Spring is on a mission to build a technology engine that gives scientists superpowers to accelerate drug discovery. 

Inflammasome biology is among the many biological processes that involve complex and convergent pathways, making target identification and drug action difficult to resolve with traditional approaches. Here, we demonstrate how machine learning and advanced multi-dimensional data analysis can be used to advance drug discovery by facilitating analysis of complex, physiologically relevant data at very high throughputs and extremely low costs.

We designed a series of high content imaging-based screening experiments on primary human PBMCs under physiological inflammasome activation conditions (flagellin and ATP) to develop novel machine learning based analytical tools. 

Specifically, we built unique phenoprints for different biological compounds and pathways, identified novel single-cell phenotypes, and advanced 300+ drug candidates for 7+ targets to potential clinical programs.

Spring’s tooling enabled analysis of over 110M images of cells were analyzed across 100k unique donor, drug, and treatment condition combinations in order to develop multiple scoring rubrics designed to categorize compounds into novel inflammasome inhibitor classes whose mechanisms of action can be uniquely matched to relevant human pathologies.