CHICAGO — Korean rare disease diagnostics startup 3billion is less than 5 years old, but the firm is already planning on going public in early 2022.
According to CEO Changwon Keum, the KOSDAQ exchange and the South Korean government have prioritized technology companies for IPOs for several years, creating a fast-track process for listing on the exchange. Plus, he noted that the KOSDAQ has been a bull market since recovering from the initial COVID-19 dip that reached its nadir on March 20, 2020.
"[That] kind of mood made us think that we can go public faster than we expected," Keum said.
The firm, founded in October 2016 as a spinoff from Korean genomic services provider Macrogen, has already raised money through the private equity market this year. Last month, Seoul-based 3billion closed a $13 million Series C funding round, bringing its total venture capital haul to $30 million since its seed round in 2017.
3billion now provides rare disease diagnostic services based on whole-exome and whole-genome sequencing in 32 countries. The firm processed 8,000 tests in 2020, up fourfold from the previous year, and expects to handle 20,000 cases this year.
Keum, a bioinformatician who previously designed clinical genome tests for Macrogen, said that the rapid growth could continue beyond 2021. "We are expanding our services to more and more hospitals [around] the world," he said.
The company will use the new funding to improve its software, hire more people, move into new markets, and, if demand is there, open offices abroad, he added.
South Korea and the US remain the company's primary markets, but 3billion is looking to target "anyone who needs our services at this point," Keum said.
3billion has its own sequencing lab in South Korea and provides its own informatics for interpretation services. Keum said that the firm will be looking to open a lab in the US in a couple of years.
"We do everything from the sequencing to interpretation and final diagnostic reporting to the medical doctors and to patients," he explained. "But our main expertise is in the software engineering technologies," including AI for diagnostics.
The firm's diagnostic software has won regulatory clearance in South Korea, and 3billion is in the process of obtaining CAP/CLIA certification for its lab, which it expects to obtain before the end of this year.
3billion's primary software product automates the 28 criteria in the American College of Medical Genetics and Genomics, or ACMG, and Association for Molecular Pathology guidelines for classifying human germline variants. "We try to cut the cost for the ACMG interpretation part," Keum said.
While 3billon's software and services are aimed at the 7,000 known human rare diseases, Keum said that the company's "indication target" includes all genetic diseases, regardless of whether they are rare or not.
The company started by offering whole-exome sequencing for diagnostics but added whole-genome sequencing in late 2020 in hopes of improving the diagnostic yield. There are no plans to offer panel sequencing.
Keum said that 3billion provides reanalysis services for no additional charge to patients who initially cannot be diagnosed with the company's tests in order to resolve as many cases as possible over time.
In addition to its core software, the firm has developed artificial intelligence-based technology called 3Cnet that addresses one of the ACMG rules, namely prediction of pathogenicity of variants. In a preprint article posted on BioRxiv in October, 3billion claims that the 3Cnet variant interpretation algorithm is 14 percent more accurate at predicting pathogenicity of missense mutations and twice as accurate in identifying causal variants than previous methods.
"If a prediction algorithm can predict the disease-causing variants in advance, it can reduce the time and cost required for diagnosis considerably," the 3billion researchers wrote in the preprint.
3billion achieved these gains by training the deep-learning 3Cnet algorithm on more than just the public ClinVar database but also incorporating clinical and conservation data. The company said that 3Cnet is the first recurrent neural network-based pathogenicity prediction tool that evaluates the effects of variants in the context of protein sequences.
As 3billion continues to collect data from its sequencing and analysis of patient samples, Keum also wants to expand beyond diagnostics. "There's a lot of things that we can do with the cumulative data," he said.
One of those things is supporting drug discovery. He said that the company is now building a software platform to help pharmaceutical companies research and identify novel drug targets for rare ailments. The software performs target discovery, drug candidate searches, and safety validation, each based on independent AI models that draw on databases of genomic and phenotypic data from patients with rare diseases.
Keum said that 3billion is talking with several pharma companies about piloting the new technology when it is ready. He did not name any potential partners but acknowledged that 3billion will need to prove the efficacy of its drug discovery platform before it becomes a viable product.
The firm has had some setbacks along the way.
In 2017, 3billion started a pilot in hopes of offering a direct-to-consumer genomic test for rare disease screening to patients in the US who had been unable to obtain a molecular diagnosis elsewhere. A closed beta test involving 100 patients provided whole-exome sequencing and annotation of variants linked to more than 4,000 rare diseases.
"We got some feedback from the testers," Keum said. "They wanted diagnostics, not the screening or knowledge about their genomes."
Plus, 3billion was unable to overcome some regulatory issues related to offering DTC genetic testing in the US.
"We decided it is not the right time to target the patient directly," Keum said. "We focus on the diagnostic tests, and now we are providing all services through medical doctors."