Vanguards of Healthcare by Bloomberg Intelligence

ProFound CEO John Lepore discusses AI-driven dark proteome mapping and novel drug targets

57 min Listen to the original Episode page
BiotechArtificial IntelligenceOncologyDrug DiscoveryGenomicsProteomicsVenture CapitalPharma Partnerships

Summary

John Lepore, CEO of ProFound Therapeutics, explains how the company uses AI to identify thousands of previously unknown proteins, expanding the druggable proteome beyond the ~600 targets identified by the Human Genome Project. He highlights challenges in rodent modeling due to evolutionary divergence and advocates for human-relevant systems like iPSCs. ProFound leverages this data for oncology programs using ADCs and TCEs, and has secured a $35M Gates Foundation grant for women's health research. The company also discusses strategic partnerships with major pharma and the use of human genetics to prioritize targets.

Key takeaways

  1. 1
    Lepore claims only ~600 of the ~20,000 proteins identified by the Human Genome Project have been successfully drugged, suggesting a massive untapped target space for new medicines.

    Unidentified speaker (speaker_3) · 9:39

Episode sections

  1. 0:00 – 1:57

    Sponsor Read: Vanguard and Samsung

    Sponsor read for Vanguard institutional quality bond funds and Samsung Galaxy S26 Ultra on Bell's 5G network.

  2. 1:57 – 9:06

    Introduction and Speaker Background

    Host Sam Fazeli introduces guest John Lepore, CEO of ProFound Therapeutics, who discusses his transition from academia to GSK and then to ProFound.

  3. 9:06 – 14:40

    The Dark Proteome and Human Genome Project Limitations

    Lepore explains that restrictive criteria during the HGP missed many functional proteins. ProFound uses ribosome profiling and mass spec to identify thousands of additional, often smaller, proteins.

  4. 14:40 – 19:29

    Conservation and Disease Modeling Challenges

    Newly identified proteins are highly conserved in primates but less so in rodents. ProFound prioritizes human-relevant systems like iPSCs and organ-on-a-chip models to avoid false positives.

  5. 19:29 – 24:20

    AI Platform and Data Scale

    ProFound has mapped thousands of new proteins across 500+ cell types. The company uses agentic AI to analyze interactions within this expanded dataset, providing a competitive advantage.

  6. 24:20 – 29:35

    Therapeutic Strategy: ADCs and TCEs in Oncology

    ProFound initially focused on oncology using ADCs and TCEs because these modalities rely on surface expression rather than deep biological mechanism knowledge.

  7. 29:35 – 32:41

    Gates Foundation Grant and Women's Health

    ProFound received a $35 million grant from the Gates Foundation in August 2026 to study preeclampsia and eclampsia, leveraging their platform to identify biomarkers and drug targets.

  8. 32:41 – 34:41

    Gates Foundation Collaboration on Pre-eclampsia

    Lepore discusses a collaboration with the Gates Foundation's AI Enabled Cures Group to target pre-eclampsia, leveraging the disease's clear patient definition and accessible placental tissue.

  9. 34:41 – 36:12

    Pipeline Strategy and IND Timelines

    Lepore addresses the company's pipeline structure across oncology, cardiovascular, metabolic, immune, and neuro fields, confirming early molecules exist in oncology.

  10. 36:12 – 38:47

    Genetic Prioritization and GWAS Insights

    The speaker explains using human genetics to prioritize targets, noting strong genetic associations increase the likelihood of positive outcomes by 2-10 fold.

  11. 38:47 – 39:00

    Collaborations with Major Pharma

    ProFound has collaborations with GSK for COPD and IPF, Pfizer for obesity, and Novartis for cardiovascular disease, based on genetic validation of novel proteins.

  12. 39:00 – 41:41

    Neuroscience and Disease Sub-types

    Lepore argues that diseases like Alzheimer's and Parkinson's are molecularly distinct entities rather than single syndromes, having identified dozens of novel, genetically associated proteins for Parkinson's.

  13. 41:41 – 50:43

    Intellectual Property Strategy

    The company protects IP through strong methodological moats, confidentiality of target identities, and rapid composition of matter patents. In some cases, the new protein itself is the drug.

  14. 50:43 – 56:40

    AI Integration in Drug Discovery

    Lepore views AI as an enabler integrated into the existing drug discovery process. He emphasizes the need for curated data sets and notes ProFound's single data lake gives it an advantage.

Claims worth verifying

Only ~600 of the ~20,000 proteins described by the Human Genome Project have been successfully drugged.

Unidentified speaker (speaker_3) ·9:35

How to verify: Review public databases of approved drugs and their targets (e.g., DrugBank, FDA Orange Book) and compare against the UniProt protein database to count drugged targets. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

The Human Genome Project used restrictive criteria for defining genes/proteins, including a minimum length of 100 amino acids and requirement for canonical start/stop codons.

Unidentified speaker (speaker_3) ·12:00

How to verify: Review the original Human Genome Project publications and the current UniProt documentation on gene annotation rules and filtering criteria. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound has mapped thousands of new proteins across more than 500 different human cell types and tissues.

Unidentified speaker (UNKNOWN) ·21:35

How to verify: Check ProFound Therapeutics' website, press releases, or published scientific papers detailing their platform's scope and cell type coverage. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

The newly discovered proteins are highly conserved in primates but less conserved in rodents.

Unidentified speaker (speaker_3) ·15:00

How to verify: Perform a bioinformatics analysis comparing the amino acid sequences of ProFound's identified proteins against primate and rodent genomes using tools like BLAST or UCSC Genome Browser. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound received a $35 million grant from the Gates Foundation in August 2026.

Unidentified speaker (speaker_2) ·29:55

How to verify: Search the Gates Foundation's public grant database or press releases for awards issued in August 2026 to ProFound Therapeutics. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound is developing ADC and TCE programs against novel tumor antigens that are unique to cancer.

Unidentified speaker (speaker_3) ·25:36

How to verify: Review ProFound Therapeutics' pipeline announcements, clinical trial registrations (ClinicalTrials.gov), or patent filings for specific ADC/TCE candidates. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

Novartis had a failed cardiovascular trial (LP) recently.

Unidentified speaker (speaker_2) ·17:42

How to verify: Search ClinicalTrials.gov or Novartis investor presentations for recent cardiovascular trial results and outcomes. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

The Human Genome Project identified approximately 20,000 proteins.

Unidentified speaker (speaker_2) ·18:35

How to verify: Review the final reports of the Human Genome Project or current counts in the UniProt database for annotated human proteins. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound's AI platform uses agentic AI to establish relationships among proteins based on interactions, genetics, and expression.

Unidentified speaker (speaker_3) ·22:10

How to verify: Review ProFound's technical white papers or publications describing their AI methodology and platform capabilities. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Insulin and other important proteins are less than 100 amino acids and would have been excluded by HGP criteria.

Unidentified speaker (speaker_3) ·12:03

How to verify: Check the UniProt entry for human insulin (P01308) to verify its mature protein length and post-translational processing. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

The Gates Foundation has a scientific arm called the AI Enabled Cures Group.

Unidentified speaker (speaker_3) ·32:41

How to verify: Check the Gates Foundation website or press releases for the existence and name of this specific group. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Pre-eclampsia is characterized by hypertension with a specific definition in women late in term, with accessible placental tissue.

Unidentified speaker (speaker_3) ·33:01

How to verify: Verify against medical literature or clinical guidelines for pre-eclampsia diagnostic criteria. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Targets with strong genetic association have a 2-10 fold higher likelihood of positive clinical outcomes.

Unidentified speaker (speaker_3) ·36:51

How to verify: Search for literature from the speaker's group or predecessor group at GSK citing this statistical correlation between genetics and clinical success. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound has collaborations with GSK for COPD and IPF, Pfizer for obesity, and Novartis for cardiovascular disease.

Unidentified speaker (speaker_3) ·38:47

How to verify: Check public press releases or partnership announcements from GSK, Pfizer, and Novartis regarding collaborations with ProFound Therapeutics. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

There are approximately 17 protein targets for Parkinson's disease that every pharma company has.

Unidentified speaker (UNKNOWN) ·41:27

How to verify: Review scientific literature or patent databases for known drug targets in Parkinson's disease to assess the count of commonly targeted proteins. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound has identified dozens of novel proteins with genetic association with Parkinson's disease.

Unidentified speaker (speaker_3) ·50:03

How to verify: Check ProFound Therapeutics' published research or patent filings for claims of novel Parkinson's-associated proteins. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

Large pharmaceutical companies struggle to integrate data from acquired companies and distributed research sites into a single searchable database.

Unidentified speaker (speaker_3) ·52:30

How to verify: Review industry reports or case studies on data integration challenges in large pharmaceutical conglomerates. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Roche expects 50% of its assets to be externally sourced.

