NEW YORK – As the venture capital market has cooled in the wake of rising interest rates, personalized health software company SelfDecode has "set aside" a planned new VC round and quietly raised a $2 million add-on to an earlier seed round, according to COO Ralph Kenney.
The Miami-based firm has also pulled in approximately $250,000 on crowdfunding platform WeFunder in recent months. To date, SelfDecode has raised nearly $1.4 million from 1,288 investors on WeFunder, and that campaign remains open.
Near the end of 2021, SelfDecode raised $8 million in venture capital that the firm is now calling a seed round and moved into the business-to-business market.
Initially, SelfDecode looked for partners making consumer-facing digital health apps. This year, the firm started talking to potential collaborators from both industry and academia to create laboratory-developed tests, according to Founder and CEO Joe Cohen. He said that LDTs for cardiovascular disease "seem like the most promising right now" because the company is touting its strength in developing ancestry-adjusted polygenic risk scores, which tend to be good predictors of cardiovascular health.
SelfDecode recently filed for a US patent on an algorithm that adjusts polygenic risk scores based on an individual's ancestry. While the application does not show up on the US Patent and Trademark Office's search function, Kenney said that it is titled, "Systems and Methods to Generate Local Ancestry Determinations, Polygenic Risk Scores, and Other Useful Information From Genomic Data."
Cohen noted that SelfDecode has a research partnership with the Scripps Research Institute that will soon produce a paper on the benefits of ancestry-adjusted polygenic risk scores.
Kenney said that the company's technology has been proven to be 14 percent better at SNP recall and 16 percent more accurate than methods that do not include ancestry adjustment.. More detailed data will be presented at the European Human Genetics Conference in Glasgow, UK, this month.
"The [more comprehensive] the ancestry, the better the polygenic risk scoring," Cohen explained. "That is one of the limiting factors for getting polygenic risk scoring into the mainstream healthcare system," at least in the US, where the population is far more diverse than in many other wealthy countries.
For example, SelfDecode can combine a basic polygenic risk score for atherosclerosis with additional metrics such as LDL cholesterol and weight, according to Cohen, who further noted that this "could significantly increase the predictive power of whether somebody's going to get a disease or not."
The company uses deep learning, Bayesian machine learning, and hyperdimensional computing methods to analyze DNA, environmental factors, and medical lab results uploaded by users to produce wellness reports and personalized health recommendations to help customers build action plans. SelfDecode augments its genetic testing and analysis with lifestyle questionnaires to refine risk predictions and offer more personalized health recommendations.
Most of SelfDecode's work since the 2021 fundraise has been in building its core informatics technology, including a consumer-facing analytics engine as well as an application programming interface for businesses. However, the firm has also expanded its R&D team and begun 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.
In the B2B market, SelfDecode offers its genomics pipeline, including the ancestry-adjusted polygenic risk scoring, for developers of digital health apps, who pay a fee of about $10,000 to connect to the API.
SelfDecode continues to offer genotyping services through a partnership with Eurofins Genomics, and various blood-based health and wellness panels through a partnership with Quest Diagnostics. The genotyping partnership with Eurofins uses microarrays as opposed to whole-exome sequencing. Cohen said that a $100 microarray-based genotype that larger players such as Ancestry and 23andMe sell is actually better for variant imputation than a $500 exome sequence.
He noted, however, that exome sequencing has a place when looking for variants indicative of specific diseases, but is less useful for developing polygenic risk scores. Whole-genome sequencing, according to Cohen, "is still too expensive for the regular consumer to get."
However, anyone can upload a whole genome or exome to the SelfDecode platform for personalized analysis, including DNA analysis, on a per-use or subscription basis. Users can upload VCF files from multiple sources, though SelfDecode currently only accepts whole-genome sequencing files from Dante Labs and Nebula Genomics.
Cohen said that the firm has signed about 30 commercial contracts in the last year and a half, but added that the B2B offering is still a work in progress because partners have to build, refine, and market their apps before SelfDecode sees revenue. "We found that there's a long lead time for that," Cohen said.
For now, nearly all of the company's B2B sales are outside the US, particularly in Europe, Latin America, and Southeast Asia, according to Kenney.
Future partnerships might be with pharmaceutical companies to help them with target discovery and with predicting whether patients are more or less likely to feel side effects from various drug classes, according to Cohen.