Have you noticed that every latest company in the cancer diagnosis business comes out swinging with a new type of data that they claim offers the best way to detect cancer? Why is cancer detection so elusive? And why do companies have to keep racing to find new data types to convince themselves and the investors that their type is better? How true are all these claims?

As we have found out through our own, decade-long, research, it is not the data type that is the problem but rather the analytical paradigm with which we process that data through. All cancer data is extremely heterogeneous. Any two cancer specimens are not congruent in most of their data points, which makes it difficult to find common denominators for a type of cancer that can be used as biomarkers or even for subtyping! This fact knocks out at least two paradigms, statistics and machine learning, as valid methods for cancer analysis or diagnosis.

Add to the above fact that cancer is a dynamic disease where its cells keep changing their biochemistry on a spectrum starting with the initiation phase and ending with the metastatic phenotypes. So, is it hopeless to find the cancer-compatible analytical paradigm? The answer is a confident NO.

We have found the heterogeneity-compatible analytical method that gives us successfully early detection and diagnosis from several types of data. We tried it on many data types like metabolomics, proteomics, and gene-expression and it worked. It worked so well that we patented it.

Therefore, we are inviting you to join our startup company, Phyloncology, and see for yourself that we are able to carry out early cancer detection and diagnosis. Why are we confident in our work? For a simple reason: we have perfected the method. After all, it is the analysis and not the data type!