Datum Line: Do my cells look “normal”?

Kat Titterton, San Francisco

Evidence shown is public JUMP Cell Painting data (CC0). Click, or press the space bar or an arrow, for the next view or slide; press D for sources and figure legends; scroll to browse.

Documents: white paper, product vision, roadmap, landscape, brand

Every lab’s cells.
One “normal”.

Do my cells
look “normal”?

Each lab judges “normal” alone.

Kat Titterton

Eleven labs.
Eleven “normals”.

Send the controls.
Know it’s “normal” before the screen.

Controls catch drift
before any drug is read.

The product is the data.
One drug: nuclei ×1.7 to 2.4 in all 11 labs.

11 labs’ public images, plus yours.
Only control wells are shared.

Every lab that joins sharpens “normal”:
error 6.2 with one lab, 2.8 with five.

Every likely buyer: $84M/year.
Five-year target: a tenth, $8.4M/year.

Each tool sees one lab.
Datum Line sees every lab.

Raising $1.8M pre-seed for 18 months:
first paying members, then the seed.

There’s a tile for your lab.
Let’s map “normal” together.

The rest, for questions.

The usual correction erases most of the lab’s fingerprint.

The same drug looks less alike in another lab.

One brightfield image already names the lab.

On the map, labs agree a little more on the same drug.

Only control wells are shared.

An image goes in; a position and a correction come out.

Every lab leaves a fingerprint, so a wrong label stands out.

About $1.8M for eighteen months, mostly people.

About $84M a year, if every likely buyer joined.

The lab keeps its data and a share; anyone can check it.

Every lab’s cells.One “normal”.

Do my cellslook “normal”?