Single-molecule analysis of natural and synthetic base modifications via Nanalogue — Talk to your genomics data in plain English!
Sathish Thiyagarajan1, Conrad A. Nieduszynski1,2
1 Earlham Institute, Norwich, UK    2 University of East Anglia, Norwich, UK
✉ sathish.thiyagarajan@earlham.ac.uk
📦 QR code to nanalogue-gui github.com/sathish-t/nanalogue-gui Start using today!
AI chat interface to your sequencing
Introducing nanalogue-gui
Sequencing data processing application with AI chat features
Transparency 🔍
View system prompt, inspect tools in plain English
Available tools 🧰
Analyze data in plain text and BAM, plot data, …
Customization 🔧
Tune according to LLM’s power and task’s needs
Powerful 💪
Can do simple exploring or complex analyses
Code is optional 👌
You don’t have to read or approve code (unless you want to)
Secure Sandbox behind the scenes 🔒
Code runs locally (Pydantic’s Monty1 & Vercel’s just-bash2)
Nanalogue GUI chat screenshot
😊 Peace of mind
AI cannot escape allowed directory, cannot delete/overwrite
Use any LLM
OpenAI, Claude, Gemini or a local LLM
Use Oxford Nanopore’s GPUs
Run local LLM for nanalogue-chat when not sequencing
🔐 Security and Privacy
Use local LLMs (e.g. Qwen), no egress from your premises
💻 Multi-platform and easy to install
macOS, Linux, Windows (with Windows Subsystem for Linux)
📋 Reproducibility
/dump_llm_instructions saves sessions, /exec saved session
⬅ Actual local, private chat snapshot
From local LLM on GPU desktop of PromethION P2 solo

Powerful: Plain-English prompt —> published RNA-seq result

Just describe the analysis in plain English. nanalogue-chat can turn that request into a real multi-step workflow: count spanning reads, filter files, compute per-position coverage, and generate a publication-style plot comparable to a scripted notebook analysis.

RNA-seq SG-NEx chat example screenshot

Execution results and comparison of isoform coverage across gene

Coverage plot generated by nanalogue-chat
nanalogue-chat output on HCT116 cells
Coverage plot generated directly from the chat workflow.
SG-NEx Jupyter notebook example output
Published notebook workflow on A549 cells
Comparable style of analysis from the SG-NEx IsoTools tutorial.3,4

Plain-English request enough to drive a real genomics analysis and get a published result.3,4

Other creature comforts: QC your data

Check the essentials at a glance: read-length distributions, insertion/deletion sizes, basecalling quality, and single-molecule sequences with their modification patterns.

QC Results — Sequences

Find out more (nanalogue.com)

🦀 Rust CLI ⬢ Node.js 🐍 Python 🐳 Docker
QR code for nanalogue documentation site
Documentation and examples of AI usage
QR code for nanalogue CLI repository
Nanalogue CLI, Rust library, Docker image
QR code for nanalogue node bindings repository
Nanalogue Node.js/TS bindings
QR code for pynanalogue repository
Nanalogue Python bindings