Los Angeles' COVID-19 Response

'As technologies such as AI and machine learning continue to make an impact on the healthcare space, startups will play an increasing role in such transformation.'

For Life Summit 2020, AI LA brought together four heavy-hitters in the LA bioscience community to discuss a variety of aspects of their response to the pandemic on the panel "LA's COVID-19 Response."

Yan Liu - Associate Professor, Director of the Machine Learning Center, Computer Science Department, University of Southern California

Dr. Liu’s research team at USC focuses on machine learning, data mining and AI for social good, applications in biology, climate science, health, and social media.

Eleazar Eskin - Professor and Chair, Department of Computational Medicine, UCLA

Leveraging tools and concepts from genomics, AI, cryptography to try to transform patient care at UCLA and spread that to other health issues.

Joe Wilson - Managing Director, Undeterred Capital, Los Angeles

Undeterred Capital invests early stage AgTech, Biotech, Clean / Environmental Tech

Seed stage, $100 - $500K ($250K sweet spot)

Joe is also the Co-founder C-19 coalition that aims to create a unified supply chain for help getting ppe to front line workers.

Jo Bhakdi, founder and CEO of Quantgene

Quantgene, Santa Monica CA: Unlocking the Deep Human Genome

Quantgene is developing solutions for liquid biopsy cancer detection and liquid biopsy to get closer to the goal of preventive precision medicine.

Dr. Bhakdi opens with the first topic, which is as pertinent for LA as it is elsewhere: how to address the urgent need to increase testing capacity as quickly as possible? It is often true that it’s necessary to strike a balance between testing speed (i.e. turn-around time) and testing accuracy. However, under the circumstances, with there being so many unknowns about the SARS-Cov2 virus, accuracy is paramount. The best strategy is to use current gold-standard testing methods (in this case that would be RT-PCR) and focus on scaling up the testing and data management infrastructure. While it is certainly possible to innovate to rapidly scale up data flow pipelines and result reporting, diagnostic testing methods have been pretty stable for a century. One should not expect something “magical” to happen such that a rapid, reliable test method might be developed that is better than today’s best. This is what their company is focused on: adopting their existing technology to bring these tests on-line as a service to the community.

Dr. Eskin of UCLA confronted the same situation (the need to rapidly scale up testing capacity with limited time), but with different resources available to bring to bear, as well as a slightly different mission due to the fact that profit is not one of UCLA’s primary missions. In his case, because social distancing restrictions essentially halted most laboratory research, all of a sudden his organization had a good deal of idle sequencing capacity available. Using a type of unique “molecular barcode” technology, they can effectively combine tens of thousands of samples into a single sequencing run supported by as few as ten people. Because UCLA needs to provide sufficient testing for their students, they plan to roll out this technology there first before expanding access to other populations. Their primary aim now is to find ways to reduce the per test cost down to about $10.

Dr. Liu’s team is combating the pandemic not with chemistry, but with computers, and along several fronts. The first is to support contact tracing using real-time data streams and AI models to develop a Covid-19 exposure prediction / probability distribution for different geographical areas, the second is to combat Covid-19 misinformation on-line. There is strong evidence to suggest that many individuals and organizations actively seek to promote disinformation during times of crisis in order to advance their social, economic, and political agendas. Using AI and other models, Dr. Liu’s team is able track patterns in the spread of certain types of disinformation and to develop and target responses to push back and ensure that factually correct information also competes for the public’s attention.

Another important focus for Dr. Liu’s team is how to retain the very personal demographic, geographic, and location information that the models require to reach meaningful statistical predictions of where outbreaks may likely to occur, while protecting the privacy of those people.

Their efforts and models they produce may be very useful to policymakers and even entrepreneurs and business leaders as it may enable them to anticipate areas of greatest need in advance and respond in time to anticipated demand potentially avoiding shortages that would exacerbate a crisis.

Early in the outbreak, all four of the panelists’ focus naturally pivoted to combating the pandemic.

Dr. Bhakdi from Quantgene witnessed many interesting developments in the early days of the outbreak; for instance, that individual actions could make outsized impacts by ensuring supply of PPE to front line workers. As the months wore on, entrepreneurs were able to design and begin to implement whole new business models to address emerging needs. Of course, some of the opportunities would be “spot solutions”--that is, they would likely recede along with Covid-19 case loads, and thus would unlikely become whole new companies.

This led to some surprising realizations, from Dr. Bhakdi’s perspective as an entrepreneur. He explained this in terms of three general phases that a diagnostic company goes through as it matures:1) research and development (ensuring your test works); 2) securing FDA approval for the test; and 3) scaling up operations, sales, and marketing. Prior to the pandemic, his team was strictly focused on making sure their liquid biopsy cancer diagnostics were as successful as possible. Once the pandemic had spread and they realized there could be huge demand for their technology as is as a Covid test, they realized that they had been strictly focused on R&D, whereas to scale up testing to the levels needed forced them to develop their supply and production chains, get their test working where it was needed, and ensure they had a sufficiently robust IT infrastructure to collect and deploy the results correctly. The pivot required them to rapidly mature as a company less focused on research and more focused on delivering solutions to customers at scale.

That’s a difference in UCLA’s response: because they’re a non-profit, they don’t need to think like a business. So their response can address a different sector of need, if you will, in the pandemic. It’s obviously good to have different institutional models out there. Not always a market-based solution (says KN).

“Crisis is a time of generation.”

The panelists think that the healthcare system has been fundamentally changed, without going into details. Many changes are here to stay. Do we even know which? Telemedicine? Conducting clinical trials? For now, at least, one concrete thing about 2020 is how it’s renewed interest in diagnostics, given that the early movers in this space managed to capture some market share. Diagnostics are often avoided by investors because of their low margins and complicity. This may change, or at least there may be more willingness to consider investing in innovative approaches to diagnostics and public health.

About THE author

Carina Grunberg

Carina Grunberg is a data analyst with a background in engineering, optics, and biophysics research. She holds bachelor’s degrees from UC Berkeley in Physics and Biochemistry. After working as an engineer for many years, she transitioned to focusing on data science, AI, and ML with a particular passion for the healthcare space. She’s always looking for ways to apply her experience to her wide array of interests and curiosities. When not working, she’s either teaching yoga, exploring the outdoors, or leveling up her Khajit in Skyrim.

Are you ready?

Get Early Access to Life Summit 2021!