By Allison Proffitt
April 21, 2017 | “We’re taking a slightly broader view of the world than we perhaps have done in the past,” Hannes Smarason, WuXi NextCode’s newly-minted CEO, told Diagnostics World earlier this month. That’s almost a challenge for a company with outposts in Reykjavik, Iceland; Shanghai, China; and Cambridge, Massachusetts, and whose provenance traces back to DeCODE Genetics on the Icelandic side and WuXi AppTech on the Chinese side.
“The vision we have is to essentially enable anyone to use the genome across a multitude of applications. That could be anything from research, to clinical diagnostics, to health and wellness,” he said. The term he favors: a Contract Genomics Organization, an underlying platform for genomic data that brings together everything needed to empower healthcare.
The model is Bloomberg in the 80s and early 90s, Smarason explains. Bloomberg consolidated publicly available financial data, news stories, and analysts’ reports into a single platform. The WuXi NextCODE platform aims to do the same thing for genomic and health data, “and then for us to be able to earn a return in the form of licensing access to the software, accessing the data, as well as helping to broker transactions,” Smarason said.
Smarason is well-positioned to make it happen. He co-founded NextCODE Health in 2013 as a spinout from deCODE genetics. Previously he had been CFO and executive vice president of business and finance at deCODE from 1997–2004. He oversaw NextCODE's acquisition by WuXi AppTec in 2015 and its merger with the WuXi Genome Center to create WuXi NextCODE.
When NextCODE Health spun out of DeCODE in October 2013, the company took with it the Genomic Ordered Relations (GOR) database, a system for storing DNA data developed during Kari Stefansson’s mass genotyping campaign in Iceland in the early 2000’s, interpretation algorithms, and Clinical Sequence Miner, a sequence analysis engine.
“We’re building on this legacy technology from DeCODE,” Smarason said.
At the end of February this year the company reorganized. Smarason formally took the CEO mantle; he’d been COO since NextCODE’s acquisition. Two WuXi AppTech senior vice presidents took c-level positions as well.
John Long has served as WuXi AppTec's senior vice president of finance since 2013. He was actively involved in WuXi AppTec's privatization from NYSE in 2015 and played important roles in WuXi AppTec's subsequent corporate restructuring, supporting WuXi Biologics's recent IPO filing process in Hong Kong as well as private placements in the China capital market.
Alex Fowkes, now WuXiNextCode’s COO, joined WuXi AppTec in 2012 initially to lead corporate development and then most recently serving as senior vice president of commercial operations.
Fowkes brings valuable, global experience from Pfizer, Smarason said. Long is “helping to drive forward our own plans for both fundraising as well as future listing.”
The transfers are “tremendous assets for us,” Smarason said, “giving us the diversity and experience base to rapidly scale from where we’ve been as a company.”
Smarason calls the company, “natively global,” alluding to its reach and roots across three continents. But that makes for more than just time zone jokes and a 24-hour work day. The company’s Shanghai, Kendall Square, and Reykjavik cultures impact the products as well.
The headquarters of the company is in Cambridge. The medical genetics, clinical informatics, and AI teams are based in Kendall Square as well. Much of the IT and development team—some of whom came from DeCODE—are in Reykjavik. Shanghai houses the wet lab, CLIA/CAP and research sequencing, and the sales and marketing teams for China and Asia.
The far-flung geography, “is a distinguishing feature of this big data platform that we’re building,” Smarason said. “It has to do with being able to both build and deploy a product across all the different markets.”
And it drives the business strategy. “We are leveraging this big data platform to go after two separate opportunities, in each instance taking advantage of the penetrance and the global reach we have by being a US company in the US, a Chinese company in China, and obviously being able to serve Europe as well through the location we have in Iceland,” Smarason said.
The company’s legacy business model has driven partnerships with hospitals, institutions, and other businesses. The new initiatives and products—many launched in the past 12 months—address customers directly. “They both generate revenue and data for us, importantly, which again advances the platform,” Smarason said.
Commercializing the business has, historically, depended on pharma companies, population projects, and hospital groups, Smarason said—“all the big users of genomic information that need this kind of platform.” He calls them the foundational customers of WuXi NextCODE, and the company has done well in those relationships.
WXNC boasts several pharma partnerships and has recently WuXi NextCODE has announced several wins in the population genomics space. The platform was chosen to support Singapore’s 18-month proof-of-concept precision medicine project, focusing on cardiovascular disease. Sidra Medical and Research Center in Qatar is partnership with WXNC to co-develop comprehensive research and bioinformatics programs that will link with the Qatari national electronic medical record system. Genomics England chose the platform as its first Clinical Interpretation Partner in the field of cancer, and Genomics Medicine Ireland announced a partnership with WuXi NextCODE and AbbVie this year at JP Morgan this year to sequence 45,000 volunteers.
