Artificial intelligence startup works on technology with diagnostic potential
ViQi, a Goleta artificial intelligence startup that was born out of the COVID-19 pandemic, has received the second phase of National Science Foundation funding to continue the development of machine learning for viral cell infection image recognition.
Kathy Yeung, CEO of ViQi, told the News-Press the technology holds the potential of being diagnostic.
“AIs could be trained to diagnose which viruses are being identified in cells,” said Ms. Yeung.
She talked about the history of the work.
“In April of 2020, the National Science Foundation put out a COVID grant call looking for solutions to help with the pandemic,” she said.“ViQi developed over 10 years at UCSB.
“Our chief science officer, Dr. Ilya Goldberg, ran a lab at the National Institute of Health for over 10 years,” she said. “He proposed to train an AI to detect subtle changes in a cell that is infected with virus before it is humanly visible. In any kind of antiviral or vaccine development, you have to quantify how many virus particles there are.
“If it works, the advantages were you could accelerate the assay to a couple hours,” she said. “In addition, there are hardware imaging devices that could increase the throughput 10-100x.”
ViQi received the first phase of the NIH grant in August 2020. “We have since been able to prove it with every single virus that we have worked on.”
“We have worked with Baylor University and North Carolina State University as well as research organizations in Europe,” Ms. Yeung said. “NSF awarded Phase II of the grant. We are in the process of completing the development of the assay.”
Ms. Yeung explained the difference between first and second phase NSF funding.
First phase funding is dedicated to research and development.
Second phase funding is for commercialization.
“They support small businesses that are innovating. We have a minimum viable product to continue to innovate and develop. We will also go into the marketplace and monetize based on the technology,” said Ms. Yeung.
“We were extremely excited, along with our collaborators, that for each virus we have tried so far, we have been able to train and accurately quantify viral infectivity using this assay,” she said. “Eight virus types tested include DNA, RNA, enveloped, and non-enveloped viruses, as well as viruses that do not reliably result in detectable cytopathic effects, such as HIV and Adenovirus.
This new $1 million grant will allow us to further optimize our cloud-based solution and make it available to any lab working on applications such as vaccines, antivirals, viral clearance testing, viral gene therapy, oncolytic therapies and potentially even diagnostics,” said Dr. Goldberg, the chief science officer.
The continued development of this technology is expected to assist in pandemic preparedness, as well as for research for antivirals and vaccines.
“Viruses are being used as vectors for gene therapy,” Ms. Yeung said. “This opens a large market opportunity for us with viral vector gene therapy. There are subtle changes in the morphology of the cells when they are infected.”