celliverse

celliverse 0.0.2

Initial CRAN release

This is the first CRAN release of celliverse, an R toolkit for clustering, marker discovery, cell-type annotation, and downstream analysis of single-cell RNA-sequencing data.

Core single-cell analysis

• Added clustoCell() for data-driven identification of major clusters and sub-clusters together with ranked positive and negative marker discovery.
• Added the MarkoCell workflow for marker discovery from pre-defined clusters and user-selected cell subsets.
• Added functionality for cluster comparison, marker inspection, label transfer, integration with Seurat objects, and visualization of clustering and marker results.

Cell-type annotation

• Added the CelliVerse MarkerDB, distributed with the package as markerDB, providing harmonized positive and negative cell-type marker information for human and mouse.
• Added typoClust() for annotation of ClustoCell/MarkoCell results using either the curated CelliVerse MarkerDB or LLM-assisted annotation.
• Added ceLLMarkup() for direct LLM-assisted annotation from marker panels, clustering results, or compatible marker tables.
• Added typoPrompt() for generating structured, model-independent cell-type annotation prompts that can be used with any preferred chatbot or LLM.
• Added saveTypoPrompt() for exporting TypoPrompt objects as plain-text or self-contained HTML documents.

CelliVerse Agent

• Added an optional LLM-powered CelliVerse Agent that provides a browser-based natural-language interface to CelliVerse workflows.
• Added install_celliverse_agent() and run_celliverse_agent() for Agent setup and launch.
• Added support for common single-cell input formats, including R objects, delimited matrices, Matrix Market/10x inputs, zipped 10x triplets, and HDF5 inputs when the required optional dependency is available.
• Added support for both cloud-based model providers and local model backends such as Ollama and LM Studio.

Documentation and reproducibility

• Added a comprehensive package vignette covering clustering, marker discovery, annotation, visualization, TypoPrompt, and the CelliVerse Agent.
• Added expanded README documentation, installation guidance, interactive examples, and links to the browser-based ClustoCell demonstration.
• Added links to the dedicated CelliVerse-Project reproducibility repository containing manuscript analysis scripts and prepared MarkerDB resources.
• Computationally intensive and external-service-dependent examples are not executed during package checks; reproducible precomputed results are used where appropriate in the vignette.