AI Agents in Transcriptomics

AI agents represent the frontier of transcriptomics analysis, integrating large language models with specialized computational pipelines to automate complex, multi-step analytical workflows. These systems can understand natural language instructions from users, decompose analysis tasks into actionable steps, execute specialized tools, and iterate based on feedback—without requiring users to manually chain individual methods. Our review identifies 5 emerging AI agent systems designed for transcriptomics applications.

Table D: AI Agents in Transcriptomics

Column Descriptions: Tool - Agent system name | Computational Requirements - Resource needs for operation | Usage - Deployment mode (Local/API/Web) | Online Service - Web interface availability | Features - Key capabilities and characteristics | Code - GitHub repository link

Tool Computational Requirements Usage Online Service Features Code
STAgent Medium Local deployment/API Not provided Enables human-AI collaboration and supports full spatial biology workflows, including experimental design, multimodal data analysis, and hypothesis generation. GitHub
SpatialAgent Usage under construction Usage under construction Not provided Supports human-AI collaboration and is capable of handling the entire spatial biology research workflow, from experimental design and multimodal data analysis to hypothesis generation. GitHub
CellAgent Low Web mode Provided Employs multi-agent collaborative decision-making to simulate a "deep reasoning" process, enabling task decomposition, execution, and optimization in a closed-loop manner. It also incorporates the sc-Omni toolkit for efficient tool integration. Not public
CompBioAgent Medium Local deployment/API Provided Fully operated through natural language with zero programming requirements. Integrates tools such as Cellxgene VIP and CellDepot for querying and visualizing various diseases and cell types. GitHub
AutoBA Low (API)/ High (Local Deploy) Local deployment/API Not provided Applicable to both spatial transcriptomics and multi-omics; highly automated, user-friendly, and compatible with emerging bioinformatics tools. GitHub

Key Insights: AI agents represent an emerging paradigm in transcriptomics, with 5 systems currently available offering diverse accessibility models. Code Accessibility: 80% (4/5) provide public GitHub repositories (STAgent, SpatialAgent, CompBioAgent, AutoBA), though CellAgent remains proprietary. Deployment Models: 40% (2/5) offer web-based interfaces (CellAgent, CompBioAgent) enabling zero-setup access, while 60% (3/5) support local deployment for data privacy and customization. Computational Requirements: Range from low (CellAgent web mode) to high (AutoBA local deployment), with most supporting flexible API-based access. Spatial Biology Focus: 60% (3/5) explicitly target spatial transcriptomics workflows (STAgent, SpatialAgent, AutoBA), reflecting the field's growing emphasis on spatial context. Human-AI Collaboration: Multiple agents (STAgent, SpatialAgent, CellAgent) emphasize collaborative decision-making rather than full automation, enabling researchers to guide and refine analyses iteratively. Note that SpatialAgent is under construction, and accessibility details may evolve as the field matures.

📊 AI Agent Landscape Analysis

Typical AI Agent Workflow

1. User Input

Natural language analysis request

→
2. LLM Processing

Parse task requirements

→
3. Tool Selection

Choose methods/models

→
4. Execution

Run analysis pipeline

→
5. Results & Insights

Visualizations & reports