MORGANTOWN – GATC West Virginia played a role in a groundbreaking AI-powered analysis of a new liver cancer drug in development – an analysis that saved inestimable hours and dollars for the client.
GATC West Virginia, headquartered in the WVU Innovation Corp. building, is a subsidiary of Irvine, Calif.-based GATC Health and collaborated with GAC teams in California and Utah to conduct the analysis for A28 Therapeutics, a California biopharmaceutical company focused on the development of cancer treatments.
GATC analyzed A28’s AT-101 cancer drug candidate. We talked with GATC Chief Business Officer Ty Lam and A28 Therapeutics founder and CEO Stanley Lewis about the job and what it means.
As we’ve reported before, GATC is streamlining the creation of pharmaceuticals using it artificial intelligence Multiomics Advanced Technology (MAT) platform, which simulates human physiology and biochemistry, and can predict what is a safe drug, what is an efficacious drug, what drug doesn’t have off-target side effects even before it gets into the lab.
Lam and Lewis – an internal medicine physician by training with 25 years of experience in clinical research, drug development, patient care and teaching – knew each other before the GATC days, they said. Lewis serves on GATC’s advisory board.
Lewis’s prior employer developed a drug, AT-100,that had the potential to treat cancer. “It was a very fascinating drug to me,” he said. But that company didn’t want to pursue the cancer angle, so Lewis formed A28, bought AT-100 and went to work on it for cancer treatment.
AT-101 is an intravenous infusion administered twice a week for about two hours, Lewis said. Trials proceed in three phases. Phase 1 tests the safety of the drug. Phase 2 focuses on patients who have the disease, to determine dosing and timing and early signs of efficacy – called proof of concept. Phase 3 is the true measure of whether it works.
The GATC analysis, he said, was later than Phase 1 and into the first part of phase 2.
The arrangement came about, Lam said, because they knew each other and the MAT AI platform’s ability to run a clinical trial “in silico,” meaning in a computer. Lam told Lewis, “Why don’t we collaborate and let A28 be the first client to ever use our platform for this type of predictive analysis.”
Lewis liked the idea. “This would be really cool to be able to apply this type of technology to the drug I was developing and understand it in even in more detail than what had been discovered during the [initial] clinical trial,” to understand its potential – side effects, other indications – and ways to use the AT-101 molecule.
The whole process took about two weeks, Lam said. They met together to determine what Lewis wanted to learn from the report, and what kind of data he had available to feed into the AI to run the analysis.
GATC is a virtual company, Lam said, and teams from Morgantown and other cities participated. When everything was plugged in, the actual AI analysis took less than three hours.
We asked Lam how long it would have taken without the AI. He said, “It couldn’t be done.” It would have required clinical trials – with all the time and expense and people and trial-and-error that requires. From a technology standpoint, there is no other way to come up with this level of detailed analysis.
Lewis characterized the analysis as the equivalent of an army of PhD’s pooling their talents. It was helpful to learn things that weren’t obvious that might help them tweak the developmental protocol or timing or dosage.
“If this can give you a little bit more confidence, de-risk your program, you can save a lot of time, you can save a lot of money by learning from this,” instead of trial and error. “We’re excited to introduce this new AI-enabled approach to drug development. We think it’s going to improve the efficacy in the way we develop drugs and help us de-risk and avoid some pitfalls in this process.”
Lewis said that roughly 90% of drugs that go into development fail. All the mistakes cost money. “By the time you get one that works you spend a lot of money on others that fell by the wayside. This is one of the ways that I think AI can really really help.”
It avoids all the expensive potholes that and those costs that get transmitted into the price of the drug that emerges, he said.
In the case of AT-101, GATC’s analysis generated predictions for such things as he way the drug moves through the body and the way it stays in the body after it’s administered; bioaccumulation, meaning whee the drug settles (they’re looking at treating liver cancer and AT-101 settles in the liver and persists there for three days); ADME considerations, meaning absorption, distribution, metabolism and excretion; and an important one, endpoint matching.
They explained that testing a drug requires establishing a clinical endpoint – a lab measure of whether the drug improves or fails to improve the disease. An endpoint has to be set before entering a clinical trial.
With endpoint matching you can take the endpoint, look at what the drug does, and the AI can help confirm the endpoint is the right thing to look at based on how the drug works. The AI might reveal a better endpoint.
AT-101 had already been tested on 85 patients and demonstrated tolerability and “remarkable anti-cancer activity.” From here, Lewis said, AT-101 heads into getting the dose optimized and proof of concept in patients with liver cancer.
Data from the clinical trial will be available to GATC, he said, and A28 is interested in returning to GATC to learn what else the molecule can be used for. “I’m looking forward to engaging GATC in the future.”
Lam said GATC’s view is that “every pharma client we work with is a client for life.” The MAT AI tool can give them insights for all of their research as they collaborate in discovering new drugs.”
Lam pointed out that the A28 collaboration was only part of what the MAT platform can do. It also does early discovery – to locate and validate targets in the body that need to be affected to treat disease, and it can discover and validate compounds that could be developed to affect the targets.MORGANTOWN – GATC West Virginia played a role in a groundbreaking AI-powered analysis of a new cancer drug in development – an analysis that saved inestimable hours and dollars for the client.
