SAN FRANCISCO (GenomeWeb) – The Mount Sinai Health system this week spun out a new company, Sema4, that will offer the same genetic tests that had been available to patients at Mt. Sinai, but will now do so nationally.
It also plans to expand the testing menu and will look to integrate machine learning and data science to move into predictive medicine, said Eric Schadt, who will serve as CEO of Sema4. Schadt will also maintain a faculty position at Mt. Sinai, as well as an active lab that will collaborate with Sema4 on building predictive models of disease.
Around 250 Mt. Sinai employees will move to Sema4. Aside from Schadt, Lisa Edelmann, who was executive director of the Mt. Sinai Genetic Testing Laboratory, will now serve as chief diagnostics officer of Sema4; George Diaz, director of the Program for Inherited Metabolic Diseases at the Icahn School of Medicine at Mount Sinai, is serving as interim chief medical officer; Jun Zhu, professor in the Department of Genetics and Genomic Sciences at the Institute of Genomics and Multiscale Biology, will serve as head of data sciences; and Rong Chen, who was previously director of the Clinical Genome Laboratory at Mt. Sinai, will now be vice president of research bioinformatics at Sema4.
In addition, the firm has hired around 50 people not previously employed at Mt. Sinai on the sales and marketing and product and software development sides. Over the next year to year and a half, Schadt said that Sema4 would look to hire an additional 100 to 150 people primarily focused on product and software design and data analytics.
He said that Sema4 launched with a large investment from Mt. Sinai that will allow it operate for the next three years. Mt. Sinai is the sole investor in Sema4, Schadt added, and the two also have a technology development agreement that includes single-cell sequencing, long-read sequencing, and data analytics.
"The move to put Sema4 on the outside [of Mt. Sinai] was driven by the desire to expand our services more broadly to beyond the New York tri-state area with products we think are very competitive," Schadt said.
From a practical point of view, the experience of patients at Mt. Sinai who receive genomic testing will not change. The same tests will still be available and the same billing procedures, Schadt said. But now, those tests will also be available to patients outside of Mt. Sinai.
The Mt. Sinai Genetic Testing Laboratory has been running between 100,000 and 150,000 samples per year, Schadt said. With Sema4, he said the goal would be to eventually quadruple those testing volumes.
The lab will be equipped with the same array of genomic technologies as Mt. Sinai, including Illumina's HiSeq 4000 and 2500 instruments, which are the "workhorse" machines, Schadt said. It also plans to use Thermo Fisher Scientific's Ion Torrent instruments primarily for somatic cancer panels. "We're among the largest users of the Pacific Biosciences technology, and that will continue," Schadt added. In addition, Sema4 will acquire a few of Illumina's NovaSeq instruments.
Currently, Mt. Sinai's genetic tests focus on reproductive health and oncology. Sema4 will offer the same suite of tests in those areas, including newborn screening, noninvasive prenatal testing, and carrier testing, as well as a range of somatic cancer panels.
For NIPT, Sema4 currently sends samples to reference laboratories for testing, but Schadt said that the company is also developing its own NIPT, which it anticipates launching in the next several months. Schadt declined to disclose further details about the in-house NIPT. It also offers four different carrier screening tests, formerly known as NextStep. Those include a standard panel that screens for four disorders, an expanded panel that screens for 281 disorders, a Comprehensive Jewish panel that looks at 96 disorders, and a panel that looks at 10 high-frequency disorders.
In the near term, Schadt said that the company plans to expand its footprint in the hereditary cancer testing space as well as to build more comprehensive somatic cancer panels that are better able to identify relevant therapies.
Sema4 also plans to develop more hybrid panels in both the oncology and reproductive health spaces that use the PacBio technology in combination with short-read sequencing technology. For oncology testing, Schadt said that the long reads of PacBio's Sequel instrument are useful for analyzing repetitive regions or looking at structural variants, while short-read sequencing technology can better identify point mutations. Similarly, for carrier screening testing, long reads are useful when screening for Huntington's disease or Fragile X, which involve repeat expansions. "Those can't be effectively tackled with short-read sequencing," he said, "so we're looking to supplement with PacBio technology to create a more comprehensive screen."
Schadt said he anticipates the first customers will be physicians, including Ob-Gyns, maternal-fetal medicine specialists, and oncologists. In addition, he said, the company is looking to partner with entire health systems like it has with Mt. Sinai, to "better use all the information that's being built up around patients," he said. "We aim to help manage the information and structure it in a way that can be integrated into the diagnostic testing and better inform on whatever conditions are of interest," he said.
As such, he said that one major new push for Sema4 will be on the data analytics and machine learning side, Schadt said. "We want to leverage the growing testing business to engage patients," Schadt said, so that Sema4 provides more than just a single test. For instance, Schadt said, if a woman comes in for carrier testing, the goal would be to engage her and provide not only that carrier screening test, but to aggregate other data — like electronic medical record data, data that can be collected from mobile or wearable devices like physical activity and sleep patterns, [and] pharmacy records — in order to provide better interpretations of the genomic tests and be able to better predict health outcomes, Schadt said.
"Machine learning will derive deeper meaning for diagnosing and treating diseases," Schadt said. However, in order to move from this testing model to a predictive model, Schadt said it would be necessary to both collect larger amounts of data and build out and improve in the data sciences space, including the algorithms and machine learning processes. Moving from an academic clinical lab to a commercial lab will enable both larger datasets to be collected as well as the investment necessary to build the algorithms, Schadt said.
Another key to this will be in convincing patients to share their data with Sema4. Schadt said that Sema4 is looking to position itself as key in empowering patients to take charge of their own data, including to share it, in order to take advantage of machine-learning tools that would ultimately serve patients by being able to better predict health outcomes and individualize treatments, Schadt said.
Ultimately, he said, Sema4 would move from being a testing company to a health information company. "Down the road it matters less and less whether we're the ones generating the data, if we can aggregate the data and provide the best interpretations for physicians," he said.