Unidentified speaker (UNKNOWN) ·44:00

How to verify: Check Roche's investor presentations or earnings calls for statements regarding their external asset strategy. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Argenx has a T-cell engager (L3) for solid tumors.

Unidentified speaker (UNKNOWN) ·47:08

How to verify: Check Argenx's product portfolio or clinical trial registries for information on their L3 T-cell engager. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

ProFound's platform includes a single data lake containing experimental, in silico, and proteomics data.

Unidentified speaker (speaker_3) ·53:05

How to verify: Review ProFound Therapeutics' technical documentation or white papers describing their data infrastructure. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify the number of druggable proteins by reviewing FDA-approved drug targets and cross-referencing with UniProt or ChEMBL databases.
  • Check ProFound Therapeutics' patent filings at the USPTO or EPO to confirm composition of matter patents for novel proteins.
  • Review the Gates Foundation's public announcements or press releases to verify the $35 million grant amount and scope for pre-eclampsia research.
  • Analyze clinical trial success rates for targets with strong human genetic associations versus those without, using data from sources like Nature Reviews Drug Discovery or company filings.
  • Assess the validity of rodent vs. human-relevant model predictions by reviewing recent high-profile clinical trial failures in cardiovascular or oncology sectors cited by the speaker.
  • Examine ProFound's partnerships with GSK, Pfizer, and Novartis for public disclosures or press releases detailing the specific indications and milestones.
  • Verify the conservation data for the newly identified proteins by comparing primate and rodent genomic databases.
  • Review the methodology of ribosome profiling and mass spectrometry used by ProFound to ensure the identified proteins are genuinely translated and not artifacts.
Full transcriptexpand
0:00
Unidentified speaker (speaker_0): Today's show is brought to you by Vanguard. Advisors, your clients count on bonds for income and stability, not unwanted surprises. That's why Vanguard builds institutional quality bond funds. They're designed to help portfolios remain steady when markets don't. Whether
0:20
Unidentified speaker (speaker_0): index or active, Vanguard aims for consistent bond performance with a low cost edge, which helps. clients earn returns that compound over time. A rigorous approach to risk management holds it all together. The goal? Keeping income instability in and unwelcome surprises
0:39
Unidentified speaker (speaker_0): out. Because fixed income shouldn't feel like a roller coaster. Explore institutional quality bond funds from Vanguard. Learn more at vanguard.com. vanguard.com/slash audio. That's vanguard.com/slash audio. While investing is subject to risk. Vanguard Marketing Corporation
0:59
Unidentified speaker (speaker_0): distributor.
1:00
Unidentified speaker (speaker_1): Oh, group shot.
1:02
Unidentified speaker (speaker_2): I'll take it.
1:03
Unidentified speaker (speaker_1): No, we're good. There's one in every group: the notographer. Best intentions, worst shots. But there's hope. The Samsung Galaxy S26 Ultra on Bell. Super steady videos and perfectly framed shots. with my fan cam and nitography for stellar picks even in low light.
1:18
Unidentified speaker (speaker_2): I want that.
1:19
Unidentified speaker (speaker_1): Yep, postworthy with just a click. All on Bell, Canada's fastest 5G plus network. Visit a Bell store or bell.ca slash Samsung for details. Bell. Connection is everything.
1:57
Unidentified speaker (speaker_0): Welcome to another episode of Bloomberg Intelligence's Vanguards of Healthcare podcast, where we speak with the leaders at the foreign. forefront of change in the healthcare industry. My name is Sam Fazilli. I'm one of the pharma analysts, pharma biotech analysts. I wear several hats at Bloomberg Intelligence, the in-house research arm of Bloomberg.
2:16
Unidentified speaker (speaker_2): Founded in 2019, as usual, for a flagship pioneering company, Uh, Profound Therapeutics is focusing on an untapped area of biology. The dark genome or dark proteome. This has recently become more of a discussion topic. I think John might correct me if I'm wrong, especially within the biota community, especially after
2:36
Unidentified speaker (speaker_2): the success of um the Moderna Cancer vaccine. I'll remind you that's also another flat chip company. Given that the dark proteom might actually be a rich source of antigens or or for for vaccines, if especially there's a difference between tumors and et cetera. So that's an interesting area. I mean, I I don't think profound is is
2:56
Unidentified speaker (speaker_2): particularly looking at vaccines. We can talk about that if we need to, but um the company's zoning in on unknown genetic drivers of disease, um like immune modulators, uh circulative metabolic factors, etc. So today I have the pleasure of speaking with John Lapore, uh CEO of the company. After leading
3:15
Unidentified speaker (speaker_2): GSK's research for uh the research at GSK for a while, John joined Profound. So I'm gonna ask him to just give us a little introduction to himself, how he ended up, where he's ended up, obviously, in terms of the career in general, because people are always interested in how people end up in different walks of life. Just
3:35
Unidentified speaker (speaker_2): tell us a little bit about that and how has the leadership of biotech, in your view, changed over the past couple of decades? Over to you, John. Thank you for joining us.
3:44
Unidentified speaker (speaker_3): Great Sam. Well listen, uh absolutely uh spectacular to uh be here with you and with your listenership and thanks for the opportunity to share our story with the Bloomberg audience. Yeah I'm John Lapore. Um as Sam has introduced I'm uh a CEO partner within the flagship pioneering uh ecosystem. ecosystem, but I'm
4:04
Unidentified speaker (speaker_3): primarily CEO of this company called Profound Therapeutics that I'm just absolutely excited about. And it relates to your question, Sam, about you know, how did I get here? Um, I'm initially trained as a cardiovascular physician scientist. Uh so I went through medical school and postdoctoral training and uh medical
4:23
Unidentified speaker (speaker_3): residency and cardiovascular uh clinical. training. Um and ultimately it was very interested throughout in understanding the science of human disease and the mechanistic underpinnings of human disease and sought in my career to find the
4:43
Unidentified speaker (speaker_3): best pathways to leverage that to make medicines and help people. So I started as an academic scientist. at the University of Pennsylvania. I had a lab there. I was NIH-funded. I was studying something very basic: the molecular mechanisms that regulate cardiovascular development, because they're related to
5:03
Unidentified speaker (speaker_3): the mechanisms that drive cardiovascular disease. And that was really cool. But for me, it was a little bit distant from actually helping patients after all the medical training I had had. had. And um I noticed that at the time many of the leading cardiovascular scientists were moving into the industry,
5:24
Unidentified speaker (speaker_3): most of which I worked with in some way. So Mark Fishman and Jeff Light and and uh others. Um and I thought, well, that's interesting. If if these prominent scientists, And Sean Coughlin, are doing that, um, maybe that's a really cool career. And somebody approached me from GSK and I thought, well, that's interesting. Like, would I really go work for a company
5:44
Unidentified speaker (speaker_3): like that? And 'cause I didn't in those eras, you just didn't do that. You were a scientist. And I I learned more about what the opportunity was. And, you know, as a young scientist,
5:52
Unidentified speaker (UNKNOWN): when was this?
5:53
Unidentified speaker (speaker_2): When was this, John? About when?
5:55
Unidentified speaker (speaker_3): That is a really good question, Sam. Long enough that we'd have to I'd have to calculate. It was about twenty five years ago.
6:02
Unidentified speaker (speaker_2): Okay.
6:03
Unidentified speaker (UNKNOWN): Yeah.
6:03
Unidentified speaker (speaker_2): That's about the time one of my uh my boss, uh Frank Walsh, went to I think he went to s uh Smith Klein at the time
6:12
Unidentified speaker (UNKNOWN): in GSK Smith.
6:13
Unidentified speaker (speaker_2): Yeah,
6:13
Unidentified speaker (UNKNOWN): yeah.
6:14
Unidentified speaker (speaker_3): Yeah. In fact, a mutual uh colleague, friend of ours, uh Perry Neeson was instrumental in uh dragging me to GSK. In fact, was it my first uh hiring manager uh and provided just enormous opportunities. opportunity to learn about the industry. And I was hooked, had a range of roles
6:34
Unidentified speaker (speaker_3): from, you know, just single point accountable person for projects through to when I left, as you've indicated, I was running the research organization in early clinical development across many therapeutic areas and having a blast. And it was outstanding. And what a fabulous company. to work with and for uh GSK. But you know, other
6:54
Unidentified speaker (speaker_3): opportunities get put in front of you. And um someone called me about this opportunity at profound. I said, well that's crazy. Like I don't even think those proteins are real. Like why on earth would I do that?
7:05
Unidentified speaker (UNKNOWN): Uh
7:05
Unidentified speaker (speaker_3): and they called back multiple times and I said, okay, well let me let me take a look at that. Let me understand the science. And it's just a fascinating
7:15
Unidentified speaker (speaker_3): I think, principle, really, that
7:20
Unidentified speaker (speaker_3): science keeps unfolding, and there are opportunities that we don't realize the complexity of human biology. And what I saw in the company was proof that there indeed are many other proteins produced from the human genome
7:38
Unidentified speaker (UNKNOWN): than
7:38
Unidentified speaker (speaker_3): those. that had been characterized and categorized by the Human Genome Project. And I saw some very early evidence that they had biological activities that were consistent with being drug targets.
7:52
Unidentified speaker (UNKNOWN): And
7:53
Unidentified speaker (speaker_3): I thought, wow, that's an opportunity to break open an enormous space and be involved in creating many, many medicines across multiple areas. Uh and I took a leap and here I am. So that's a long-winded answer to how I got here, Sam.
8:08
Unidentified speaker (speaker_2): That's a good answer. It's interesting because the image that just conjured up in my head was human genome essentially pushed us down the road of looking just around where the light shines under
8:21
Unidentified speaker (UNKNOWN): the
8:21
Unidentified speaker (speaker_2): lamppost. And we we ended up I mean we had we had to start somewhere. and that was an immense project and frankly if it hadn't been for the human genome would probably project we wouldn't have been here but this this idea of the expanded human proteome is quite fascinating to me as I said as I'm going
