Genomenon’s Software Seeks to Give Geneticists Faster Insights

2/22/2017

Genomenon, a University of Michigan spinout developing analytics and data visualization software for the genomics industry, released its first product last week: software designed to simplify the process of interpreting gene variants.

Called Mastermind, the software is a tool that automatically combs through millions of journal articles and medical publications; its goal is to drastically reduce the time pathologists and geneticists spend examining medical data and to provide faster insights.

Mike Klein, Genomenon’s CEO, says the software uses algorithms to sift through the medical literature on publicly available sources, such as PubMed, and determine the relationship between specific diseases and gene mutations. Mastermind then organizes the data into clinical categories prioritized by the strength of those relationships.

“Mastermind doesn’t draw conclusions, it takes clinicians directly to the literature,” Klein says, noting that PubMed searches titles and abstracts, but not the article’s full text.

The challenge geneticists face, Klein says, is that only a fraction—about 6.6 percent—of the disease-gene mutation relationships explored in full-text articles are covered in the titles and abstracts that PubMed searches. Mastermind solves this problem by performing a full-text search of the primary literature, improving the accuracy of interpretation and reducing the time clinicians spend searching for articles by up to 80 percent, he explains.

Genomenon was co-founded by Mark Kiel, a molecular pathologist at U-M who knew this challenge firsthand. Before he focused his attention on Genomenon, he spent the majority of his day searching Google, PubMed, COSMIC, and HGMD for information—which didn’t leave much time for interpreting the results.

“It was very frustrating—he had all this knowledge, but was spending his time doing labor-intensive searches,” Klein says. “He’s a true entrepreneur; he ran headlong into his problem, took that experience, and figured out how to automate a rote process. People had been looking at it from a standard IT perspective, but they needed the secret sauce of Mark’s genetic insight.”

In developing Mastermind, Klein says Genomenon spent three years poring over 3.3 million full-text articles and found every disease-gene mutation combination covering somatic cancer, hereditary cancer, cardiomyopathy, and infertility. The company plans to continuously expand its database, adding between 250,000-500,000 articles per month, and hopes to have 7 million articles in its database by mid-year.

Klein says Genomenon is going after a global market that is expected to be worth $1.3 billion by 2021. Although there are other genomic data platforms, such as Watson Health, he believes his company is forging new ground. He says there are two other approaches to what Genomenon does: manual curation, or “doing what [Genomenon does] but not automatically,” and population studies that look for correlations. “We’re bringing a third way,” Klein says.

The 10-person company, which is based in Ann Arbor, MI, plans to sell access to Mastermind on a subscription basis. To help get Mastermind to market, it closed on a $1.8 million seed round in November, with investments from two U-M funds and several angel investors.