The startup using AI to help rare disease families develop custom treatments

CNBC Cures is proudly underwritten by Alexion, AstraZeneca Rare Disease, whose support enables our journalism to elevate stories that advance research, innovation and patient care in rare diseases.

Jacalyn Lee first sensed something was off with her daughter, Isla, just before her first birthday.

"She was just under a year old, and nobody believed me, including the pediatrician," she said.

Isla was missing some of the developmental milestones that her sisters had hit at her age. Doctors told Lee that her two other daughters were just advanced, and that she should not be measuring Isla's development against theirs.

But by the time Isla was 15 months, Jacalyn had grown more insistent.

"I said to my husband, because I couldn't go to the pediatrician appointment, 'You do not leave that pediatrician appointment without getting some type of referral for an eval. Something is different, I'm telling you.'"

A few months later, Isla was diagnosed with autism. It wasn't until her parents received the results of a genetic test when Isla was 3 years old that they realized the cause of her struggles. Isla was diagnosed with DEAF1-Associated Neurodevelopmental Disorder, or DAND — an ultra-rare group of genetic conditions characterized by developmental delays, intellectual disability, and autism. Only about 200 patients have been diagnosed with the disease worldwide.

"We were handed this diagnosis, and little else. There was no roadmap, barely any research, no community that I was aware of at the time," Jacalyn said.

For the five working moms, life had already become an unsustainable series of meetings with researchers. Lee said the group needed to find a partner who understood the work that they were doing, and knew how to help move the project forward.

"We were trying to figure out where should we spend money, knowing that we only have so much. But really we wanted this final document that was board-ready, that we could use to fundraise against, that we could use to operationalize against… almost like a pipeline with sequencing of, okay, you need A to get to B to get to C," Lee said. "We still needed to kind of untangle and create a path forward that was easy to understand, easy to digest in terms of what we needed to focus on right now."

The group turned to Nome for help, a startup trying to position itself as a contract research organization, or CRO, dedicated to serving smaller rare disease groups that the pharma industry does not often cater too.

"We act with patient-level urgency and are highly focused on providing a roadmap forward for an underserved part of the market," said Stevie Ringel, Nome's founder and CEO.

In the case of DAND Alliance, Nome provided the group with a 53-page report detailing next steps and outlining everything from animal studies to trial design.

"They helped us figure out what the gaps are, what work streams we need to prioritize, and even potential researchers or vendors that we need to think about engaging with," Lee said.

Lee said Ringel delivered it all within a timeframe, and at a price point, that surprised her.

"I think he got it in terms of the speed in which we wanted to move and how urgent this was for us, the fact that we were so new, we didn't have a lot of funding," Lee said.

Ringel understood the sense of urgency for Isla because, in many ways, her journey mirrors that of his own.

A company built on understanding

As a teenager, Ringel and his sister were both diagnosed with a type of retinal dystrophy caused by a mutation of the KIZ gene — a rare inherited genetic eye disorder that impacts less than 200 patients globally. There are no FDA-approved treatment options.

Ringel launched the Kizuna Foundation, and was able to raise money towards developing an individualized treatment, but the work required to get there, coupled with a lack of certainty around the process, was taxing.

Ringel designed Nome with that in mind. He said he built the company to be exactly what his own foundation needed — a service for small patient groups looking to develop personalized treatments. One that provides guidance after funds are raised, but before significant capital is deployed.

"As a patient, it would have made a world of difference to me to hear, you know, actually there's something we can do. It's gonna be long and hard and potentially expensive, but there's a pathway here that's credible, here's what it is, and here's how you can take action — versus, sorry, go learn Braille," Ringel said.

That's what inspired Ringel to launch Nome, which he calls a "white glove service" for developing treatments for underserved patient populations.

"People come to us one family at a time, or as a collection of families."

The goal is to bring down the cost of individualized therapies and to make them more attainable for smaller patient groups — maybe even one day providing a path for insurance to pay for them.

Here's how it works. After a patient receives a genetic test that identifies a genetic disorder, they can upload the test results into Nome's system. Nome uses an AI platform that it built — one that it says is more accurate than readily available consumer models — to search for potential treatment options. Nome then provides a free detailed report on the findings.

Ringel sees Nome as providing insight after a diagnostic test. He wants to work in tandem with clinicians and genetic counselors — identifying pathways toward a treatment shortly after a genetic disorder is identified.

"We've built an AI system that can do that analysis in about 10 minutes. And we have a PhD in the loop on top of that before we send anything back to a patient. But it really allows us to give back that answer in a few minutes versus dozens of hours."

Ringel said Nome currently does between 80 and 100 of these reports per month. Of the roughly 5,000 cases Nome has looked at to date, Ringel said the company has identified a programmable medicine or an existing custom medical therapy that fits the known mutation about 25% of the time.

Nome makes money when customers hire them to help design clinical trials for those treatment options, or to act as a project manager to usher the drug to a point where it can be built and delivered to the patients that need them.

"Once we tell you that it's possible, you contract with us to do all the actual hard work and heavy lifting, like a concierge drug developer working for you," Ringel said. "We don't want to send people off into the wilderness. We want them to say, 'no, please help me take this all the way through to dosing.'"

Nome, which oversees a little more than 10 genetic medicine programs, uses AI models to help design trials and manage logistics of those program operations. Ringel said about 25% of the work done at Nome is automated using its AI systems. He believes that as the company's AI agents train and get more advanced over the next one to two years, Nome can automate 60%-80% of the process.

In personalized medicine, where human capital can account for so much of a drug's development costs, Ringel believes a better use of AI will be key in making these kinds of treatments more accessible.

"We know that a customized ASO is between $1.2 and $1.4 million," Ringel said. "If we can bring the cost down by 50%, we hope we can create a pathway where insurance might even start to pay for them."

He compared it to CAR-T cancer therapies. Once seen as too expensive for insurance to pay for, approved CAR-T therapies are now covered by most public and private insurance plans.

Connecting patients to science

Ringel acknowledges that Nome is very early to all of this, and that there's still a long way to go before the company is able to fully accomplish what he envisions for it. To skeptics, who say Nome provides an added layer of cost to an already expensive process, Ringel says he's just trying to provide a service to patient groups who desperately need it.

"I don't know how you can expect these families to move this process forward without someone playing quarterback," Ringel said. "We think we'll ultimately save these groups money, and we'll save them time."

He considers Nome a logistics play. "We call it drug development operations," he said. "We're just trying to connect great science to the people that need it."

When it comes to the science, Nome turns to the more than 80 partners that it works with — groups like La Jolla Labs and Dyno Therapeutics — who have expertise in drug development.

For Jacalyn Lee, the promise of scientific development is what initially gave her hope following Isla's diagnosis.

"I really believe that we are in this renaissance era of AI and gene therapy and patient advocacy," Lee said.

"A nurse at the doctor's office told us, 'the slow curve of progress is now a vertical line.' So I really try to hold onto that."

And Lee hopes Nome could potentially offer a shorter path towards the progress that she desperately clings to.

"In a really acute time, where you just get this devastating diagnosis, and there's no roadmap, there's no cure that you know of, and there's barely any research out there, it's very isolating. And so, if you can have a service that will run through the possibilities...it can give you options."