8:41
Unidentified speaker (speaker_2): around particularly after this and and way ahead of it there was a whole there are people who keep talking about the dark genome the dark proteome etc et cetera. so uh has become more of a conversation piece. So and obviously profound seems to be one of the key companies in that space. So wh why do you believe that this expanded
9:00
Unidentified speaker (speaker_2): uh or or we say expanded, but the reality of what it is, it is what it it is there already.
9:06
Unidentified speaker (UNKNOWN): Right. It's our
9:06
Unidentified speaker (speaker_3): expanded understanding. Yeah.
9:08
Unidentified speaker (speaker_2): Right. How do you how do you believe this is going to help really push the um
9:13
Unidentified speaker (UNKNOWN): the
9:13
Unidentified speaker (speaker_2): the frontier of direct drug development?
9:15
Unidentified speaker (speaker_3): Yeah, let me go back a half a step and double down on your point that the Human Genome Project was by any standard um just an enormous leap forward for our scientific understanding of human disease. And it was fundamental for uh enabling essentially a new generation of drug targets. And you know,
9:35
Unidentified speaker (speaker_3): I've seen different estimates, but it's somewhere around 600.
9:39
Unidentified speaker (speaker_3): of the 20,000 or so that have been described by the Human Genome Project have been successfully drugged. Okay. So that's on one hand enormous uh contribution for society. But if there were an equivalent or even greater number of proteins that were underappreciated and that our platform is now
9:59
Unidentified speaker (speaker_3): starting to define, then you could think of a proportional number of additional new medicines, which is measurable in the hundreds, if not thousands. Okay. I'm not saying we've done that. I'm just saying that's the potential, you know, on a comparative basis. So that's what drove me to get in the space and so to see how much of that we could unlock and how
10:19
Unidentified speaker (speaker_3): rapidly.
10:20
Unidentified speaker (UNKNOWN): The
10:21
Unidentified speaker (speaker_3): second point I wanted to make though, uh anchoring on your comments, was that you mentioned like looking under the lamppost. And we should also put that in the context of the lamppost or the tools that were available at the time. So it wasn't like a human genome project did anything wrong or incomplete. It's the tools and
10:40
Unidentified speaker (speaker_3): the understanding that were available at the time enabled identification of about 20,000 genes and proteins.
10:48
Unidentified speaker (UNKNOWN): Now
10:48
Unidentified speaker (speaker_3): there are new tools and new opportunities. and that's what we're leveraging to unlock the rest of the human genome. Okay. And then I think uh your question was somewhere along the lines of why would we believe that this has opportunity for medicines? So a couple of things. Let's start with um proteins are proteins.
11:08
Unidentified speaker (speaker_3): Whether or not we knew about them before or defined define them newly is a little bit irrelevant from a drug discovery point of view. And once we identify bona fide, proven new proteins, they're on the table for evaluation as potential drug targets in the same way as the quote known proteins are. The only
11:27
Unidentified speaker (speaker_3): thing that's new is not there now that they're also known. Okay. So there's no black magic here. And that's why I I sometimes try to get away from the the dark proteome because it it almost
11:37
Unidentified speaker (UNKNOWN): implies it. It just
11:38
Unidentified speaker (speaker_2): sounds good.
11:39
Unidentified speaker (speaker_3): It sounds cool,
11:39
Unidentified speaker (UNKNOWN): but it
11:39
Unidentified speaker (speaker_3): almost sounds
11:40
Unidentified speaker (UNKNOWN): like, well,
11:41
Unidentified speaker (speaker_3): there's something strange or unusual. They they don't defy the laws of nature and biology and physics and everything else. They're just proteins we didn't know were there. And the the catch is that when the Human Genome Project was um done, they used a series of rules to define what would be a gene in a protein. And some of them
12:00
Unidentified speaker (speaker_3): ended up being rather restrictive. So for example.
12:03
Unidentified speaker (speaker_3): the protein encoded had to be at least 100 amino acids. And there's some very important proteins like insulin and others that when processed are like less than that and probably would not have been characterized. They also had to have a very canonical start and stop codon, but it's very well described that translation can occur with
12:23
Unidentified speaker (speaker_3): different start codons and end with different stop codons.
12:27
Unidentified speaker (speaker_3): And then another very interesting rule used was that a newly defined protein had to have some degree of homology to something previously described. Okay. So if you loosen all of those, you get to the possibility that there are many more proteins. And what unlocked this, back to the lamppost and the tools, is
12:47
Unidentified speaker (speaker_3): the proper tools to examine the human. genome to find these. And the unlocking feature, and it's a major part of our platform, is a series of proprietary
13:01
Unidentified speaker (speaker_3): methodologies for examining translation events in cells. So what we're doing is looking at the level of the ribosome and asking which RNAs can we capture in the process of being translated, and therefore we know there's a protein produced, and then we show up by mass spec.
13:21
Unidentified speaker (speaker_3): That's a lot different than saying, can we find the 20,000 previously described? It's an open-ended question, and it's enabled us to find um many thousands of additional proteins than were previously described.
13:34
Unidentified speaker (speaker_2): Right. So thinking through this a little bit, um w when you look at this the the variety of undescribed, previously undescribed proteins, the way that you find them, are they characterizable as all the sorts that we we expect? Are they are they big enough to be kinases and enzymes and
13:54
Unidentified speaker (speaker_2): receptor bound, transmembrane? Membrane, Do the uh all of the above? Yeah.
14:00
Unidentified speaker (speaker_3): They're all good questions. And as I alluded to before, they don't defy any of the uh sort of rules of uh what a protein is. They do tend to be smaller. Um the majority of what we're finding are smaller hundred amino acids uh or fewer, um, but some are bigger as well. But they do have uh a range of features that overlap with the the known
14:20
Unidentified speaker (speaker_3): proteins. For example, there are many that are transmembrane, there are many that have signal peptides indicating that. they're likely secreted. We found proteins that are uh signaling molecules that signal another cell to uh act. We've found ones that are clearly intracellular proteins and involved in some kind of signaling
14:40
Unidentified speaker (speaker_3): cascade. So they have biological activities um similar to the gnome proteins, although generally a little bit smaller. We've also um started to map two different things. One is genetics, which we'll come on because that's key to this whole story. But we've also mapped their conservation in other species. And
15:00
Unidentified speaker (speaker_3): they're very highly conserved broadly in primates. And they're a little bit less conserved in rodents than are the 20,000 gnome proteins. So our view is that these are likely a relatively new evolutionary. addition. And this is important for drug discovery because conservation in in
15:19
Unidentified speaker (speaker_3): monkeys enables toxin drug discovery programs.
15:22
Unidentified speaker (speaker_2): It does. It does um I didn't know that. It does throw up a a potential wrench though, in that a lot of disease models are
15:31
Unidentified speaker (UNKNOWN): uh
15:31
Unidentified speaker (speaker_2): rodents.
15:32
Unidentified speaker (speaker_3): They are.
15:33
Unidentified speaker (speaker_2): And a lot of people do their mice and and and rats. And I'm assuming that if you try to generate these animals, I don't know whether you can genetically engineer them so that they
15:43
Unidentified speaker (UNKNOWN): I
15:43
Unidentified speaker (speaker_2): I don't I tell you that's that's a fascinating topic. Can you talk us through that? How would you build disease models
15:49
Unidentified speaker (UNKNOWN): for
15:49
Unidentified speaker (speaker_2): these?
15:49
Unidentified speaker (speaker_3): We could we could do another whole hour on this, Sam, if you want at some point. Uh because the second half of your sentence is many people your first half was many people use uh rodent models for their disease modeling. The second half of that the sentence is and very often those models are not informative. And um this is a hot topic, as you well know, across the
16:09
Unidentified speaker (speaker_3): industry. Um I think FDA are really thoughtfully now enabling TOX programs that get past that, but also pharmacology um that isn't as reliant on um rodent animal models. And you know, I even when I was working at GSK, as
16:29
Unidentified speaker (speaker_3): other companies are in the known proteome, we were guiding the scientists much more towards human relevant systems to do the target validation and understanding of the clinical potential. So, and we're doing the same thing within profound. I'm not changing course on that, which is does human genetics really
16:49
Unidentified speaker (speaker_3): tell us there's a likely causality for the proteins, biology, and the disease? But also, can we study its biology and pharmacology in human-relevant cells and cell systems? So IPS differentiated cells to specific human uh cellular phenotypes that mimic
17:09
Unidentified speaker (speaker_3): disease, um, multicellular human models that mimic the disease, organs on a chip, these things that are more human relevant. Honestly, Sam, I think provide much more insight than a rodent model that has so many false positives.
17:23
Unidentified speaker (speaker_2): Yeah. Yeah. I mean, we've had one little story recently, not little, big one, Novartis with their LP little A trial for cardiovascular that failed. Um, we don't know exactly why yet, but there was there has been now a little question. thrown up as to how easy is it to
17:42
Unidentified speaker (speaker_2): always rely on human genetics to make a connection to the probability that your intervention will help you succeed as a therapeutic. There's a lot to unpack there. We're not going to go there, but it does make it more complex and more interesting if it ends up being that the genetics were telling you something that.
18:03
Unidentified speaker (speaker_2): you didn't draw correctly, that you should have gone at a different time point in the disease, et cetera. So it would be interesting. Now I'm beginning to get the feel, the urge to start using the phrase artificial intelligence, um, or whatever you want to call it, machine learning, software packages that help you.
18:21
