CHICAGO – In the midst of migrating its Colox assay for early colorectal cancer development onto a new next-generation sequencing-based platform, Swiss MDx firm Novigenix last month enlisted Belgian startup BioLizard to develop artificial intelligence to support the product.
For BioLizard, the deal could serve as a proving ground for its business model, which relies on an array of academic bioinformatics expertise to tackle a variety of omics-related informatics needs, including in the molecular diagnostics and therapeutics spaces.
BioLizard, headquartered in Ghent, has only been in business since late last year, but its bioinformatics roots are deep. Much of the braintrust, including Cofounder and de facto CEO Gerben Menschaert, comes from the laboratory for computational genomics and bioinformatics at Ghent University.
BioLizard is backed by Belgian venture fund Novalis Biotech Incubation, a business incubator that specializes in bioinformatics and genetics. The Novalis portfolio also includes Dutch biotech startup Epify and Doc.ai, a US-based maker of a natural-language processing platform for quantified biology, as well as Ghent-based Ohmx.bio, which produces a multi-omics R&D platform. Epify raised €500,000 ($552,000) last month to develop a portfolio of biomarkers to detect colorectal cancer.
Wim Van Criekinge, who heads the bioinformatics and computational genomics labs at the university, cofounded Novalis along with entrepreneur and bioengineer Jan Van den Berghe. The latter is a director Epify, Doc.ai, Ohmx.bio, and BioLizard, according to his LinkedIn profile.
"We have strong links to academia with predictive modeling groups and biostatistics groups, and that's really our stronghold," Menschaert said. "We can build our pipelines or further develop pipelines based on already available academic knowledge."
That is exactly what BioLizard is trying to do for Novigenix by developing an algorithm to support the migration of Colox, an 29-gene mRNA assay,, to the new LITOseek system, which relies on next-generation sequencing rather than RT-PCR. LITOseek, which stands for liquid immunotranscriptomic sequencing, previously had been named Colox+.
The algorithm will also underlie unspecified Novigenix products under development, according to BioLizard.
Novigenix CEO Jan Groen, who joined the company in May, described Colox as a diagnostic workflow that has been implemented in two major commercial laboratories in Switzerland, Unilabs and LMZ Dr. Risch Gruppe, which perform Colox tests in-house. They upload the results to Novigenix's server, and Novigenix technology automatically generates reports.
Data released two weeks ago at the European Society for Medical Oncology conference showed early positive results of this NGS-based testing, Groen said, though the technology has not yet been optimized with machine learning. Now when users upload full immunotranscriptomes, Novigenix can release more than just the results from colon cancer testing because there is more data available.
"Every analysis can be significantly improved if you apply bioinformatics analysis, machine learning, and AI capabilities," Groen said. He called BioLizard and up-and-comer in this sector and the right partner to help Novigenix develop LITOseek, initially for colon cancer.
"We'll have to see [which] tools, be it just predictive modeling tools or more neural network-based or deep learning tools, will be the most appropriate in this situation," BioLizard's Menschaert said. Then the algorithms will need to be tested.
The collaboration with Novigenix is still in the developmental phase, and will be followed by a testing phase that should take at least six months for colon cancer alone, Menschaert said.
BioLizard also has agreements with other, undisclosed academic medical centers to build AI for tests for prostate and bladder cancers, he said, but declined to specify partners.
Kjell Mortier, fund manager at Novalis and interim manager of BioLizard, said that the startup also has some collaborations with biotech companies, but likewise did not specify these partners.
Menschaert said that the company has a few strengths it wants to focus on in bioinformatics, including the emerging fields of multi-omics analysis and microbiome analysis, particularly from next-generation sequencing, though he is not ruling anything out. "We are a bioinformatics company and we can discuss all kinds of collaborations in that field," he said.
Kathryn Morrissey, hired by BioLizard as business development manager a month ago, said that the company has expertise in proteomics, transcriptomics, genomics, and whole-genome sequencing and is looking to harness the combined talents of its people to build "a one-stop shop for omics analysis and machine learning."
The BioLizard name comes from a bit of an inside joke. The group of experts that came together to form the company called themselves "lizards" for their perceived chameleon-like ability to adapt their complementary sets of skills to their surroundings, Menschaert explained. Some are experienced in machine learning, while others are stronger in biostatistics, for example. "By grouping them all together, we feel that we can tackle a lot of problems or issues that are in this genomics bioinformatics world right now," he said.
"We try to keep the pack together … so that we can work towards better solutions together, of course in collaboration with the client and the customers. We don't outsource our people," Menschaert added.
Morrissey said that there are three full-time "lizards" now, plus three part-time consultants. Including support and marketing staff, BioLizard has a team of about a dozen people.
The company currently has offices in Ghent and in Amsterdam. Within a year or two, Menschaert would like to have a presence in the US, the UK, and perhaps other European countries, but the hub will still be in Ghent, with "satellites" elsewhere, he said.
"We see that in life sciences in general, things are becoming more and more data-driven and the bottleneck is more at the site of the data analysis and the learning from the data rather than the data processing itself," Menschaert said. "There is a lot of demand for really highly educated, experienced people in that field that combine bioinformatics knowledge plus the machine learning knowledge in order to derive information from that specific data in life sciences."
Mortier said that BioLizard wants to create custom technology for each client. "using these custom solutions, we grow smarter and smarter and develop tools ourselves that can be used to help the healthcare community in general."
Menschaert said that BioLizard is open to collaborating with a wide range of companies and institutions in diagnostics, RNA therapy, or other biotech areas.
He expects that the knowledgebase BioLizard is building based on its projects will become one of the company's most prized assets. "The knowledge pool we will have at one point will be very valuable," Menschaert predicted.
In some ways, BioLizard is similar to the Hyve, a Dutch open-source software development company that acts as a talent hub as well. The Hyve, however, has chosen to concentrate on building technology to support the Cancer Genome Atlas and has been around since 2011.