Skip to main content
Premium Trial:

Request an Annual Quote

SelfDecode Moves Into B2B Services as it Seeks to Diversify, Increase its Reach


CHICAGO – Backed by new venture capital and crowdfunding, personalized health software company SelfDecode is gearing up to move into the business-to-business market, part of a larger diversification effort.

Last month, the Miami-based company closed a Series A investment round worth about $8 million. While $7 million came from private placements, more than $1 million was the result of a crowdfunding campaign through WeFunder.

Even after the venture capital round, SelfDecode is continuing its crowdfunding as a way to build a community of user-investors, including physicians as well as consumers. "That's what they want, and we want them to feel more engaged with what we're doing," said Joe Cohen, the company's founder and CEO.

SelfDecode currently analyzes DNA, environmental factors, and medical lab results uploaded by users with deep learning, Bayesian machine learning, and hyperdimensional computing methods to produce wellness reports and personalized health recommendations to help customers build action plans.

Cohen said the firm seeks to "empower individuals to uncover mysteries about their body that allow them to improve their health." However, he said that the B2B strategy is more sustainable in the longer term than marketing directly to consumers.

SelfDecode said it plans to expand its research and development team and begin partnering with dietary supplement companies, health app developers, and independent practitioners with the aim of reducing healthcare costs via disease prevention and providing clinical decision support systems to medical providers.

Specifically, it will look to deliver customized reports and polygenic risk scores to its partners. This work will mostly involve building a new interface because the back-end technology like variant imputation, collection of ancestry information, and calculation of polygenic risk scores is already there, and should be ready in about two months, according to Cohen.

He said the company's variant imputation has reached an accuracy of 99.7 percent, but it needs to get closer to 99.99 percent to be "medical grade" for sussing out rare single-nucleotide polymorphisms.

He also touted SelfDecode's commitment to "legit science," including genome analysis, and, soon, peer-reviewed research and independent validation of its technology and methods.

SelfDecode plans on targeting corporate clients by offering personalized health recommendations to those firms' customers based on polygenic risk scores and other genetic data.

Cohen mentioned as potential partners Nebula Genomics, the George Church-founded DTC genetic testing firm that was acquired by ProPhase Labs in August. ProPhase has subsidiaries in molecular diagnostics, precision medicine, and dietary supplements. He also referred to DTC vitamin and nutrition purveyor Care/of, which sold a majority stake to Bayer a year ago.

"Supplements are a commodity. These companies are usually marketing companies, and they want to stick out in some way, so they want some kind of precision health model," Cohen said, but they are missing laboratory and genomic tests, as well as data related to environmental risk.

SelfDecode wants to offer both health app and supplement firms back-end support for genomic data and analytics. It recently began selling white-label versions of its technology so customers can add their own branding to its reports.

As it seeks to win the trust of healthcare professionals, SelfDecode plans to publish peer-reviewed research on its polygenic risk scores, imputation method, ancestry profiles, and data compression. In a taste of what is to come, the firm reviewed various imputation methods in a preprint posted to BioRxiv last month.

SelfDecode started in 2016, but Cohen said it took until 2019 to assemble the right group of geneticists, informaticians, and software developers. He said he created SelfDecode because he had numerous chronic health issues, including fatigue, pain, brain fog, and anxiety, that his doctors could not fully explain, and he wanted to explore potential genetic causes.

Years before SelfDecode came to be, he created a site called SelfHacked, where he wrote about how individuals can take better control of their health and healthcare. That site still exists as a repository of consumer health information that the company says is "medically reviewed."

Cohen said SelfDecode had $2.7 million in revenues over the past year and currently has 16,000 active, paying members, including 1,000 healthcare practitioners. The firm has had about 70,000 customers since its inception.

SelfDecode has 80 full-time and 10 part-time employees, some of whom are partially compensated in stock. One prominent recent hire is Chief Technology Officer Jason Merkoski, who was a member of the team at Amazon that created the Kindle e-reader.

The firm offers its own mail-in array-based genotyping tests through Eurofins Scientific and can ingest blood test reports from labs including Quest Diagnostics. The company also has a series of software tools to provide actionable steps people can take to improve their health based on test results and genotype data.

Cohen claimed that SelfDecode has built infrastructure that might cost $100 million if done in a different fashion, but the firm only spent a fraction of that amount by reinvesting profits, hiring people overseas, and having a mostly virtual workforce to save on overhead.

The company is also actively working on data compression, he said, since uncompressed genomic data is too costly to store and analyze. According to Puya Yazdi, the firm's CSO, SelfDecode designed an algorithm it plans to patent that takes into account patterns found in genetic information to preserve the most important details of annotated sequences while radically reducing the size of data files.

Preliminary in-house testing has shown that SelfDecode can achieve about 400X compression with no loss of data fidelity, according to Yazdi, who is a physician by training. As new genomic sequences are added, the compression algorithm looks for population-level patterns to allow the database to grow at a slower rate than if all the sequences were stacked on top of each other. SelfDecode is benchmarking its compression algorithm against public software utilities that only reach 10X lossless compression, he added.

Looking two to five years into the future, Cohen wants the SelfDecode platform to move into "mainstream healthcare," specifically health systems and insurance companies. He pledged not to share individual data with payors but to help them with population health initiatives.

Eventually, he said he wants the firm to seek US Food and Drug Administration clearance for lab tests that combine genetics with environmental data to improve the accuracy of disease risk scores.