Unidentified speaker (speaker_2): And now one thing I haven't asked you is what's the magnitude we're talking about here? I can't remember what the latest number is. for that known protein atlas. Are we at fifteen thousand, twenty thousand? I don't remember what the number is. But
18:34
Unidentified speaker (UNKNOWN): where are
18:35
Unidentified speaker (speaker_2): we here?
18:35
Unidentified speaker (speaker_3): Yeah, there's about twenty thousand proteins identified that everyone has access to in the databases from originally from Human Genome Project. It's 20,000-ish. Right. Uh Which is one or two percent of the human genome, by the way.
18:48
Unidentified speaker (speaker_2): Right, which is shocking, right? And
18:49
Unidentified speaker (UNKNOWN): always always 'cause i it would
18:50
Unidentified speaker (speaker_2): have told you evolution. has created this setup where we only the rest of it's being made and copied for nothing. Yeah,
18:56
Unidentified speaker (speaker_3): even though the DNA
18:57
Unidentified speaker (UNKNOWN): is
18:57
Unidentified speaker (speaker_3): highly conserved. So that was
18:59
Unidentified speaker (UNKNOWN): always
18:59
Unidentified speaker (speaker_3): a very suspect as the premise
19:00
Unidentified speaker (UNKNOWN): for how and
19:01
Unidentified speaker (speaker_3): if you think about the complexity of human biology and even interaction the way we are, and the fact that that would be controlled by twenty thousand discrete entities, it seems a stretch. And it's not true. There are many more. Um look it it depends very much how you characterize these. and what degree of evidence you have. But I would say there are there are tens of thousands of additional proteins
19:21
Unidentified speaker (speaker_3): beyond those twenty thousand.
19:23
Unidentified speaker (speaker_2): Mm-hmm. Which is why this int artificial AI or sophisticated modeling comes in.
19:29
Unidentified speaker (speaker_3): Exactly.
19:29
Unidentified speaker (speaker_2): Because the data volume sounds like it's going to be enormous.
19:33
Unidentified speaker (speaker_3): Yeah. And so let me tell you a little bit about that. I'm making myself a note because I want to come back to the genetics. because and the and the example you use, because it's it's
19:42
Unidentified speaker (speaker_2): was that provocative?
19:44
Unidentified speaker (speaker_3): Well, yes, and it's highly important for how we're thinking about um characterizing these proteins, not thinking about how we are prioritizing them. Um but on the AI, so so yes, our our company is e has evolved. It's almost like stages, like stage one, and you know, you you kindly uh introduced us as part of the flag. flagship ecosystem.
20:03
Unidentified speaker (speaker_3): And what flagship is, not many of your listeners will already know, but some they might be new to, uh, is a is a venture firm, but we don't invest in other companies. It's a venture creation firm where we use the venture money to create uh companies within flagship. And the principle is to use the money
20:23
Unidentified speaker (speaker_3): to evaluate
20:25
Unidentified speaker (UNKNOWN): an
20:25
Unidentified speaker (speaker_3): area that's new. and it's not adjacent to something known. It's a completely new idea.
20:31
Unidentified speaker (UNKNOWN): And
20:31
Unidentified speaker (speaker_3): we don't think of it as a high-risk idea. We think of it as an idea that it's not known whether the risk is high or low. So what we do is put money into the area and decide, is it a high or low risk? And we start with like a pretty provocative question. And here the question was, are there many more proteins in the human genome?
20:51
Unidentified speaker (speaker_3): And uh there was an initial investment, um, and the answer was a resounding yes.
20:57
Unidentified speaker (UNKNOWN): But
20:57
Unidentified speaker (speaker_3): then it was, well, what do we do with them? And that's how I got in the company, because you need expertise about how do you now convert them into real drug discovery programs. And um where I'm going with this is the first bit of the company was finding the proteins, figuring out how many there are.
21:15
Unidentified speaker (speaker_3): And we we've done just I think a spectacular job. Most of it was before I got here, so I could call it spectacular, um, of finding the proteins. Okay. And we've done this systematically at an industrial scale across now more than 500 different human cell types and tissues. And we've probably got,
21:35
Unidentified speaker (UNKNOWN): you know,
21:35
Unidentified speaker (speaker_3): asymptotically starting to get close to having many or all of them.
21:40
Unidentified speaker (UNKNOWN): So
21:41
Unidentified speaker (speaker_3): now the question isn't can we find them?
21:43
Unidentified speaker (UNKNOWN): The
21:44
Unidentified speaker (speaker_3): question is, what do they do? And you already sort of uh presaged this.
21:49
Unidentified speaker (UNKNOWN): If
21:50
Unidentified speaker (speaker_3): you think about that enormous data set, we uniquely have access to that data set. And what we're doing now is, if you can imagine, what almost every other company is doing is taking their turn. target space and using agentic AI to start to establish relationships among proteins,
22:10
Unidentified speaker (speaker_3): network analyses, what can we infer the protein does based on what it interacts with and its genetics and where it's expressed in. However, almost every other company is limited in that they only have the 20,000 in their model.
22:25
Unidentified speaker (UNKNOWN): Well,
22:25
Unidentified speaker (speaker_3): What we uniquely have is those same 20,000, the same AI model, but also this new data set in there. And that enables us to understand the function or the potential function, not just of our new proteins, but also to inform the
22:42
Unidentified speaker (speaker_3): function of the known proteins because we know how the networks interact. So this is a great um application of AI that's going to be a major part of the how we now sift through these many proteins and bias towards the ones that have really uh strong evidence of disease-like uh properties.
23:03
Unidentified speaker (speaker_2): So, I mean that you've kind of answered the question I was gonna ask is this could be like drinking from a fire hose.
23:09
Unidentified speaker (speaker_3): Yeah.
23:09
Unidentified speaker (speaker_2): Um, you know, I have come across many biotech companies who have a platform technology which can throw off many opportunities and I and deciding what to do is probably your hardest job. Uh I is that fair in terms of what to take on, which what therapeutic area? So
23:29
Unidentified speaker (speaker_2): there has been a a development recently which I'll come back to. But if you wanted to just touch on that, and you did to a degree, but do you intend what is the uh uh the intention here? Developing small molecules against them and their potential interactions, developing antibodies, which of course keeps you to a degree external. Uh d do S I
23:48
Unidentified speaker (speaker_2): RNAs of kind of work here. Uh and of course in the end you have um uh other or whatever way. Would you like to just talk us through w how you see this is gonna evolve?
23:59
Unidentified speaker (speaker_3): Yeah, sure. And that's exactly the the challenge. Um is is how to prioritize that. It's no different than any pharma company has in their own target strategy and you know what's their prioritization. And there's a combination of things uh that you need to consider. In this case, we as we the companies evolved have parsed this out in several different ways. Um
24:20
Unidentified speaker (speaker_3): initially it was where are we most likely going to understand the biology in a disease-related. disease-relevant way and be able to move quickly towards a drug discovery program? Because we want to get our legs under us and really have evidence that there's progress. And that led us to cancer in the first instance. Because if you think
24:39
Unidentified speaker (speaker_3): about
24:41
Unidentified speaker (speaker_3): cancer targeting therapeutics, like antibody drug conjugates, ADCs, or T cell engagers, TCEs,
24:49
Unidentified speaker (UNKNOWN): you
24:49
Unidentified speaker (speaker_3): don't need to know a lot about the bio. biology of the protein.
24:53
Unidentified speaker (UNKNOWN): You
24:53
Unidentified speaker (speaker_3): need to know a lot about the expression of the protein.
24:55
Unidentified speaker (UNKNOWN): And
24:56
Unidentified speaker (speaker_3): if the protein is on the cell surface, highly enriched in cancer, and expressed at a very low level in healthy tissue, then you can create a targeting molecule like an ADC or a TCE against the cancer, kill the cancer cells safely, and never have to know a lot about the bio. biology.
25:15
Unidentified speaker (speaker_2): That
25:16
Unidentified speaker (speaker_3): was a starting point. And we now have several programs that are in relatively advanced stages of making ADC and TCE molecules where it's a novel. We're literally the only person that has it. And if you put it in the search engines, it's not described anywhere. Multiple
25:36
Unidentified speaker (speaker_3): different programs uh where it's a unique. tumor antigen that we're making an ADC or a TCE against. That that was a rapid exemplification of how you can
25:45
Unidentified speaker (UNKNOWN): move quickly
25:45
Unidentified speaker (speaker_3): in the space.
25:46
Unidentified speaker (speaker_2): You use the phrase unique
25:49
Unidentified speaker (speaker_2): really? I.e., how sure are you that you don't get because one of the problems with, you know, we we what we we look at uh you know, I've just come back from Like ESMO, which is the second biggest oncology conference uh globally. They, you know. ADCs are everywhere. B7H3, CES6, uh, TROP2, I mean you you name it.
26:09
Unidentified speaker (speaker_2): But they all have, of course, the part of the issue is is not the antibody or the antigen. Part of the issue is how stable the the drug is on the
26:17
Unidentified speaker (UNKNOWN): antibody, etc.
26:18
Unidentified speaker (speaker_2): etc.
26:18
Unidentified speaker (speaker_3): Yeah.
26:19
Unidentified speaker (speaker_2): But uh you still get some side effects because your your Clodine 18.2 eighteen point two is expressed in the normal gut, for example, and you're going after gastric. cancer. How do you how are you f uh do you find that these are m more unique than than than the antigens that we've had access to so far?
26:37
Unidentified speaker (speaker_3): Yeah. So s several things there. And just so uh let's parse them out. One is um there's uniqueness in the sense that we've newly identified the protein and described it as a cancer antigen. and not the visible to anybody
26:52
Unidentified speaker (UNKNOWN): else.
26:53
Unidentified speaker (speaker_3): That's one set of kind of uniqueness. The other uniqueness is unique in cancer versus healthy tissue. Look, you hit the nail on the head. You named four or five antigens. There's not that many more in cancer. There's maybe 10, 15, something like that, that
27:07
Unidentified speaker (UNKNOWN): have
27:08
