Predictive Cancer Classification Algorithm Based on Similar Gene-Expression Signatures

Smith Yoav, HUJI, School of Medicine - IMRIC


  • Current cancer diagnostic-methods rely on descriptive histopathological data.
  • New technology measures patient genomic data against detailed classification of cancer types from databases of measured  tumor -genomic expression profiles.
  • The method monitors genetic changes enabling better diagnostics and therapeutics.

Our Innovation

This new tool measures the similarity between gene data derived from DNA gene testsof a patients malignant tissue with sets of gene test data from pre-classified malignancies (microarray, deep sequencing, Nanostring, etc.) . The actual similarity distance obtained provides a powerful and sensitive diagnostic tool for each patient.


Key Features

  • Increased diagnostic and detection accuracy, reclassification into new subgroups. A clear similarity value to different clinical states.
  • Enables cancer diagnosis, prediction of clinical outcomes and formulation of therapeutic approach.
  • General method may be used for any type of disease if an adequate size sample of microarrays from previously classified disease is available.

The Opportunity

Seeking industry cooperation for further development with companies that can prepare datasets


Researcher Information          bioinfo.md.huji.ac.il/contact.asp



Contact for more information:

Mel Larrosa
VP Business Development Healthcare
Contact ME: