Skip to main content
Back to timeline
BiotechnologySource publication:

Illumina wins the Virtual Cell Challenge with its Billion Cell Atlas: CellularAI predicts perturbation responses in unseen cell lines

Synopsis

Illumina's BioInsight AI Lab took first place in the 2026 Virtual Cell Challenge hosted by Arc Institute, with results powered by data from the Billion Cell Atlas and a new cellular response prediction algorithm, CellularAI. The challenge asked teams to predict knockdown responses in unseen cell lines using only each cell's unperturbed state, and among roughly 1,500 competitors Illumina's predictions most closely matched hidden experimental data generated by Arc. An open-source version of the algorithm was released today so researchers can examine the methodology.

Source-provided article image: Illumina Wins Virtual Cell Challenge, Powered by Billion Cell Atlas
PR Newswire

Interpretation

Illumina's BioInsight AI Lab won the 2026 Virtual Cell Challenge hosted by Arc Institute, a contest to predict cellular responses to genetic perturbations in unseen cell lines. Teams were given only the cells' unperturbed state and had to predict knockdown responses from it, and Illumina's predictions were powered by Billion Cell Atlas data together with the new CellularAI algorithm. Illumina's predictions were scored against hidden experimental data generated by Arc, and among roughly 1,500 competitors CellularAI's predictions most closely matched real, validated experimental results across a range of biologically meaningful metrics.

CellularAI predicts how cells respond to genetic perturbations and can produce predictions in cell types not yet experimentally tested. The Billion Cell Atlas uses CRISPR to capture how individual cells respond to genetic changes across more than 200 disease-relevant cell lines, with data generated by Illumina in a consortium with select pharmaceutical partners. An open-source version of the algorithm was released today, giving researchers a window into the methodology behind the challenge results, and resulting hypotheses can be tested experimentally through direct CRISPR capture in the Illumina Single Cell Preparation workflow.

Perspective

The report covers Illumina's placement in the 2026 Virtual Cell Challenge, the data and algorithm behind it, and the company's description of intended applications for the Billion Cell Atlas and CellularAI; scoring details and specific metric values are not given in the material.

Illumina describes the Billion Cell Atlas as becoming a foundational dataset and as positioned to help unlock predictive biology, which are forward-looking company statements; the release also notes inherent challenges in developing, validating and advancing AI models, and that the performance, utility and adoption of CellularAI and related technologies may not meet expectations.