Unidentified speaker (speaker_3): been sort of successfully therapeutically or in in the process of being therapeutically evaluated. And so what you see and it's I want to be really clear about this, it's important and valued. It just we're doing something different. You see a lot of companies iterating on modality to go after that same 10 or 15
27:27
Unidentified speaker (speaker_3): targets and improve either potency, selectivity, you know, something. Okay. Very important thing for companies to be doing great for patients. But what we're putting on the table is an additional, an alternative, a complementary approach, which is to start with completely novel antigens and open up new target space.
27:47
Unidentified speaker (speaker_3): Okay. And we we have the luxury of learning from that experience on the predecessor targets and molecules, many of which, as you say, are expressed in healthy tissues. And what we're doing is being very selective about ones that have a much more differential expression between
28:07
Unidentified speaker (speaker_3): cancer and healthy. So we're starting in in a in a different space.
28:11
Unidentified speaker (UNKNOWN): Well,
28:11
Unidentified speaker (speaker_3): you never know to do the clinical experiment, right?
28:13
Unidentified speaker (speaker_2): Right. Of course, of course. Uh the and they they are clearly have to be cell surface antigens. And um given the size, do they end up being relatively because antibodies can do magic things, you know. as develop them, but
28:29
Unidentified speaker (speaker_2): do they end up being relatively easy to find uh an antibody against?
28:35
Unidentified speaker (speaker_3): Um
28:35
Unidentified speaker (UNKNOWN): I I
28:36
Unidentified speaker (speaker_3): don't think my team would appreciate if I said it was easy what they're doing.
28:41
Unidentified speaker (speaker_3): I would say it's equally easy to
28:42
Unidentified speaker (UNKNOWN): do. I said I said
28:43
Unidentified speaker (speaker_2): relatively.
28:44
Unidentified speaker (speaker_3): Yeah, I'd say it's equally easy and equally hard to the known ones. This is an example of what we sort of chatted about at the beginning of the call. which is they're not there's nothing magic about them other than that they're newly defined. So, you know, finding the first HER2 antibody wasn't easy either. And, you know, because you have to develop the reagents, you have to develop the amino histochemistry
29:04
Unidentified speaker (speaker_3): assays, you know, uh so there's there's a development, technical development aspect. Um, but they're equally amenable to making an antibody as the known. protein.
29:15
Unidentified speaker (speaker_2): Mm-hmm. So I I I want to talk about I'm gonna just step out for a minute. I'll come back to your developing pipeline, uh for want of a better phrase. You recently uh did a um got a grant from the Gates Foundation for the use of your platform and and everything else that goes with it, data sets,
29:35
Unidentified speaker (speaker_2): algorithms, etc., to find new drugs against and and and biomarkers, of course, because these could also be biomarkers. That's the that's the beauty of it. And just as HER2 is to a degree for a subset of breast cancers. But pre-eclampsia and eclampsia. Now, this takes us into this was $35 million announced at the back end of
29:55
Unidentified speaker (speaker_2): August. This takes us into a very interesting conversation in that women's health has been a, and I'd love to hear. your take on this as to why, has been a relatively ignored, despite the fact they are, I would say, probably roughly exactly 50% of the global population. In these sort of things, that
30:15
Unidentified speaker (speaker_2): our diseases are probably I mean, nobody would ever consider preclampsia or reclampsia as not a disease or not a condition that should could benefit from treatment. And so first question is why do you think that is? There's other things that happen in in in you know um in women
30:35
Unidentified speaker (speaker_2): specifically that don't happen in men, but it seems to be ignored. So there's that one point. And then secondly, talk us through this particular deal, because it it does sound quite interesting. It almost also tells me that it's Gates Foundation that you still signals that pharma companies are not necessarily the ones to go to go to in that space.
30:52
Unidentified speaker (speaker_3): Yeah, it's lots of overlapping points there, uh, all of which uh you know directionally I think I'm I agree with. So let's parse out. First off, uh, roughly you asked why has women's health lagged as an area of investment in RD. I think it's multifactorial. Part of it is the
31:12
Unidentified speaker (speaker_3): science hasn't been clear what the drug targets are, uh, and we'll get come up on that. And that's that's a critical factor for this preoclampsia uh uh project we're working on with Gates. There's also been some particularly when it relates to pregnancy, some trepidation about doing development plans in pregnant women and risks
31:32
Unidentified speaker (speaker_3): to the both mom and the fetus. I think that's that's driven some of the um issues. And I I'd imagine there's also some market analysis that's been done over the years that, you know, I think it's con confluence of those several factors that have led to it being um under invested in. But the one that's that's I think now emerging is a
31:52
Unidentified speaker (speaker_3): deeper scientific understanding of these diseases. So we could come on endometriosis, which is cool as well. But what we're working on now is preeclampsia, pre eclampsia,
32:01
Unidentified speaker (UNKNOWN): where
32:01
Unidentified speaker (speaker_3): again I'm a cardiovascular physician scientist. I've heard a lot about preclampsia of t you know, taking care of patients. And I got permission to say the following, uh, my wife actually had preocclampsia during the birth of our first daughter. And so, you know, I I'm aware of like the the angst for dad, but it's and the risk for mom and baby. Yeah.
32:21
Unidentified speaker (speaker_3): And, you know, I was frustrated as a cardiovascular person, like, uh really, we don't know any more about this other than let's deliver the baby and it'll go away? Like, that's not exactly a treatment. But you know, there it there wasn't like a lot of insight to the biology. Flash forward this collaboration, um, a scientist you may have had on before, uh, Skip Virgin, who leads um
32:41
Unidentified speaker (speaker_3): a scientific arm of uh the Gates Foundation, is called the AI Enabled Cures Group. And that's v very highly pertinent to the conversation. And he and I were chatting about where our platform might be applicable. And we rapidly got to pre-eclampsia because it has several features. One,
33:01
Unidentified speaker (speaker_3): you know exactly who the patient is. It's a woman who's um late in term. Two, you know exactly what the disease looks like. It's hypertension with a very specific definition. Three, you know a tissue. that's accessible that's essentially donated in every case, which is placenta. And that
33:21
Unidentified speaker (speaker_3): confluence that allowed us to think, well, okay, well, why don't we take placenta from people, women with and out preoclampsia, and examine what are the proteins that are differentially expressed, including this whole new expanded protein space.
33:37
Unidentified speaker (UNKNOWN): And the
33:37
Unidentified speaker (speaker_3): two of us well, got so excited. about that opportunity and, you know, figured out how to get it funded and organized. And that's what we're doing now. And as you say, it enables finding novel drug targets, but also novel biomarkers. But a key part of it is the AI.
33:53
Unidentified speaker (UNKNOWN): So
33:53
Unidentified speaker (speaker_3): Skip and that group are really keen on how to accelerate things. And the Gates Foundation has a lot of, you know, positive uh energy on um. getting rapidly to solutions in the in global health space. And one way to do that rapidly is using an AI to enable your programs. So what we're doing is creating essentially
34:13
Unidentified speaker (speaker_3): an AI agent that is trained on the biology of pre acclamsia that could then be used not only to prioritize the targets in this collaboration.
34:27
Unidentified speaker (speaker_3): but also used by the foundation going forward in their work on preoclampsia. So the very exciting confluence of science, ambition, making a dent in women's health and using AI in the space.
34:40
Unidentified speaker (UNKNOWN): So,
34:41
Unidentified speaker (speaker_2): John, you spoke kind of alluded to obviously to the to the oncology uh work that you're doing. It's Is one way of dealing with this problem of having so much that you could do to create proper. profound one, profound two, profound three, profound four with therapeutic uh specialties? Or is it that just keep going as you are and as the
35:00
Unidentified speaker (speaker_2): pipeline grows, I mean, you know, we've been listening to Moderna, they've been working hard to find continued applications for their mRNA platform, which are there are s you know, vaccine
35:15
Unidentified speaker (speaker_2): aspect of it is the way to go forward there, whether it's in oncology or in infectious diseases, uh or some of these other diseases where you can get expression in a specific uh tissue. So, what what else do you think we're going to see from you? And when would you hope to file the first
35:35
Unidentified speaker (speaker_2): IND for anything?
35:36
Unidentified speaker (speaker_3): Yeah, so this is a good segue into the genetic. genetics, which I wanted to come back to. Yeah. And um you know the the Cardization in cancer was along the lines that I described, cell surface proteins, targetable, straightforward pathway to the clinic. And look, we haven't disclosed exact timelines, but I I think you could infer those things move pr pretty quickly. And I
35:56
Unidentified speaker (speaker_3): told you we've got um uh a very interesting target space and we've got r uh early molecules and you'll be seeing that progress. Okay. In the other areas, cardiovascular metabolic immune neuro is really cool
36:11
Unidentified speaker (UNKNOWN): as
36:11
Unidentified speaker (speaker_2): well.
36:12
Unidentified speaker (speaker_3): Um what we're using uh is prioritization based on human genetics. So if you'll indulge me, I'm gonna do a slight divergence on genetics. So a principle that's been used, and you alluded to the uh LPA um situation and you kind of inferred, well, the genetics didn't pan
36:32
Unidentified speaker (speaker_3): out. But you know, I don't think any of us working in the field would say, well, it's a hundred percent. Like if it's got genetic association, well it's definitely going to work. The point is having human genetic linkages that are strong biases you toward a positive result. It doesn't guarantee one. And you know, there's been really good
36:51
Unidentified speaker (speaker_3): literature on this, much from the group, um, my group and the predecessor group at GSK, on targets that have strong genetic association have somewhere between two and tenfold more likelihood of being a positive outcome. Okay. That doesn't mean 100%, but
37:11
Unidentified speaker (speaker_3): they're biased towards it. So in sifting through this
37:15