GATC West Virginia, headquartered in the WVU Innovation Corp. building, is a subsidiary of Irvine, Calif.-based GATC Health and collaborated with GAC teams in California and Utah to conduct the analysis for A28 Therapeutics, a California biopharmaceutical company focused on the development of cancer treatments.
GATC analyzed A28’s AT-101 cancer drug candidate. We talked with GATC Chief Business Officer Ty Lam and A28 Therapeutics founder and CEO Stanley Lewis about the job and what it means.
As we’ve reported before, GATC is streamlining the creation of pharmaceuticals using it artificial intelligence Multiomics Advanced Technology (MAT) platform, which simulates human physiology and biochemistry, and can predict what is a safe drug, what is an efficacious drug, what drug doesn’t have off-target side effects even before it gets into the lab.
Lam and Lewis – an internal medicine physician by training with 25 years of experience in clinical research, drug development, patient care and teaching – knew each other before the GATC days, they said. Lewis serves on GATC’s advisory board.
Lewis’s prior employer developed a drug, AT-100,that had the potential to treat cancer. “It was a very fascinating drug to me,” he said. But that company didn’t want to pursue the cancer angle, so Lewis formed A28, bought AT-100 and went to work on it for cancer treatment.
AT-101 is an intravenous infusion administered twice a week for about two hours, Lewis said. Trials proceed in three phases. Phase 1 tests the safety of the drug. Phase 2 focuses on patients who have the disease, to determine dosing and timing and early signs of efficacy – called proof of concept. Phase 3 is the true measure of whether it works.
The GATC analysis, he said, was later than Phase 1 and into the first part of phase 2.
The arrangement came about, Lam said, because they knew each other and the MAT AI platform’s ability to run a clinical trial “in silico,” meaning in a computer. Lam told Lewis, “Why don’t we collaborate and let A28 be the first client to ever use our platform for this type of predictive analysis.”
Lewis liked the idea. “This would be really cool to be able to apply this type of technology to the drug I was developing and understand it in even in more detail than what had been discovered during the [initial] clinical trial,” to understand its potential – side effects, other indications – and ways to use the AT-101 molecule.
The whole process took about two weeks, Lam said. They met together to determine what Lewis wanted to learn from the report, and what kind of data he had available to feed into the AI to run the analysis.
GATC is a virtual company, Lam said, and teams from Morgantown and other cities participated. When everything was plugged in, the actual AI analysis took less than three hours.
We asked Lam how long it would have taken without the AI. He said, “It couldn’t be done.” It would have required clinical trials – with all the time and expense and people and trial-and-error that requires. From a technology standpoint, there is no other way to come up with this level of detailed analysis.
Lewis characterized the analysis as the equivalent of an army of PhD’s pooling their talents. It was helpful to learn things that weren’t obvious that might help them tweak the developmental protocol or timing or dosage.
“If this can give you a little bit more confidence, de-risk your program, you can save a lot of time, you can save a lot of money by learning from this,” instead of trial and error. “We’re excited to introduce this new AI-enabled approach to drug development. We think it’s going to improve the efficacy in the way we develop drugs and help us de-risk and avoid some pitfalls in this process.”
Lewis said that roughly 90% of drugs that go into development fail. All the mistakes cost money. “By the time you get one that works you spend a lot of money on others that fell by the wayside. This is one of the ways that I think AI can really really help.”
It avoids all the expensive potholes that and those costs that get transmitted into the price of the drug that emerges, he said.
In the case of AT-101, GATC’s analysis generated predictions for such things as he way the drug moves through the body and the way it stays in the body after it’s administered; bioaccumulation, meaning whee the drug settles (they’re looking at treating liver cancer and AT-101 settles in the liver and persists there for three days); ADME considerations, meaning absorption, distribution, metabolism and excretion; and an important one, endpoint matching.
They explained that testing a drug requires establishing a clinical endpoint – a lab measure of whether the drug improves or fails to improve the disease. An endpoint has to be set before entering a clinical trial.
With endpoint matching you can take the endpoint, look at what the drug does, and the AI can help confirm the endpoint is the right thing to look at based on how the drug works. The AI might reveal a better endpoint.
AT-101 had already been tested on 85 patients and demonstrated tolerability and “remarkable anti-cancer activity.” From here, Lewis said, AT-101 heads into getting the dose optimized and proof of concept in patients with liver cancer.
Data from the clinical trial will be available to GATC, he said, and A28 is interested in returning to GATC to learn what else the molecule can be used for. “I’m looking forward to engaging GATC in the future.”
Lam said GATC’s view is that “every pharma client we work with is a client for life.” The MAT AI tool can give them insights for all of their research as they collaborate in discovering new drugs.”
Lam pointed out that the A28 collaboration was only part of what the MAT platform can do. It also does early discovery – to locate and validate targets in the body that need to be affected to treat disease, and it can discover and validate compounds that could be developed to affect the targets.
