mLLMCelltype2.0.0 package

Cell Type Annotation Using Large Language Models

annotate_cell_types

Cell Type Annotation with Multi-LLM Framework

AnthropicProcessor

Anthropic API Processor

BaseAPIProcessor

Base API Processor Class

CacheManager

Cache Manager Class

calculate_simple_consensus

Calculate simple consensus without LLM

check_consensus

Check if consensus is reached among models

clean_annotation

Clean annotation text by removing prefixes and extra whitespace

combine_results

Combine results from all phases of consensus annotation

compare_model_predictions

Compare predictions from different models

configure_logger

Set global logger configuration

create_annotation_prompt

Create prompt for cell type annotation

create_consensus_check_prompt

Create prompt for checking consensus among model predictions

create_discussion_prompt

Create prompt for additional discussion rounds

create_initial_discussion_prompt

Create prompt for the initial round of discussion

create_standardization_prompt

Create prompt for standardizing cell type names

custom_providers

Custom model manager for mLLMCelltype

DeepSeekProcessor

DeepSeek API Processor

dot-onAttach

Package startup message

dot-onLoad

Package load message

dot-qwen_endpoint_cache

Qwen API Processor

execute_consensus_check

Execute consensus check with retry logic

extract_labeled_value

Extract numeric value from line containing a label

facilitate_cluster_discussion

Facilitate discussion for a controversial cluster

filter_valid_responses

Filter out error responses from model round responses

find_majority_prediction

Find majority prediction from response lines

GeminiProcessor

Gemini API Processor

get_api_key

Utility functions for API key management

get_initial_predictions

Get initial predictions from all models

get_logger

Get the global logger instance

get_model_response

Get response from a specific model

get_provider

Determine provider from model name

GrokProcessor

Grok API Processor

identify_controversial_clusters

Identify controversial clusters based on consensus analysis

initialize_logger

Reinitialize global logger with a specific directory

interactive_consensus_annotation

Interactive consensus building for cell type annotation

list_custom_models

Get list of registered custom models

list_custom_providers

Get list of registered custom providers

logging_functions

Convenience functions for logging

MinimaxProcessor

Minimax API Processor

mllmcelltype_cache_dir

Get mLLMCelltype cache location

mllmcelltype_clear_cache

Clear mLLMCelltype cache

mLLMCelltype-package

mLLMCelltype: Cell Type Annotation Using Large Language Models

normalize_annotation

Normalize annotation for comparison

normalize_cluster_gene_list

Prompt templates for mLLMCelltype

OpenAIProcessor

OpenAI API Processor

OpenRouterProcessor

OpenRouter API Processor

parse_consensus_response

Parse consensus response from model

parse_flexible_format

Parse flexible format consensus response

parse_standard_format

Parse standard 4-line consensus response format

parse_text_predictions

Parse text-format model predictions into a named list

prepare_models_list

Prepare list of models to try for consensus checking

print_consensus_summary

Print summary of consensus results

process_anthropic

Process request using Anthropic models

process_controversial_clusters

Process controversial clusters through discussion

process_custom

Process request using custom provider

process_deepseek

Process request using DeepSeek models

process_gemini

Process request using Gemini models

process_grok

Process request using Grok models

process_minimax

Process request using MiniMax models

process_openai

Process request using OpenAI models

process_openrouter

Process request using OpenRouter models

process_qwen

Process request using Qwen models

process_stepfun

Process request using StepFun models

process_zhipu

Process request using Zhipu models

register_custom_model

Register a custom model for a provider

register_custom_provider

Register a custom LLM provider

resolve_provider_base_url

URL Utilities for Base URL Resolution

select_best_prediction

Select the best prediction from consensus results

standardize_cell_type_names

Standardize cell type names using a language model

StepFunProcessor

StepFun API Processor

UnifiedLogger

Unified Logger for mLLMCelltype Package

ZhipuProcessor

Zhipu API Processor

Automated cell type annotation for single-cell RNA sequencing data using consensus predictions from multiple large language models. Integrates with Seurat objects and provides uncertainty quantification for annotations. Supports various LLM providers including OpenAI, Anthropic, and Google. For details see Yang et al. (2025) <doi:10.1101/2025.04.10.647852>.

  • Maintainer: Chen Yang
  • License: MIT + file LICENSE
  • Last published: 2026-02-08 10:50:09 UTC