Unidentified speaker (speaker_3): large new database of proteins, what we're doing is using genetics to help us bias towards the ones that are going to be more likely to be effective. And if you think about human genetics, it's largely driven by genetic variants or GWAS variants that are associated with a particular protein.
37:35
Unidentified speaker (speaker_3): And then you infer that protein must have causal biopics. biology because there's a mutation that links to, you know, coming the disease. And then you study it and et cetera, et cetera. The challenge is that most GWAS variants exist in segments of DNA where it's not clear what the associated protein is.
37:55
Unidentified speaker (speaker_3): And they've been like a mystery. Well how and then we kind of come up with a story, well, maybe it regulates in trans a protein that's very far away. or a gene that's very far away. And you know, it's sort of a little bit unsatisfactory. It turns out what we're doing is systematically looking at these all the GWAS variants. And in many cases, we're
38:15
Unidentified speaker (speaker_3): finding that there's an open reading frame at that location of the variant. There's
38:20
Unidentified speaker (UNKNOWN): actually
38:20
Unidentified speaker (speaker_3): the new protein we've described.
38:22
Unidentified speaker (speaker_2): Ah, interesting.
38:23
Unidentified speaker (speaker_3): So it's basically a new pool of genetically validated.
38:27
Unidentified speaker (speaker_3): targets that are explaining the old GWAS variant. Everyone knows those GWAS variants. They don't know that there's a protein there. It's really, really cool. So those are the targets that we're prioritizing. And that's been the substrate. That line of thinking has been the substrate for several collaborations we have, both
38:47
Unidentified speaker (speaker_3): uh including with GSK for C COPD O P D and IPF. uh with Pfizer in obesity and with Novartis in cardiovascular disease.
38:56
Unidentified speaker (UNKNOWN): And
38:56
Unidentified speaker (speaker_3): you know, watch the space. We'll be we'll be doing more of those types of collaborations.
39:00
Unidentified speaker (speaker_2): But let's talk about neuro, because I'm I'm fascinated by this. There's also another angle here. I wonder whether that's been part of the conversation at Profound. Is that um one of the I was visiting uh Charité in in in Berlin and there is a uh neurologist there who
39:19
Unidentified speaker (UNKNOWN): uh
39:19
Unidentified speaker (speaker_2): has a B in his bonnet about amyloid plaques being actually directly relevant to Alzheimer's in that he says I can take you downstairs and show you sample after sample of of brain tissue of folks who died never been
39:35
Unidentified speaker (speaker_2): so the opposite almost there's something in those people that's protecting them from disease, assuming there is biology, that that the biology that we have for those sort of things is exactly correct. That that that aminoid is a causative agent. But let's say that's true. So there's something that's protecting those folks. And there are others who who
39:55
Unidentified speaker (speaker_2): can have um cardiovascular markers but never have a a
40:00
Unidentified speaker (UNKNOWN): uh
40:00
Unidentified speaker (speaker_2): heart attack or a disease. And does that also feature in in your thinking and your approach?
40:05
Unidentified speaker (speaker_3): Oh for sure. Yeah. It's the same um it's the same process. It's the same uh thought experiment on how to evaluate genetic data sets, phenotypic analysis coupled with GWAS analysis,
40:22
Unidentified speaker (UNKNOWN): um
40:22
Unidentified speaker (speaker_3): coupled with biomarkers, but being open to the answer not being one of the 20,000 known proteins, but actually one of these newly described proteins. And you know, it's the same storyline like I mentioned with the oncology. It's it's not that we're saying companies that are working on their
40:42
Unidentified speaker (speaker_3): tau-related proteins or you know, known own mechanisms, and they're th doing a really good job in neuroscience evaluating those proteins.
40:51
Unidentified speaker (speaker_3): Super.
40:53
Unidentified speaker (UNKNOWN): But
40:53
Unidentified speaker (speaker_3): they've developed assays, reagents, um, target validation approaches, de-risking strategies for those known targets.
41:02
Unidentified speaker (UNKNOWN): What
41:02
Unidentified speaker (speaker_3): we're putting on the table is they use all of those same tools to now evaluate this new set of targets that has equal genetic validation as the ones they're working on, but they're new. And it's it's a very um sort of uh inspiring opportunity to like not just do what everyone else is doing. Like you could Google the two of us, Kevin, your organization
41:22
Unidentified speaker (speaker_3): will know this in great depth. Like what are the protein what are the targets for Parkinson's disease? You're gonna come
41:27
Unidentified speaker (UNKNOWN): up
41:27
Unidentified speaker (speaker_3): with a list of 17 that every pharma company has. I've not the exact number, but figuratively. It's in that range. What I'm saying is there are those, plus they're now a new panel that we've identified. And they're equally uh important to evaluate.
41:41
Unidentified speaker (speaker_2): What about intellectual property? So there's obviously two ways of doing this. One is you're spending all this energy and effort. I I I mean, how much of this no is know-how in terms of of protecting yourselves?
41:53
Unidentified speaker (UNKNOWN): And
41:53
Unidentified speaker (speaker_2): then of course, once you have a molecule, you erase something to it, you're an antibody to it. Talk me through what you're thinking is uh an IP and protection.
42:01
Unidentified speaker (speaker_3): Yeah, look, I mean this n it's not a huge secret that protecting IP on targets is challenging. So we have an extraordinarily strong position on the methodologies for uh identifying these proteins. It's not just a single tool, it's a very uh complex suite of both experimental and machine
42:20
Unidentified speaker (speaker_3): learning and computational steps that enable us not just to find these proteins, but to characterize them and be confident what's the start point, what's the end point, what's the full sequence, et cetera. So that's a very strong moat for sort of uh protecting uh the space. Um the second
42:40
Unidentified speaker (speaker_3): is um, you know, we keep the identity of these uh confidential.
42:45
Unidentified speaker (speaker_2): Uh-huh.
42:45
Unidentified speaker (speaker_3): And the third is we get composition of matter as you've alluded to uh quickly. And one thing we haven't touched. on here is in in some of the cases, if not many of them, the new protein itself is the drug. If you think about like an insulin and glucagon, you know, without and and they're
43:02
Unidentified speaker (speaker_2): that is patentable.
43:03
Unidentified speaker (speaker_3): Yeah, you essentially start with composition of matter and a method of use and you know, off to the races.
43:08
Unidentified speaker (speaker_2): Okay. Interesting. So uh obviously we have a um we have a world which is becoming quite complex with regards to its ability to pay for our medicines and therefore the pharma companies want to be best in class first in class. This seems to me to be a an
43:28
Unidentified speaker (speaker_2): excellent route to get
43:31
Unidentified speaker (speaker_2): potentially best in class, but definitely first in class agents, especially if you keep them secrets. So you have a few deals, Pfizer and Novartis, GSK. What how much more room do you have to collaborate? And how do those deals uh work in terms of the future development? Are they uh early stage partnerships? Tell us a little bit about that
43:51
Unidentified speaker (speaker_2): and what your ambitions are. And how much does that actually tell us about the um um innovation that's going on within Pharma? You know, we just heard Roche tell
44:00
Unidentified speaker (UNKNOWN): us
44:00
Unidentified speaker (speaker_2): that I expect 50% fifty percent of their assets to be expensive. external.
44:03
Unidentified speaker (UNKNOWN): And
44:04
Unidentified speaker (speaker_2): there's nothing wrong with that, because they are they have super strength in development and commercialization. So again, I've thrown a whole bunch of things at you in the same kind of sentence. So just parse that out for me and tell me
44:17
Unidentified speaker (UNKNOWN): what
44:17
Unidentified speaker (speaker_2): you think. Yeah,
44:18
Unidentified speaker (speaker_3): sure thing. A lot it there is a lot there, but it's all very pertinent to our business strategy and our opportunity. Let me just step back. You mentioned governments, you know, funding uh medicines. The biotech pharma industry has been extraordinarily successful at creating medicines that have transformed many diseases. However,
44:38
Unidentified speaker (speaker_3): there's just horrendous burden of disease that persists. Even in diseases where we have great medicines, there are people don't respond well, there their are people people that intolerant, intolerant, there are people people who lose responsiveness, you know, all of that. Uh and there's patients who don't. you know, have diseases that we don't have anything for yet. So there's tons of space. So, and governments are always going to
44:58
Unidentified speaker (speaker_3): spend money on health care of their populations. Now, does it go up or down? And how you know there's a lot of politics, whatever. But it's going to be a key factor. What's going to be important is we as a company, but we as an industry develop things that provide meaningful benefits. benefit and that's measurable. And one way to
45:18
Unidentified speaker (speaker_3): do that is to work on a completely new target. Because if you're working on an old target, you're looking for okay, it had that much efficacy. I'm I'm we're not on video. I'm showing uh you
45:27
Unidentified speaker (UNKNOWN): know
45:28
Unidentified speaker (speaker_3): one foot space between my hands. You could do, you know,
45:31
Unidentified speaker (UNKNOWN): fourteen
45:33
Unidentified speaker (speaker_2): better.
45:33
Unidentified speaker (speaker_3): Yeah.
45:33
Unidentified speaker (speaker_2): Five percent
45:34
Unidentified speaker (speaker_3): increment better. Which might be important statistically, but also might be important clinically. But how many more could we find where it's the first thing that unlocks a completely novel biology in a disease? That's what the huge opportunity of what we're putting on the table is, and what the pharma companies that are collaborating with us see is an
45:54
Unidentified speaker (speaker_3): opportunity for them to break into new space first in class commercially, but also first in class. in terms of a huge opportunity for patients. So so that's how I think of those
46:07
Unidentified speaker (speaker_0): things. So
46:08
Unidentified speaker (speaker_3): we're we have the luxury of not having to find something that's gonna have a 20% margin. I'm gonna pass on that one. We're gonna work on the ones that like it's it's demonstrably a different approach to the disease that has an opportunity to have a very significant impact.
46:24
