Package: CellODE 1.0.0

Introduction | Mathematical Framework | Problem Formulation | Model Architecture | Encoder Network | Time Inference | Latent Space Inference | Neural ODE | Continuous Dynamics | Integration | ODE Function Architecture | Decoder Network | Reconstruction Likelihood | 1. Negative Binomial (NB) | 2. Zero-Inflated Negative Binomial (ZINB) | 3. Mean Squared Error (MSE) | Loss Function | Components | Time Direction Determination | Demonstration | Key Innovations | 1. Automatic Time Inference | 2. Continuous Dynamics | 3. Unified Framework | References | Session Info

Introduction | Hyperparameter Tuning | Latent Space Dimensionality | ODE Network Architecture | VAE Network Architecture | Loss Function Selection | Loss Weight Balancing | Training Configuration | Learning Rate Scheduling | Batch Size and Epochs | Data Subsampling | Handling Complex Trajectories | Branching Trajectories | Cyclic Trajectories | Prediction on Query Data | Coarse vs Fine Mode | Predicting Future States | Integration with Seurat Workflows | Adding Results to Seurat Object | Downstream Analysis | Performance Optimization | GPU Acceleration | Memory Management | Batched Inference | Model Persistence | Saving and Loading | Model Inspection | Troubleshooting | Training Issues | Quality Issues | Memory Issues | Reproducibility | Best Practices Summary | Session Info

Introduction | Installation | From R-universe (Recommended) | From GitHub | Basic Workflow | 1. Load Libraries | 2. Prepare Data | 3. Create and Train Model | 4. Extract Pseudotime | 5. Get Latent Space | 6. Compute Vector Field | 7. Save Model | Quick Example with Synthetic Data | Parameter Guidelines | Loss Mode Selection | Latent Dimensions | Next Steps | Session Info

Introduction | Quick Start: Using CellODE Functions | Create Example Data | Simulate CellODE Model Output | Vector Field Computation with CellODE | Step 1: Compute Cosine Similarity | Step 2: Compute Embedding Velocity | Step 3: Grid-Based Vector Field | Visualization with CellODE | Plot Pseudotime with Direction Arrows | Plot Vector Field - Grid Mode | Plot Vector Field - Stream Mode | Plot Vector Field - Raw Mode | Customization Options | Arrow Styling | Clean Publication Figure | Mathematical Background | From Latent Dynamics to Velocity | Cosine Similarity | Transition Probability | Embedding Velocity | Comparison with RNA Velocity | Best Practices | 1. Neighbor Selection | 2. Visualization Density | 3. Self-Transition | Summary | Session Info