tested and working on localhost:8001
Ready for production deployment on cloud platforms
Fixed line length E501 by breaking long function signature
Removed trailing whitespace W291
Fixed continuation line indentation E128
Removed blank line with whitespace W293
All flake8 checks now pass with --max-line-length=100

- HTML
Published by Tumo505 7 months ago

heartmap - v0.1.0-alpha pre-release

Release Notes - v0.1.0-alpha

🎉 Initial Alpha Release

This is the first alpha release of the GRAIL-Heart cell_comm module, a foundational component of the broader GRAIL-Heart project for analyzing cell-cell communication patterns in human heart tissue using single-cell RNA sequencing data.

✨ New Features

🔬 Core Analysis Pipeline

Complete end-to-end analysis pipeline with 7 modular scripts
Data preprocessing and quality control with automated filtering and normalization
Cell type annotation and clustering using Leiden algorithm
Cell-cell communication analysis with LIANA integration
Advanced communication analysis including pathway enrichment and temporal analysis
Comprehensive visualization and reporting with interactive plots

📊 Quality Control & Preprocessing

Automated cell and gene filtering based on quality metrics
Mitochondrial, ribosomal, and hemoglobin gene detection
Data normalization and log transformation
Highly variable gene identification for downstream analysis

�� Cell Type Analysis

Principal Component Analysis (PCA) for dimensionality reduction
Neighborhood graph computation using UMAP
Leiden clustering for cell type identification
Marker gene identification using Wilcoxon rank-sum test
UMAP visualization of cell clusters

💬 Cell-Cell Communication

Integration with LIANA framework for ligand-receptor analysis
Communication pattern visualization with heatmaps and networks
Pathway enrichment analysis for heart-specific pathways
Communication specificity analysis between cell types
Hub cell identification in communication networks

📈 Visualization & Reporting

Static plots (QC metrics, clustering results, pathway heatmaps)
Interactive UMAP plots using Plotly
Communication network visualizations using NetworkX
Automated report generation in Markdown format
Comprehensive analysis summaries

🛠️ Technical Features

🔧 Pipeline Management

Automated pipeline runner with error handling and progress tracking
Modular script architecture for easy customization
Comprehensive logging and error reporting
Data persistence at each pipeline stage

📦 Dependencies & Compatibility

Scanpy 1.9.0+ for single-cell analysis
LIANA 0.1.0+ for cell communication analysis
Plotly 5.0.0+ for interactive visualizations
NetworkX 2.8.0+ for network analysis
Python 3.9+ compatibility

��️ Data Management

H5AD format support for AnnData objects
Automated data organization with clear directory structure
Intermediate data saving for pipeline resumption
Git-friendly with appropriate .gitignore patterns

📋 Dataset Support

🫀 Human Heart Dataset

Source: Single Cell Portal (SCP498) - healthy human 4-chamber heart map
Size: ~287,269 cells, ~33,694 genes (pre-filtered)
Reference: He et al. (2020) Genome Biology
Format: H5AD (AnnData) format

🚀 Getting Started

Quick Start

```bash

Clone repository

git clone https://github.com/Tumo505/GRAIL-Heart-cell-cell-communication.git cd grail-heart

Setup environment

conda create -n cellcomm python=3.9 conda activate cellcomm pip install -r requirements.txt

Download data and run pipeline

python scripts/run_pipeline.py ```

Manual Execution

```bash

Individual pipeline steps

python scripts/01datapreprocessing.py python scripts/02qualitycontrol.py python scripts/03cellannotation.py python scripts/04communicationanalysis.py python scripts/05_visualization.py ```

⚠️ Known Issues & Limitations

🔴 Alpha Release Limitations

Large dataset processing may require significant computational resources
Memory requirements can be high for full dataset analysis
Dependency conflicts may occur with certain Python versions
LIANA integration requires additional setup for some systems

🟡 Performance Considerations

Highly variable genes detection can be slow on large datasets
Clustering algorithms may require tuning for optimal results
Visualization generation may take time for large datasets

🔮 Future Roadmap

Planned Features for v0.2.0

Spatial transcriptomics integration
Multi-omics data support
Deep learning model integration
Cloud deployment support
Enhanced visualization options

Research Integration

GNN model preparation for spatial modeling
RNN framework integration for temporal analysis
Hybrid GNN-RNN architecture development

📄 Documentation

README.md: Comprehensive project overview and setup instructions
CITATION.cff: Proper citation information for academic use
NOTICE: Attribution and licensing information
Apache 2.0 License: Open source licensing

�� Contributing

This is an academic research project. For questions, contributions, or collaboration: - Contact: Tumo Kgabeng - Institution: UNISA Biomedical Engineering Research Group - Supervisors: Prof. Thanyani Pandelani, Prof. Lulu Wang, Prof. Harry Ngwangwa

📊 Citation

If you use this work in your research, please cite: Kgabeng, T., et al. (2024). GRAIL-Heart: Graph-based Reconstruction of Artificial Intercellular Links (cell_comm module). GitHub: https://github.com/Tumo505/GRAIL-Heart-cell-cell-communication

⚠️ Research Preview: This is an alpha release for ongoing academic research. The code, models, and results are under active development and may change. Use with caution in production environments.