10x Genomics has launched Sentira Single Cell, a cloud application where researchers can ask questions about single-cell data in plain English and have a Claude-based AI agent plan and run the analysis. The useful shift is that researchers can inspect and revise the plan, then get the code and an editable notebook behind the results, rather than waiting for custom analysis code.
Watch Desk analysis
What happened
The application launched on 2 November for researchers using 10x’s Chromium single-cell platform, according to R&D World. Users can bring a hypothesis or ask Sentira to suggest one from their data; the agent proposes a plan, runs it and supports follow-up questions. The analysis runs on 10x Cloud.
Each run produces an editable notebook recording the methods and code, written in Python with packages including Scanpy. 10x says Sentira Single Cell is priced by consumption and offers promotional credits to get users started.
Why it matters
Single-cell researchers can face a coding bottleneck when comparing samples or testing a biological question. A natural-language interface that exposes its plan and code could let more researchers explore their own data directly, while leaving a trail they can inspect and edit. That is a more useful promise than “the AI did some science”, which is not much of a methods section.
There is an important limit: as of 4 November, 10x had not published accuracy figures for its tests against datasets previously analysed by a computational biologist. The company says it plans to publish the results and evaluation framework. Until then, the workflow is tangible, but the reliability evidence remains forthcoming.
Our read
The combination of plain-language analysis and inspectable code makes this a genuinely practical AI application for a specialist research workflow. The notebook is the right kind of receipt; the next question is whether the results stand up across more than the company’s own tests. Researchers should treat it as a way to explore and accelerate analysis, not a substitute for checking the biology.
What to watch
- Whether 10x publishes the promised accuracy results and evaluation framework.
- How researchers assess and revise the agent’s plans and outputs in practice.
- Whether consumption pricing and promotional credits make repeated analyses practical for labs.
Discussion spark: Would you trust a natural-language AI workflow for exploratory single-cell analysis if it exposed editable code and methods, or should researchers wait for published accuracy results first?
Sources and evidence
- Source update (5 November 2026, 00:35 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.