Every tumor genome carries genome-wide structural patterns that conventional single-gene tests discard. TESA Research's platform was built to discern them.
Most biomarkers read one gene: a narrow view. But cancer behavior emerges from interactions across the genome. We define a tumor by its whole genomic architecture: the position and type of every event, read together with molecular data.
One variant is measured; everything else in the genome is ignored. Useful, but blind to the larger structure that shapes tumor behavior.
Genes act in concert. Disease course and drug response are properties of the network, not of any single node within it.
We read the tumor genome as a pattern: where every event occurs and what kind it is, integrated with molecular and clinical data.
The output of our platform is not a score, it's a set of tumor subgroups. Each is defined by a reproducible genomic architecture implicating a biological process which ties to prognosis and likely therapy response.
Cytogenetic and structural data, sequence and expression data, and clinical outcomes, from any platform, liquid or tumor biopsy.
Our proprietary encoding preserves the location and type of genomic events which are often discarded by other methods.
The model surfaces biologically coherent, reproducible tumor subgroups tied to disease states, prognosis, and drug response.
The expensive multi-omic work happens once, in training. Classifying each new patient requires only DNA.
Working with leading academic collaborators, we have identified statistically significant prognostic and predictive signatures across multiple cancer types, including colorectal cancer, breast cancer, brain tumors, chronic lymphocytic leukemia, and pediatric Ewing sarcoma.
Prognostic tells you how aggressive the disease is. Predictive tells you who benefits from therapy. Our signatures do both.