30-Second Command Line Analysis
Analyze your heart data with a single command:
heartmap your_heart_data.h5ad
- Automatically detect your data format
- Perform quality control
- Run cell type annotation
- Generate visualizations
- Save results to
./results/
This uses default settings optimized for typical heart datasets. For custom analysis, see the 2-minute Python example below.
2-Minute Python Analysis
For more control, use the Python API:
from heartmap import Config
from heartmap.pipelines import ComprehensivePipeline
# Quick analysis with defaults
config = Config.default()
pipeline = ComprehensivePipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
print("✓ Analysis complete! Check 'results/' directory.")
- Cell type clusters and annotations
- Quality control metrics and plots
- UMAP visualization
- Comprehensive analysis report
- Processed AnnData object saved for further analysis
Custom Output Directory
# CLI
heartmap your_data.h5ad --output-dir my_results/
# Python
results = pipeline.run('your_data.h5ad', 'my_results/')
Choose Analysis Type
Basic
Communication
Multi-Chamber
Comprehensive
Runtime: 5-10 minutes | Best for: Quick QC and cell typingheartmap your_data.h5ad --analysis-type annotation
from heartmap.pipelines import BasicPipeline
pipeline = BasicPipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
- Cell type clusters (Leiden clustering)
- QC metrics (mitochondrial genes, cell counts)
- UMAP plot colored by cluster
- Marker genes per cluster
Runtime: 10-15 minutes | Best for: Cell-cell interaction analysisheartmap your_data.h5ad --analysis-type communication
from heartmap.pipelines import AdvancedCommunicationPipeline
pipeline = AdvancedCommunicationPipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
- Ligand-receptor interaction networks
- Communication hub scores
- Pathway enrichment analysis
- Interactive network graphs
Runtime: 15-20 minutes | Best for: Chamber-specific analysisheartmap your_data.h5ad --analysis-type multi-chamber
from heartmap.pipelines import MultiChamberPipeline
pipeline = MultiChamberPipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
- Chamber-specific markers (RA, RV, LA, LV)
- Cross-chamber correlations
- Chamber composition analysis
- Chamber-specific visualization
Runtime: 20-30 minutes | Best for: Complete analysisheartmap your_data.h5ad --analysis-type comprehensive
from heartmap.pipelines import ComprehensivePipeline
pipeline = ComprehensivePipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
- All of the above
- Comprehensive dashboard
- Automated HTML report
- Publication-ready figures
Memory-Optimized Analysis
If you have limited RAM (8-16GB), use a custom config:
from heartmap import Config
# Create memory-optimized config
config = Config.default()
config.data.max_cells_subset = 10000 # Analyze 10K cells
config.data.max_genes_subset = 2000 # Top 2K variable genes
pipeline = ComprehensivePipeline(config)
results = pipeline.run('large_dataset.h5ad', 'results/')
Memory Guidelines:
- 8GB RAM:
max_cells_subset = 10000, max_genes_subset = 2000
- 16GB RAM:
max_cells_subset = 30000, max_genes_subset = 4000
- 32GB+ RAM:
max_cells_subset = 50000, max_genes_subset = 5000
Using Custom Config File
# Create your config
cat > my_config.yaml << EOF
data:
min_genes: 200
min_cells: 3
max_cells_subset: 30000
max_genes_subset: 4000
analysis:
resolution: 0.5
n_neighbors: 10
use_leiden: true
EOF
# Run with custom config
heartmap your_data.h5ad --config my_config.yaml
# Python
config = Config.from_yaml('my_config.yaml')
pipeline = ComprehensivePipeline(config)
results = pipeline.run('your_data.h5ad', 'results/')
Working with Results
After analysis, explore your results:
import scanpy as sc
# Load processed data
adata = sc.read_h5ad('results/heartmap_complete.h5ad')
# View cell clusters
print(adata.obs['leiden'].value_counts())
# Access hub scores (from communication analysis)
if 'hub_score' in adata.obs:
print(f"Average hub score: {adata.obs['hub_score'].mean():.4f}")
# Plot specific genes
sc.pl.umap(adata, color=['NPPA', 'MYH7', 'CD36'])
Expected Outputs
After running quick analysis, you’ll find:
results/
├── figures/
│ ├── umap_clusters.png # Cell type visualization
│ ├── qc_metrics.png # Quality control plots
│ └── marker_genes.png # Top markers per cluster
├── tables/
│ ├── cluster_summary.csv # Cluster statistics
│ └── marker_genes.csv # Marker gene list
├── heartmap_complete.h5ad # Processed AnnData
└── analysis_report.html # Interactive report
Troubleshooting
Memory Error: Cannot allocate array
Reduce dataset size:config.data.max_cells_subset = 5000
config.data.max_genes_subset = 1000
Use absolute paths:from pathlib import Path
data_path = Path('/full/path/to/your_data.h5ad')
results = pipeline.run(str(data_path), 'results/')
Enable test mode for quick validation:config.data.test_mode = True # Uses small subset
Next Steps
Python API
Explore comprehensive API examples
Custom Pipelines
Build your own analysis workflows
Configuration
Learn about all configuration options
Output Reference
Understand all output files