Welcome to Cell2Fate’s documentation!

For installation instructions please see:

• Installation

For reporting bugs or other issues with Cell2Fate please use GitHub Issues.

The standard recommended workflow for using Cell2Fate can be found in the quick start tutorial.

Quick start tutorial:

• Cell2Fate analysis of mouse pancreas development dataset
  • Loading packages
  • Loading single cell RNA sequencing data
  • Initializing and training the model
  • Exporting and visualizing posteriors
  • Visualizing summary statistics and module activations
  • Visualizing velocities and module specific velocities
  • Plotting technical variables and other downstream analyses
• Assessing confidence in cell2fate predictions

Cell2Fate + Cell2Location Workflow

Cell2Fate RNA velocity modules can also be mapped spatially when used together with Cell2Location. To use this workflow, follow these steps:

1. Run Cell2Fate and save modules, as shown in here.

2. Install Cell2Location by following the instructions in the Cell2Location GitHub repository.

3. Run Cell2Location, with the Cell2Fate modules as input, as shown in here.

Cell2Fate + Cell2Location:

• Running cell2fate and saving modules in human developing brain dataset
• Running cell2location with the saved modules from the human developing brain dataset

See other notebooks to reproduce publication figures.

Reproduce publication figures:

• Cell2Fate analysis of mouse bone marrow dataset
• Cell2Fate dynamical model analysis of PBMC dataset
• Cell2Fate analysis of human bone marrow dataset
• Cell2Fate analysis of mouse erythroid dataset
• Cell2Fate analysis of mouse dentate gyrus dataset

API:

• Prepare data
  • Prepare anndata for training the model
    • utils.get_training_data()
• Cell2Fate: RNA velocity (scvi-tools/pyro)
  • Cell2Fate RNA velocity modules for prediction of cell fates model class (scvi-tools BaseModelClass)
    • Cell2fate_DynamicalModel
      • Cell2fate_DynamicalModel.setup_anndata()
      • Cell2fate_DynamicalModel.train()
      • Cell2fate_DynamicalModel.plot_history()
      • Cell2fate_DynamicalModel.compute_module_summary_statistics()
      • Cell2fate_DynamicalModel.export_posterior()
      • Cell2fate_DynamicalModel.export_posterior_quantiles()
      • Cell2fate_DynamicalModel.view_history()
      • Cell2fate_DynamicalModel.get_module_top_features()
  • Pyro and scvi-tools Module classes
    • Pyro Module class (defining the model using pyro, math description)
      • Cell2fate_DynamicalModel_module
        • Cell2fate_DynamicalModel_module.create_plates()
        • Cell2fate_DynamicalModel_module.list_obs_plate_vars()
        • Cell2fate_DynamicalModel_module.forward()
        • Cell2fate_DynamicalModel_module.training
• Plotting
  • Methods to compute and visualize velocity graphs and trajectories
    • Cell2fate_DynamicalModel.compute_velocity_graph_Bergen2020()
    • Cell2fate_DynamicalModel.compute_and_plot_module_velocity()
    • Cell2fate_DynamicalModel.compute_and_plot_total_velocity()
    • Cell2fate_DynamicalModel.visualize_module_trajectories()
    • Cell2fate_DynamicalModel.compute_and_plot_total_velocity_scvelo()
    • utils.compute_velocity_graph_Bergen2020()
    • utils.plot_velocity_umap_Bergen2020()
  • Methods to plot module summary statistics and activations
    • Cell2fate_DynamicalModel.compare_module_activation()
    • Cell2fate_DynamicalModel.plot_module_summary_statistics()
  • Other methods to plot top features, genes, and technical variables
    • Cell2fate_DynamicalModel.plot_top_features()
    • Cell2fate_DynamicalModel.plot_genes()
    • Cell2fate_DynamicalModel.plot_technical_variables()
• General utils
  • Calculate max number of modules with leiden clustering
    • utils.get_max_modules()
  • Robust optimization of cell2fate model
    • utils.robust_optimization()
  • Other methods for cell2fate model definition
    • utils.G_a()
    • utils.G_b()
    • utils.mu_alpha()
    • utils.mu_mRNA_continuousAlpha()
    • utils.mu_mRNA_continousAlpha_globalTime_twoStates()
• Pyro and scvi-tools infrastructure classes
  • Base mixin classes (AutoGuide setup, posterior quantile computation, plotting & export)
    • init_to_value()
    • expand_zeros_along_dim()
    • complete_tensor_along_dim()
    • AutoGuideMixinModule
    • MyAutoHierarchicalNormalMessenger
      • MyAutoHierarchicalNormalMessenger.call_new()
      • MyAutoHierarchicalNormalMessenger.get_posterior_quantile()
      • MyAutoHierarchicalNormalMessenger.quantiles()
      • MyAutoHierarchicalNormalMessenger.training
    • PyroTrainingPlan_ClippedAdamDecayingRate
      • PyroTrainingPlan_ClippedAdamDecayingRate.training
    • QuantileMixin
      • QuantileMixin.posterior_quantile()
    • PyroAggressiveTrainingPlan1
      • PyroAggressiveTrainingPlan1.training
      • PyroAggressiveTrainingPlan1.change_requires_grad()
      • PyroAggressiveTrainingPlan1.training_epoch_end()
      • PyroAggressiveTrainingPlan1.training_step()
  • Scvi-tools Module classes (initialising the model and the guide, PyroBaseModuleClass)
    • Cell2FateBaseModule
      • Cell2FateBaseModule.model
      • Cell2FateBaseModule.guide
      • Cell2FateBaseModule.is_amortised
      • Cell2FateBaseModule.list_obs_plate_vars
      • Cell2FateBaseModule.init_to_value()
      • Cell2FateBaseModule.training
  • Auto amortised hierarchical normal messenger for amortized inference
    • FCLayersPyro
      • FCLayersPyro.training
    • AutoAmortisedHierarchicalNormalMessenger
      • AutoAmortisedHierarchicalNormalMessenger.weight_type
      • AutoAmortisedHierarchicalNormalMessenger.get_posterior()
      • AutoAmortisedHierarchicalNormalMessenger.encode()
      • AutoAmortisedHierarchicalNormalMessenger.median()
      • AutoAmortisedHierarchicalNormalMessenger.quantiles()
      • AutoAmortisedHierarchicalNormalMessenger.training