July 11, 2024
Member Spotlight: Yatiri Bio

Yatiri Bio Leverages Proteomics, Bioinformatics and Precision Medicine to Transform Cancer Drug Discovery
Pilgrim Jackson is on a journey to address the gap between how cancer drugs are discovered and developed and their clinical success by harnessing the power of proteomics, bioinformatics, and AI and machine learning through his company Yatiri Bio. He says Yatiri Bio’s ProteoModels, a portfolio of cellular models that are tailored to match molecularly defined cancer subtypes, can be used to guide better patient matches for clinical trials and explore more options for drug discovery to advance precision medicine in oncology. The company is currently making advancements in AML (Acute Myeloid Leukemia) and ovarian cancer.
Pilgrim, who says he is “a motivated drug hunter,” received his Ph.D. in biophysical chemistry at UC Santa Cruz and had a long career at Celgene, where his research in drug discovery delivered six therapeutics to the clinic and four patents, including Revlimid (lenalidomide), a medication for multiple myeloma and myelodysplastic syndromes. In just the few short years since he blazed his own trail and founded Yatiri Bio in 2020, the company has been a finalist at the San Diego Angel Conference, was named one of Connect’s Cool Companies in 2023, and announced a collaboration with Fred Hutch Cancer Center and other organizations to discover new therapies for AML.
We spoke with Pilgrim about his career path, why the recent resurgence of proteomics makes this an exciting time in biotech and how the company’s technology is addressing current inefficiencies in cancer drug discovery.
How did you get into the life science field?
From a very young age, I knew I wanted to be a scientist. I was raised on a sailboat and early on I wanted to be a marine biologist. I have a deep love of the ocean and I still sail, although since I launched a startup I have less time to do so! [laughs]
As I went through college, I realized that I liked hypothesis-driven science as opposed to observational science. I really got interested in the chemistry and biochemistry side and narrowed in on biochemistry and biophysics as where I wanted to apply myself.
Why did you decide to start your own company?
I like filling a gap. That’s the part of entrepreneurship that I enjoy—solving problems. Celgene, when I joined, was a fabulous place to work. We had big pharma revenue and a biotech spirit, and we pushed some really good science forward. As the organization grew, things started to move more slowly. Then BMS bought us and that was a huge merger. I’ve been through big mergers before, and I just knew it was going to be years before early research was contributing again—large companies tend to focus on marketed drugs and then on clinical compounds, and then eventually look at what is going on at early research. After staying there for a year, I decided it was time to leave.
Yatiri Bio was founded to address current inefficiencies that exist in cancer drug discovery and development, particularly that patients cannot be treated homogenously. Can you provide us with an overview of how you are addressing this?
In most cases, you spend nearly a decade getting a drug all the way to the clinic. And in the case of oncology, 95% of those that make it to clinic fail to get approval—that’s an abysmal success rate. That’s where the $2 billion to $3 billion price tag comes from for drug discovery.
This is where we thought innovation could have the biggest impact, both for drug discovery and for the patients, because lots of these patients are getting therapies that don’t benefit them. If you can separate the patients that do benefit from the ones that are having a hard time, it’s good for the clinical trials, it’s good for the patients on clinical trials and it’s good for the healthcare ecosystem as a whole. We are bringing these empirical models that you can test into a knowledge space, which is unique and key to driving the science forward.
With our ProteoModels, we are doing individual indication by indication. It’s so important that you work on very solid data for any AI machine learning process, that the data going in is high quality. We take patient samples and we run proteomics on those samples, and we create a landscape of the patients from that. And then we use machine learning to position empirical model systems, the proteome models in that landscape. We have proteome models for AML acute myeloid leukemia—that’s the first one that’s completed and we’re commercializing it. Ovarian cancer is very close behind and will be completed middle of this quarter or next quarter.
Why the focus on these two specific types of cancers?
