NEW YORK – With a recent infusion of private equity funding under its belt, Watershed Informatics, a Cambridge, Massachusetts-based startup, is trying to bridge the gap between the so-called "high-code" and "no-code" philosophies of bioinformatics to support multiomic analyses.
There is no shortage of commercial or academic bioinformatics software developers touting the power of multiomics.
A study that appeared in Nature just last week discussed a method to generate a synthetic multiomic dataset for the UK Biobank, which then supported a phenome-wide association study using PheCodes. Also last week, a pan-European research team announced plans to develop biomarkers for diagnosing hypertension from multiomic datasets.
On the commercial side, dozens of companies are talking about multiomics or single-cell analysis combined with spatial transcriptomics. A sign of that sector's growth is the growing wave of consolidation among vendors of multiomics research informatics.
In January alone, Pierian, Seven Bridges Genomics, and UgenTec merged to form Velsera, and Precision for Medicine subsidiary QuartzBio purchased SolveBio.
The founders of Watershed believe that all of their would-be competitors are on the wrong track to some degree because they are taking either a high-code or no-code approach.
High-code products are flexible and scalable to handle all kinds of workflows but require significant bioinformatics expertise to customize, maintain, and adapt software installations. "The problem is you need to be some unicorn systems administrator/bioinformatician/software engineer if you want to get that [approach] off the ground," said Watershed Cofounder and CSO Mark Kalinich.
No-code is a relatively new phenomenon in the world of web development that allows people without specific programming knowledge to build software via graphical user interfaces. Startups LatchBio and IBM Research spinoff Almaden Genomics are among those following this approach.
The drawback with no-code software is that it is not very flexible for the 10 to 20 percent of research projects that differ from a laboratory's normal procedures, according to Kalinich. He said that Watershed's software is flexible like high-code products but is fast and user-friendly like no-code applications.
"You don't need slightly better tools. You don't need slightly better [computing] infrastructure. You don't need slightly better sequencers. Those things are mature," said Cofounder and CEO Jonathan Wang. "We're providing the thing that's missing, versus just assuming that strengthening one part of it is going to solve the entire problem."
Watershed's product is Omics Bench, a cloud-based software platform to support multiomics research. Omics Bench includes workflow and visualization tools for whole-genome sequencing, single-cell and spatial transcriptomics, epigenomics, proteomics, metabolomics, microbial sequencing, and protein folding.
Wang called Omics Bench "a platform that unifies the ecosystem" of raw sequencing data, fairly commoditized computing infrastructure, and disjointed software tools, free and otherwise.
The firm last week closed a $14.5 million Series A round led by Canvas Ventures, with participation from previous investors Bessemer Venture Partners and Accomplice Ventures. Watershed also raised a $1.7 million seed round in 2021.
The company currently has 17 employees. With the Series A funding, Watershed will be looking to accelerate software development as well as hire sales, marketing, and technical support staff, according to Wang.
Omics Bench was released early last year after a lengthy beta period, but the technology has been in development since 2019, a year before Watershed was incorporated. The concept goes back much farther.
Kalinich and Wang have known each other since they were freshmen at Massachusetts Institute of Technology in 2009. Kalinich studied chemical and biological imaging and gained experience with targeted drug delivery research in the laboratory of Robert Langer.
Kalinich went on to an M.D.-Ph.D. program at Harvard University in hopes of pursuing a career in cancer diagnostics.
"I was shocked to find out that making data had become surprisingly easy. I could [assemble] a whole genome in a couple of days," he recalled. "But it was really the analysis and the interpretation of that data that prevented me from building my diagnostic."
In talking to friends and colleagues in biotechnology and biopharma, Kalinich realized that others were having the same problem. "This had become the defining obstacle preventing scientists from building better drugs in diagnostics faster," he said.
"The tools are there, the tools are good. The problem is that the tools and the data and the expertise are fragmented," Kalinich said. Omics Bench attempts to bring everything together.
"The reason we know there is an urgent unmet need is that I had an urgent unmet need when I was trying to perform these exact manipulations," Kalinich said.
Wang studied electrical engineering and computer science at MIT, then cofounded a high-frequency-trading hedge fund, where he encountered a similar problem in the financial world that Kalinich did in biotech.
That firm, called Domeyard, subsequently built a data-driven research platform to support automated trading. "We thought that if we just hired a bunch of smart people, bought a lot of really expensive computing servers, and gathered huge amounts of market data that we would somehow eventually be able to arrive at very powerful trading insights," Wang said.
It turned out that these "smart" people who knew hedge funds did not know how to run experiments with big data, so the company had to bring in external IT professionals.
"Very often, [researchers] were sitting around waiting for software developers to come by and help them build these distributed computing [applications]," Wang said. So the company built a platform that automated the behind-the-scenes work that goes into assembling data-intensive experiments.
Wang said that he and Kalinich soon realized that biologists faced similar issues with lack of software development skills limiting their ability to use big data. End users just want actionable insights. "Everything else … gets in the way," Wang said.
Watershed's target market theoretically is anyone working "at the intersection of computation and biology," though the current customer base is in data-intensive, omics-focused preclinical R&D, according to Kalinich. He said that Watershed will soon be releasing a module for managing imaging data, which would put it in competition with firms like Owkin and PaigeAI.
The user base includes academic researchers and early-stage biotechs. Watershed has named SalioGen Therapeutics, Cellino Biotech, Revitope Oncology, SQZ Biotech, and Rarebase among its clients.
The company's operations are only in the US right now, and there are no immediate plans to expand internationally. "There's definitely no lack of innovative companies in the US and even in particular within a 1-mile radius of our headquarters trying to build the drugs of tomorrow," Wang said.
Wang said Watershed is focusing on preclinical R&D because there is so much interest in shortening the drug-development process. "Why is it taking 10 years to develop a drug when there's so much information locked away in data that's already been collected?" he wondered. "People simply don't have the resources to utilize it."