A Gene Expression Barcode for Normal and Diseased Tissue Classification

This invention measures absolute gene expression and shows that with hundreds of samples it is possible to demarcate expressed from unexpressed genes. The utility of the method is demonstrated by defining a gene expression barcode for human and mouse tissues, which is then used to differentiate diseased tissue from normal tissue. With clinical data, we find near perfect predictability of normal from diseased tissue for three cancer studies and one Alzheimer's disease study. The barcode method also discovers new tumor subsets in previously published breast cancer studies. This invention classifies normal and diseased tissues based on gene expression. The method is not constrained to a specific tissue or cell type, but can be used generally for any sample in any organism. The increased resolution of the barcode algorithm allows for the identification of novel subtypes in heterogeneous tissue types such as cancer. The use of this algorithm is similar to a BLAST search for DNA sequence data, except extended microarray data. Description (Set) Proposed Use (Set) The novel features of the invention are the statistical algorithm and software implementation to classify tissues. Many groups have designed these types of algorithms, but they are all based on continuous data, which is highly variable for microarrays at this point. This invention converts microarray data to discrete variables, which eliminates variability, making the algorithm simpler, more accurate, and more flexible than the previous algorithms. The algorithm can be used for diagnosis, prognosis or treatment decisions in cancers and potentially many other diseases.

Inventor(s): Irizarry, Rafael

Type of Offer: Licensing



Next Patent »
« More Software Patents

Share on      


CrowdSell Your Patent