Number one, what’s important for us was addressing the high unmet need—there is a bad survival rate among both of those groups. Second is they’re very heterogeneous diseases. That does apply to quite a few cancers, but in the case of AML, clinicians are accustomed to giving treatment dependent on genomic markers—they’re very familiar with precision oncology as a concept and treatment paradigm, so it’s not hard for them to go to another level that is more detailed and accurate for their actual treatments—that is why AML was chosen. Ovarian cancer is similar, it has differences based on genomics that decide whether they get the new PARP (poly-adenosine ribose polymerase) inhibitors versus the older chemotherapy. Clinicians are accustomed to this idea of precision oncology.
There is quite a bit of development in the industry around both of those [cancers]. There are a lot of compounds coming through clinic right now, which is perfect for us. Our clients are pharma-biotech, and we want them to have compounds ready to go into our models so that we can position them in the right patient subsets. We want what’s called a “crowded space,” so that we have a good group of potential clients. We are also excited to extend into additional indications, including head and neck, colorectal, lung and others.
Proteomics has become popular again in recent years. How does proteomics, specifically, fill the gap in oncology drug development?
Transcriptomics is now more available. Illumina and others, such as PacBio, have done a great job of that. There’s still a big gap between RNA level and protein level. The median Spearman correlation between the two across different diseases is 0.23 to 0.43—there’s a huge number of proteins that are not correlated to the RNA and proteins are the ones that do the real work.
If you think of it, RNA would be the blueprint of a house. Proteins are the actual structure. Cancer is a problem in the actual structure—your proteins are a much better measurement of that. In addition, proteins are the targets of almost all the drugs that are out there. Looking at the proteins makes a lot more sense than looking at the transcriptomics. The tools to look at the proteins have become more available and capable in the last decade.
There’s been an uptake in proteomics as a whole. Now, it’s about how to put proteomics and patient data and model systems together. Some of the AI tools and bioinformatics tools that are out there have pushed that to a new level—our company is half bioinformatician. It’s going to be an exciting decade for all of it coming together. I think we can do something good for medicine and drug discovery.
There’s been an uptake in proteomics as a whole. Now, it’s about how to put proteomics and patient data and model systems together. Some of the AI tools and bioinformatics tools that are out there have pushed that to a new level. It’s going to be an exciting decade for all of it coming together. I think we can do something good for medicine and drug discovery.
What other exciting news about the company can you share at this time?
Since the company was founded, the route our revenues are going are significant, especially for a small biotech. We have generated hybrid computational/empirical model systems that can efficiently identify patients subsets that will be more sensitive to therapies of interest. In the process, we’ve developed this amazing data set that is linked to clinical metadata. We found some interesting novel targets that we’re going to out-license.
We’re going to initiate a clinical trial with our technology as an interventional decision-maker on what therapeutic a patient should get. It’ll be the first of its kind with Fred Hutch Cancer Center and will be launched probably at the end of this year, so that’s very exciting.
We’re also adding collaborations with MD Anderson, Sharp HealthCare, Scripps Research and the Japanese Foundation for Cancer Research. We received an SBIR grant for AML work to expand that knowledge base and we’re also moving into the organoid side with studying ovarian tumors.
What is your biggest challenge right now?
Fundraising is definitely a significant one—2023 was not the kindest year to biotech as a whole and to startups, and there was a bit of a defensive time for the VCs. We want to scale to a larger number of indications, and that really boils down to the funding side.
Our other big challenges have been getting quality data. I think we’ve overcome that significantly, but the quality metadata that’s associated with patient samples—so that when we run the proteomics, we can cross the two—that’s a definite challenge.
What advice do you have for a young person who wants to work in the life science industry? And what advice do you have for someone in the industry who wants to start their own biotech company and become a CEO?
For young scientists joining the biotech community as a whole, it would be to stay true to the science. Don’t be afraid to think outside of the box. We don’t know as much as we think we know, so follow the data. Be passionate about whatever project you’re on—try and get on a project where you’re actually passionate about the end goal.
For those thinking about becoming a CEO: surround yourself with a really, really strong team—not just scientifically and as far as their stature—but for trust and confidence in each other.
