Automatic search for statistics

Komorebi | 木漏れ日
“Sunlight filtering through leaves” [1]

Can’t stand hunting through a cytometry dashboard to find the exact filters to be able to do statistically valid comparisons? Neither can we. 

Out of that frustration, we’ve just released “Automatic search for valid statistics”, now on your app.teiko.bio dashboard.

The Challenge: Clinical datasets have missing values. It’s a reality of running clinical trials: patients miss a dose, something doesn’t get measured, etc. The show must still go on. 

However, finding exactly which sets of samples you can even do a statistical comparison on is still too tough. 

For example, your favorite graphing software stumbles on a deceptively simple dataset like this:

Time	Responder
Baseline	27
On-Treatment	17

To make valid paired statistical comparisons for responders, you have to hunt for all 17 subjects that have both a Baseline and On-treatment samples, set the filter and run the test. This is both error-prone, time-consuming and difficult, because you would have to find the 10 samples (i.e. 27 Baseline – 17 on-treatment) that don’t have a corresponding pair.

The Solution: Our new automatic search tool scans our statistical test database for “tallies” – complete records based on a specified endpoint and cell types. We then auto-apply the right filters so you can see valid statistical comparisons.

In the case above, the result of the automatic filtering, you would see only the 17 baseline samples that had a corresponding non-responder pair. Compare this to trying to find all 10 samples that don’t have a corresponding pair. Huge time saver.

Available now at app.teiko.bio, this feature simplifies high-dimensional cytometry analysis for clinical trial teams. Check out the demo at app.teiko.bio.

If there are no statistics showing on the dashboard, this button becomes enabled, allowing you to do this automatic search.

[1] Literally “sunlight filtering through leaves,” this term encapsulates the natural, transient interplay of light and shadow, evoking a sense of wonder and subtle beauty.