And last year, WuXi AppTec helped launch the China Precision Medicine Cloud, although as an independent effort not sanctioned by the Chinese government. WuXi NextCODE would bring sequencing capability and provide the platform for organizing, mining, and sharing genomics big data.
The company’s aim is to have more than 2 million genomes running on the platform by 2020, and Smarason says they are close. It’s a slightly arbitrary number—“aspirational,” Smarason said—meant to underscore the power of the database and platform. But it’s also an achievable number based on the customers WXNC is engaged with. He fully expects to manage that many genomes soon.
But a rapidly growing part of the business is directed toward consumers, Chinese consumers specifically. “We’re tapping into the clinical wellness market in China; in particular we’re pursuing rare disease diagnostics and reproductive health,” Smarason said.
In 2016, WXNC launched three products for the Chinese market: RareCODE is a diagnostic test for rare disease, FamilyCODE is carrier screening, and HealthCODE is a whole-genome wellness scan. All three are built on WXNC’s genomic database, and tailored to Chinese population variants.
RareCODE launched in the second quarter of 2016 in partnership with two of the country’s largest pediatric hospitals—Fudan and Ruijin. RareCODE processed between 6,000-8,000 samples last year, but Smarason expects that number to double or triple in 2017.
HealthCODE is the company’s “expensive, executive health product,” offering whole-genome wellness scans delivered in a clinical setting by a partner healthcare institution. Smarason expects the numbers for this product to be much lower, but reports “really interesting data” coming from the first 200 cases. The scan uses WuXi NextCODE’s proprietary risk modeling to gauge each individual’s inherited risk of 28 common complex diseases, and at this year’s American College of Medical Genetics annual meeting, the company reported that average customers of the product have heightened risk for four of the diseases. For example, roughly 18% of participants are at >1.5 times average inherited risk for type 2 diabetes.
The company hopes its tests can prompt better prevention for individuals, as well as create a growing resource of whole-genome data that can be correlated with medical and outcomes data to develop an ever more refined picture of genetic risk factors and effective prevention strategies, specifically in Chinese populations.
FamilyCODE, which launched at the end of 2016, is a carrier screen for known risk factors in the Chinese population for some 200 diseases. The company expects it to be very successful. “The reproductive health area is of key interest there because of the end of the one child policy,” Smarason said. “The China market, by comparison to the US, is about five times the size. You’ve got about 20 million pregnancies a year.”
The tests were developed at WuXi NextCODE’s Cambridge headquarters, but are only available in China. “As you know in the US it’s a highly-regulated area,” Smarason said. “If we did launch in the US we’d have to take it through an FDA process. In China we don’t have that regulation as tightly in place at the moment, which gives us a little more flexibility.” Even in the absence of regulation, the test is being done to the highest US and global standards in WuXi NextCODE's lab, which is the first CLIA and CAP certified sequencing facility in China.
Currently about 80% of WuXi NextCode revenue comes from business-to-business clients; 20% comes from consumer products. But that’s changing quickly, Smarason says. By 2020, he expects 60% of the company’s revenue to come from consumer products and services. That’s won’t be a shift in attention, but will instead reflect faster growth on the consumer side.
All of this work, for business or consumer customers, expands the company’s knowledgebase, and DeepCODE is the effort to mine that knowledgebase with deep learning tools to discover novel causal and predictive connections in datasets. “[We want to be] able to take sources of information based on DNA, RNA, and arrays using our novel feature selection techniques to develop all of the important layers that are necessary for deep learning to take place, and be able to derive some novel insights from that,” Smarason said.
The company’s legacy systems were based on traditional, statistical algorithms, Smarason explained. In 2015, the company appointed Tom Chittenden, a quantitative scientist, to create the company’s Advanced AI Research Laboratory.
DeepCODE methods can both automate and improve the accuracy of tumor subtype and drug response classification, the company says. This month WXNC presented a poster at the American Association for Cancer Research meeting outlining how DeepCODE’s novel, multinomial statistical-learning classified 27 different tumor types with greater than 95% accuracy when applied across approximately 9,000 human tumors from The Cancer Genome Atlas (TCGA) collection.
“We’ve done other work looking at responders and non-responders to shelved drugs, where we’ve been able to pinpoint with 100% accuracy which patients will fall into which category. We’ve also been able to look at breast cancer—ER-positive vs ER-negative—and predict which category you would fall into with 98% accuracy. It’s becoming one of the key areas of research for us,” Smarason said. “We’re seeing interest on the pharma side for this technology, but also we’re increasingly integrating this into our core computational engine so that this is now becoming part of what makes our tools unique.”
Genomic datasets are orders of magnitude more complicated than past data problems, Smarason said. “That is where we’ve been investing most of our time. Taking the leading edge thinking in AI… and layering on top of that deep domain expertise to try to advance the field further, and then adding on top of that all of the data we have for training purposes. We can actually come up with some pretty novel insights.”