Package index • bHIVE

Core Functions

The primary interface for running bHIVE algorithms. Use bHIVE() for single-pass analysis, honeycombHIVE() for multilayer hierarchical refinement, and refineB() for gradient-based post-processing.

bHIVE(): bHIVE: B-cell Hybrid Immune Variant Engine
honeycombHIVE(): honeycombHIVE: Multilayer AIS with optional gradient-based fine-tuning
refineB(): refineB: Gradient-based fine-tuning for bHIVE antibodies with multiple loss functions and optimizers
visualizeHIVE(): Visualize bHIVE/honeycombHIVE Results

Bioconductor Integration

Run bHIVE through bluster::clusterRows() and bluster::clusterSweep(). bHIVE() also accepts SummarizedExperiment and SingleCellExperiment objects directly.

bHIVEParam() .defaultScalarArguments(<bHIVEParam>) show(<bHIVEParam>) clusterRows(<ANY>,<bHIVEParam>): bHIVE clustering through bluster

Hyperparameter Tuning

Grid search over bHIVE hyperparameters with BiocParallel support.

swarmbHIVE(): Tune Hyperparameters for bHIVE (Swarm/Grid Search)

R6 Algorithm Classes

Object-oriented interface for composing immune algorithms with injectable modules. AINet is the primary class; ImmuneAlgorithm is the abstract base.

AINet: AINet
ImmuneAlgorithm: ImmuneAlgorithm
ImmuneRepertoire: ImmuneRepertoire

Mutation & Selection Modules

Modules controlling how antibodies mutate, compete, and are selected. Inject these into AINet$new() to customize algorithm behavior.

SHMEngine: SHMEngine
GerminalCenter: GerminalCenter
VDJLibrary: VDJLibrary

Network Regulation Modules

Modules for repertoire regulation and activation control.

IdiotypicNetwork: IdiotypicNetwork
ActivationGate: ActivationGate

Adaptation & Memory Modules

Modules for environment-driven adaptation, memory, and ensemble methods.

Microenvironment: Microenvironment
MemoryPool: MemoryPool
ClassSwitcher: ClassSwitcher
ConvergentSelector: ConvergentSelector

caret Integration

Model objects for use with the caret package’s train() function.

bHIVEmodel: bHIVE classification model for caret
honeycombHIVEmodel: Multilayered honeycombHIVE classification model for caret