Unidentified speaker (speaker_2): Yeah. I mean that is very interesting. Uh in at least one sense. At least one sense. I mean, we talked about many other aspects of why this is a uh you know the white space that you're in, essentially. And one of them is that we have um oncology has ended up with companies trying so
46:44
Unidentified speaker (speaker_2): hard to push, especially with the immunoncology space, to get the immune system to just drive a little. higher and just drive a little harder. And after the revolution with PD1, CTLA4 targeting, things have kind of slowed down a bit until the T cell engagers came in. Now of course there we pretty much have only one for solid tumors,
47:04
Unidentified speaker (speaker_2): if I'm not wrong.
47:06
Unidentified speaker (speaker_2): Somebody was going to tell me you forgot this particular one. Yeah.
47:08
Unidentified speaker (UNKNOWN): I think
47:08
Unidentified speaker (speaker_3): it's
47:09
Unidentified speaker (UNKNOWN): a little bit more than that. And
47:09
Unidentified speaker (speaker_2): that's
47:09
Unidentified speaker (UNKNOWN): um
47:09
Unidentified speaker (speaker_2): Argenzium genzym Delta with the L LL3. L three. And that would that's proven to be and but it's given people hope that that we're gonna be able to get there. But here you have an opportunity to to expand that that way beyond what we already know. And whereas a lot of people are trying to create X plus Y, uh buy specifics, try specifics to try and
47:29
Unidentified speaker (speaker_2): see whether they can somehow magically increase the the efficacy. Um now back to Neuro. I wanna want to I wanted you to talk to me about Neuro a little bit and where you think that could goes again. And then obviously um and then we can end up talking about AI a little bit more about you know the j your application of AI or whatever you want to call it, your
47:49
Unidentified speaker (speaker_2): uh advanced technology is very, very spare specific. But from your perspective as somebody who's been in the industry for a longer time, where do you see the impact most coming? So two two separate questions, but we can come back to the AI afterwards. Let's let's focus on neuro.
48:04
Unidentified speaker (speaker_3): Yeah. So, you know, neuro is just a a a vast uh untapped opportunity to help patients. I mean, there are very few targeted, effective, safe drugs for neurologic disease. Okay. And if you think about Alzheimer's disease, Parkinson's disease, schizophrenia,
48:24
Unidentified speaker (speaker_3): ALS, so you know I there are not great medicines. And for some period of time, the problem has been scientific understanding of the disease process. There's also delivery has been a challenge into the CNS. And I think now those things are starting to come together where the scientific understanding is
48:44
Unidentified speaker (speaker_3): increasing. And there's been some really good inroads into uh delivery and modality that can start to target the those new mechanisms. So I I think the future is extremely bright for neuro. One of the interesting aspects is if you think about the names of the diseases I just listed, they're literally from the 19th century,
49:03
Unidentified speaker (speaker_2): I think.
49:04
Unidentified speaker (speaker_3): You know, and they're they're sort of like descriptive. Like, you know, Parkinson uh described uh a syndrome and it's got features. But in reality, it's a group. of very distinct molecularly driven entities that end up with the same type of manifestation in the patient, but they're actually distinct
49:23
Unidentified speaker (speaker_3): diseases.
49:24
Unidentified speaker (UNKNOWN): So
49:24
Unidentified speaker (speaker_3): I think our scientific understanding is going to now allow us to start to parse those diseases into their molecular entities and start to match new mod new drugs to the specific entity that has the mechanism. Okay. So understanding that in more detail through what we're discovering.
49:44
Unidentified speaker (UNKNOWN): Like
49:45
Unidentified speaker (speaker_3): what are the other components to the molecular pathways that drive Parkinson's disease is one of the things we're doing. Like so there are three or four or five pathways that are very well established in Parkinson's disease. But there's tons of missing pieces to that puzzle.
50:03
Unidentified speaker (speaker_3): And going back to my description of the genetics before, we've now found, I'll say dozens of proteins that have genetic association with Parkinson's are completely novel. We've demonstrated the by mass spec. They're expressed in relevant brain Uh areas. And
50:23
Unidentified speaker (speaker_3): some of them are differentially expressed in Parkinson's disease versus health. healthy brains. That's a pretty good story for those having some degree of causality in the disease and they're testable. I think we're going to see like enormous progress on
50:38
Unidentified speaker (speaker_3): new targets that could be testable and start to make a dent in these diseases.
50:43
Unidentified speaker (speaker_2): And the last couple of minutes on your your your view as to where AI is most useful in the continuum of drug development?
50:52
Unidentified speaker (speaker_3): Yeah, you know, this is a long uh it could be a
50:55
Unidentified speaker (UNKNOWN): long
50:55
Unidentified speaker (speaker_3): topic, but you know, some of the high points for me, you y you see companies the way they talk about it and even the way they implement it almost like it's a separate process. Like this could be some like
51:08
Unidentified speaker (speaker_3): new like unlocking of drug discovery that's AI. And it's it's just not the case.
51:15
Unidentified speaker (speaker_3): It's it's an enabler for the existing drug discovery process. And no matter how you do it, you're gonna need people who really understand drug discovery, know what questions to ask of the AI, know how to interpret the output from the AI, and know how to iterate with the AI. So I think the companies that are gonna do it really well are the
51:35
Unidentified speaker (speaker_3): ones that integrate the AI into the process as opposed to have it like a as separate entity. And there's no doubt it's uniquely um enabling us to see relationships, pathways, interactions, risks that we couldn't have seen just by sort of human
51:55
Unidentified speaker (speaker_3): scouring of data. Um and it's it's really pretty dramatic and it's best applied when you've got very large curated data sets,
52:04
Unidentified speaker (UNKNOWN): which
52:04
Unidentified speaker (speaker_3): is what we've generated in this space. So we're we're very bullish on how AI is gonna help
52:09
Unidentified speaker (UNKNOWN): us
52:10
Unidentified speaker (speaker_3): um
52:10
Unidentified speaker (speaker_2): uh
52:10
Unidentified speaker (speaker_3): prioritize our activities in the space.
52:13
Unidentified speaker (speaker_2): Just picking up on your last phrase there, do you think that gives f large pharma by automatically a a a an advantage over smaller biotechs given that they have decades of data, some of
52:26
Unidentified speaker (UNKNOWN): which
52:26
Unidentified speaker (speaker_2): is not never been published. It's
52:28
Unidentified speaker (UNKNOWN): a
52:28
Unidentified speaker (speaker_3): good
52:30
Unidentified speaker (speaker_3): and and I'm I'm speaking generally not about uh my own company or companies, but just as a general principle. Um many of the large pharma, as you well know, are sort of literally conglomerates of multiple companies that they've sort of uh taken
52:45
Unidentified speaker (UNKNOWN): on
52:45
Unidentified speaker (speaker_3): over the years. And they also all have, you know, uh many discrete research sites. all over the place. Um, it it's been very hard for them to integrate all of their data into one place so that it's searching. Interesting. Extremely hard. So even though they have it, it's not, it's hard to make it searchable. We we tackled that
53:05
Unidentified speaker (speaker_3): within our company when we built it from the bottom up, and it was a principle for at the beginning that all of the data had to be in one place. There's one single data lake. So all of our experimental data, all of our in silico data, all of our proteomics data is in one place. So it's, you know, very straightforward to search it using AI and ML. And we don't have this
53:25
Unidentified speaker (speaker_3): problem with distributed data sets.
53:27
Unidentified speaker (speaker_2): Brilliant. Interesting. John, I really look forward to having you back on when you filed your first I IND. and D.
53:34
Unidentified speaker (speaker_3): Super.
53:35
Unidentified speaker (speaker_2): Or you you've f filed some phenomenal patents in different areas. Anything, because this sounds like it's going to be a very rich scene that you're um attacking there. So thank you very much for joining me. That's John Lapore, CEO of Profound Therapeutics. I've used quite a bit of his time already today. Again, thank you for joining
53:53
Unidentified speaker (UNKNOWN): me.
53:54
Unidentified speaker (speaker_2): Thank you to our audience for joining us for our last latest episode. Please don't forget to give us a review and make sure to click the follow up. button on your favorite podcast app or site so you never miss a discussion with the leaders that we are uh very lucky to have join us talking about the healthcare
54:13
Unidentified speaker (speaker_2): innovation. And before I end, I want to give John one last minute, if you have anything you want to close on.
54:20
Unidentified speaker (speaker_3): Well, first and foremost, Sam, uh absolute pleasure to uh be on your podcast and to uh have the opportunity for your listenership to hear uh our discussion and about our very exciting company. As you know, to paraphrase the whole discussion, I I remain extraordinarily enthusiastic about what the whole enterprise can
54:40
Unidentified speaker (speaker_3): do for patients going forward. There's a lot of headwinds and you know things to to tackle, but um the scientific understanding of the human body is uh increasing. We have a very uh exciting piece of that, which is the identification and characterization of the expanded proteome. And I'll take you up on your offer to come back as that story
54:59
Unidentified speaker (speaker_3): unfolds and uh we start to see what it really can do for patients.
55:03
Unidentified speaker (speaker_2): I look forward to it. Um, I'm Sam Fazelli, and you've been listening to the Vanguards of Healthcare podcast by Bloomberg Intelligence. Until next time, take
55:10
Unidentified speaker (UNKNOWN): care.
56:09
Unidentified speaker (UNKNOWN): Hi,
56:09
Unidentified speaker (speaker_0): I'm Barry Rittholz inviting you to join me for the Masters in Business Podcast. Every week we bring you conversations with the people who shape markets, investing, and business. I speak with CEOs, Nobel laureates, market innovators, and legendary investors. Whether you own stocks, bonds, real estate commodities, even
56:29
Unidentified speaker (speaker_0): crypto, these are discussions you absolutely need to hear. Subscribe to the Masters in Business Podcast on Apple, Spotify, or anywhere you listen.
This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.

