Amongst the regional pavilions and orchestrated and informal networking opportunities at BIO 2019 last week were an abundance of entrepreneurial scientists and companies showcasing innovative technologies poised to disrupt biopharma research, development, and manufacturing as well as personalized medicine.
One did not have to be in the Philadelphia convention center too long to hear about the power of artificial intelligence and machine learning to speed drug discovery and reduce costs and the potential of organ-on-a-chip technology to replace animal testing.
Computer-assisted discovery
The use of AI and ML in drug discovery and development will surely not live up to the considerable expectations in the near-term, but, already, ML is enabling complex analyses that were unfathomable just a few years ago and the long-term impact looks promising.
Evan Feinberg, founder and CEO of Genesis Therapeutics, who participated in a panel discussion on how AI is impacting biopharma R&D, also spoke to Biocompare about AI. Feinberg was quick to point out that AI is “not one monolithic technology, it is a whole category of tools.” One of those tools, feature learning, is at the heart of Genesis’s AI platform PotentialNet, which Feinberg invented as a graduate student at Stanford.
PotentialNet and other elements of the Genesis deep learning platform are intended to better inform the decisions of medicinal chemists by proposing modifications to chemical scaffolds and also enabling better prioritization of those proposed chemical modifications by applying PotentialNet as a more accurate predictor of key molecular properties. Feinberg does not see PotentialNet replacing chemists; instead, he said that his company wanted to “supercharge chemists with better tools to make them more productive.”
Echoing Feinberg’s assurance that his company wasn’t looking to replace human scientists was Gabriel Musso, CSO at BioSymetrics. The company’s Augusta framework performs predictive analytics and machine learning in target discovery and computational biology. According to Musso, BioSymetrics wants to make “data scientists more efficient by helping to identify molecules more likely to succeed.”
Augusta, which lets users combine data from multiple sources, including phenotypic, imaging, chemical structures, and genomic data sets, is reportedly easy to deploy and enables faster pre-processing and model interrogation.
Aperiomics is capitalizing on the power of AI in a different way. The company has compiled the world’s most comprehensive database of pathogens—with the genetic information of every known microorganism. In order to make sense of the sequencing information that it has collected as well as manage its considerable database of over 37,000 bacteria, fungi, viruses, and parasites, it uses analytical algorithms and AI to put context around the information it provides. Using any DNA-sequenced organic sample, the company is able to identify the genetic material of any pathogen present.
Aperiomics’s service is targeted at clinicians and their patients facing difficult-to-diagnose infectious diseases. Thus far, the company has worked with more than 1,300 patients to help their medical professionals identify and treat conditions including chronic infections, blood sepsis, and autoimmune disorders. The company has also worked with customers in the research and biotech markets.
Another company touting its use of artificial intelligence is LabTwin, which offers a voice-powered digital lab assistant. The company’s co-founder and CEO Magdalena Paluch demonstrated the technology, and it was quickly obvious how it enables faster and more reliable documentation of experimentation as well as general lab management. With LabTwin's app, researchers can record actions as they are performed, thus avoiding trying to remember precisely what you did after the fact or deciphering hastily scribbled notes. LabTwin, doesn’t just record and transcribe, it uses machine learning to guide researchers through interactive protocols, offer feedback and recommendations on actions, as well as provide on-demand access to scientific data. Although it is a “light ELN” it also connects to other platforms to facilitate use with established ELN systems.
Organs (and humans) on-a-chip
Organ-on-a-chip and microphysiological systems are another area that has received incredible hype recently. Lowry Curley, CEO of AxoSim, was one of several panelists addressing the ballyhoo in a BIO 2019 panel entitled “Humans–in-a-Dish: Hope or Hype"; NIH and FDA representatives also sat on the panel that included discussion of expectations, limitations, as well as what could realistically be expected in the next three to five years.
AxoSim spun out of Tulane in 2014 and has been breaking ground in the organ-on-a-chip space ever since, first with its Nerve-on-a-Chip® technology, which Curley said was the first all-human model of myelination, as well as its recent licensing of Mini-Brain technology from Johns Hopkins University. The addition of the Mini-Brain platform allows AxoSim to address neurological applications across both the central and peripheral nervous systems and makes the company unique in that it offers the only two commercially available platforms with 3D human-relevant myelination.
The Mini-Brain platform is composed of tiny brain-like organoids created using human induced pluripotent stem cells. The cells are stimulated to grow into brain cells, which are engineered to reproducibly form tiny, uniform spheres. After several months, the Mini-Brains develop different types of neurons and support cells as well as myelin. The Mini-Brain components interact with each other and with their environment, and they can be replicated on a large-scale. Mini-Brains are being designed to study Alzheimer's disease, Parkinson's disease, multiple sclerosis, and other neurodegenerative, functional, trauma-related, and infectious disorders.
Hesperos is one of about a half dozen companies in the human-on-a-chip or interconnected multi-organ systems space. Using its pumpless platform with serum-free cellular medium, its fully human systems have the potential to disrupt toxicology testing as well as efficacy evaluation in drug discovery, according to chief scientist James Hickman, who believes the platform will be used as a rare disease avatar and as an alternative to animal models. Hesperos can custom design its systems by allowing customers to choose which organs to use in the model, or even building the model from organ tissues from a specific patient, which is especially appropriate for rare disease efficacy models.
In a study published earlier this year, a Hesperos system with interlinking modules containing human-derived heart, liver, skeletal muscle, and nervous system cells was able to maintain cellular viability and record cellular function in real-time for 28 days. With the technology advancing rapidly, Hickman envisions a day when companies could bypass preclinical testing in animal models on the road to drug approval.
Image courtesy of Hesperos.
The companies described in this article are just a handful of the hundreds of innovative companies pitching their wares at BIO 2019. Who knows where all these nascent companies will be in five years, but it is safe to say that the technologies they are espousing are here to stay and will likely revolutionize the biomedical industry.