Definition
Plain language
A technique that measures gene activity in individual cells, producing very noisy data.
As stated in the literature
A genomics assay quantifying RNA molecule counts per gene per cell with sparse, noisy reads, often denoised via methods like MAGIC; serves as a benchmark domain for self-improving AI in the SIA paper.
Also called: scRNA-seq
Why it matters: The combination of high stakes and noisy data makes scRNA-seq a useful proving ground for whether AI methods truly improve over standard analysis pipelines.
For example, a single experiment might report how many copies of each gene's RNA were detected in tens of thousands of individual cells — but with most counts missing or noisy.
Heard on the show
“And then the third domain, which is the one we opened with — single-cell RNA-seq denoising.”Episode 088 — Two Levers for Self-Improving AI: When Rewriting Code Isn't Enough