More from Vanguards of Healthcare by Bloomberg Intelligence

Vanguards of Healthcare by Bloomberg Intelligence

Biolinq Introduces New Microsensor Tech For Next Chapter of Glucose and Analyte Detection

Episode discussion: BioLink CEO Rich Yang details Shine sensor tech, OTC clearance, and insulin dosing roadmap

Speakers’ claims: BioLink CEO Rich Yang explains that the Shine sensor uses a 2mm silicon wafer with seven micro-probes to access the papillary dermis, eliminating hypodermic needles. He highlights the company's recent FDA OTC clearance and strategy to target the 90-95% of Type 2 diabetics not on insulin. Yang discusses using real-time LED feedback for cognitive restructuring and manufacturing advantages via semiconductor foundries. The company plans a pivotal trial for insulin dosing claims, targeting FDA submission next year, while developing a second-generation device with doubled electrode density.

Medical DevicesDiabetesCGM
41 min0 assets17 claims to verify
Vanguards of Healthcare by Bloomberg Intelligence

Akido Bets AI Can Break Healthcare’s Scarcity Problem

Episode discussion: Akido Labs CEO claims AI can deliver concierge-quality healthcare to all by breaking scarcity through Scope AI.

Speakers’ claims: Akido Labs CEO Prashant Simant explains how the company evolved from a data platform to a full-stack care delivery organization to leverage predictive models for patient outcomes. He highlights Scope AI, a neural network enabling software-based cardiology visits and clinical decision support for non-medical case managers. Simant claims the system achieves 93.6% accuracy and has reduced emergency department utilization by 50% in their street medicine program. He predicts that within five years, AI and wearables will make high-quality healthcare accessible to everyone, similar to the 'Jeff Bezos model.' The company plans to expand to 50 markets, focusing on value-based care and partnerships in underserved populations.

Artificial IntelligenceHealthcareValue-Based Care
44 min0 assets20 claims to verify
All insights