Thursday 28 January 2027 | 11:30 - 13:00 | Workshop Room

A practical introduction to using agentic AI approaches for single-cell genomics analysis.

Participants will explore how to define analytical questions, design an analytical path, automate analysis workflows, and incorporate human review throughout the process.

Through a live demonstration using agentic AI harnesses such as ClawBio, Biomni and Claude Science, attendees will see how agents can be used to analyse public single-cell datasets and produce interactive results. The workshop will also explore practical considerations around efficient token use, evaluation and verification of AI-generated analyses. 

Single-Cell

Learning outcomes

By the end of this workshop, participants will be able to:

  • Understand how agentic AI can be incorporated into single-cell genomics analysis workflows 

  • Define analytical questions and translate them into structured analytical paths 

  • Identify opportunities to automate aspects of their analysis using AI agents 

  • Understand the role of prompts, agents and skills within agentic AI workflows 

  • Explore how agentic AI can be used to analyse public biological datasets 

  • Incorporate human-in-the-loop review into AI-assisted analysis workflows 

  • Understand approaches for sharing interactive analysis results and documenting analytical steps 

  • Recognise the importance of evaluations and verification when using AI for biological analysis 

  • Consider how to use AI resources efficiently, including managing token use

Who should attend

This workshop is designed for:

  • Bioinformaticians and computational biologists working with single-cell genomics data 
     

  • Genomics researchers interested in applying agentic AI to their analytical workflows 
     

  • Data scientists working with biological datasets 
     

  • Researchers looking to automate aspects of their analysis while retaining human oversight 
     

  • Principal investigators and research leaders exploring the potential of agentic AI for genomics research 

Workshop outline

  • Session 1 — Designing Agentic AI Workflows for Single-Cell Analysis 
     
  • Session 2 — Live Demonstration: Agentic AI for Single-Cell Analysis
     
  • Session 3 — Verification, Evaluation & Responsible Use

 

Prerequisites: A basic understanding of single-cell genomics and computational analysis workflows is recommended. No prior experience with agentic AI is required.

Workshop Leaders

Loading
Speaker profile image for Georgia Whitton

Georgia Whitton

Senior Bioinformatician
The Francis Crick Institute
Speaker profile image for James Campbell

James Campbell

Bioinformatics Training Developer
The Francis Crick Institute

Session 1 — Designing Agentic AI Workflows for Single-Cell Analysis (30 min)

An introduction to using agentic AI to structure and automate analytical workflows, including: 

  • Defining your analytical questions 

  • Designing an analytical path 

  • Identifying opportunities to automate analysis 

  • Using human-in-the-loop review 

  • Sharing interactive results 

  • Summarising and documenting the analysis steps 

The session will focus on how researchers can think about the overall analytical workflow before introducing automation, rather than simply using AI to generate individual pieces of code or analysis. 

Session 2 — Live Demonstration: Agentic AI for Single-Cell Analysis (25 min)

A practical walkthrough of an agentic AI workflow using real public data, including: 

  • Setting up agentic AI harnesses 

  • Exploring tools such as ClawBio, Biomni and Claude Science 

  • Working with prompts, agents and skills 

  • Applying agents to the analysis of public datasets 

  • Reviewing the resulting analysis and outputs 

This session is designed to make agentic AI workflows concrete and demonstrate how they can be applied to real biological analysis rather than remaining theoretical. 

Session 3 — Verification, Evaluation & Responsible Use (30 min)

A practical discussion of the considerations researchers need to make when incorporating agentic AI into biological analysis workflows, including: 

  • Efficient use of tokens and computational resources 

  • Evaluating AI-generated analyses 

  • Verification of analytical outputs 

  • Human review of agent-generated workflows and results 

  • Understanding where researchers need to retain oversight 

The goal is to provide a realistic understanding of how researchers can use agentic AI effectively while maintaining confidence in the resulting analysis. 

**Please register your interest in being considered for a workshop place when completing your festival ticket registration.**