Sunday, February 5, 2017

Digital Genomics: DarwinHealth as an Oncology Case Study

In early 2016, I gave a talk for the first time on Digital Genomics, creating a short white paper as a leave-behind for attendees (here; see also here).  This coming summer, I have the opportunity to chair a panel on Digital Genomics at the 1000-person NextGenerationDX conference in Washington August 15-18 (here).

The new issue of Nature Reviews Cancer contains a fascinating deep-dive article on the concept lof clinical digital genomics by authors at Columbia University and DarwinHealth, a New York-based genomics startup.

More after the break.

The article is, "The Recurrent Architecture of Tumour Initiation, Progression, and Sensitivity," by Andrea Califano and Mariano Alvarez (here; subscription).  A key illustration and "box" are open-access, here.  

The authors argue that we need to move far beyond a simple gene-drug analysis when looking at somatic mutation panels.  Rather, they argue, we will get much further both for drug discovery and for treatment by creating complex in silico cell biology models - which can be customized at the level of a single person's tumor - focused on tumor checkpoints and master regulator proteins.   The argumentation and evidence are multi-faceted and quite interesting.

In addition to this detailed academic research paper, the group hopes to implement the ideas through a new company focused on advanced genomic bioinformatics, DarwinHealth (website here; page of "offerings" is here.)   DarwinHealth is based in New York.  Appropriately, the Nature article also had two pages of coverage in The Economist (here).

In December 2016, DarwinHealth announced an international alliance with the Japanese pharma Daiichi Sankyo (press release here; D-S website here; D-S annual report in English here.)  If I convert yen correctly, D-S annual sales are circa $10B (for comparison, Amgen's are circa $20B).

DarwinHealth's plans include (product page here):
  1. DarwinOncotarget, a diagnostic platform of aberrantly active, pharmacologically active proteins "independent of DNA mutational state."
  2. DarwinOncoTreat, a "therapeutic platform" that prioritizes..agents for a patient-specific tumor."
  3. DarwinOncoDiscovery, a pharmacological drug discovery platform.
  4. DarwinOncoSphere, a "suite of products and services for biopharmaceutical companies, as well as "diagnostic and therapeutic prediction technologies that hospitals, research centers, and clinical centers of excellence can deploy to improve design, biomarker profiles, and results of clinical trials."
Cancer Proteomics

Two other quick notes on cancer proteomics.   

The same issue of Nature Reviews Cancer contains an article by Carl Borrebaeck of the Karolinska Institute on "Precision diagnostics: Moving towards protein biomarker signatures of clinical utility in cancer," (here, subscription).   

Second, billionaire biotech researcher Patrick Soon-Shiong has emphasized that his company Nanthealth could be able to upend our clinical management of cancer through proteomics.  They describe their ability to undertake very advanced protein analysis in paraffin block tissue its GPS test (here, here).   Another flag in the ground for the potential future of next-generation proteomics in oncology.  For a still-earlier effort at patient-specific pathway modeling, Vaske et al. 2010, here.


Another company in the space of predictive cancer analytics is Ariana Pharma, here.
For a 2015 paper by Creixell, Lincoln Stein, and others on in silico modeling of cancer pathways, here