39th Annual JPMorgan Virtual Healthcare Conference

Jan 12, 2021

Slides

Good morning and welcome and thanks again for joining us at the Annual JPMorgan Healthcare Conference. I'm Eric Joseph, Senior Biotech Analyst And our first presenting company this morning is Black Diamond Therapeutics. It's my pleasure to welcome President and CEO, David Epstein, to tell us a little bit about the company. Before I hand it over to David, just a programming note, for folks who want to ask a question, feel free to do so by clicking the ask a question icon and I can ask on, you're going to have at the end of the presentation. So with that, David, thanks again for sharing some of your time with us this morning. Great. Thank you. Thank you, Eric, and thank you everybody for joining this morning. You know, I want to take a moment to thank the employees and the investors in Black Diamond. It's been a great year, a challenging year for everybody, as I'm sure you all know, and so just want to acknowledge all of their great efforts. So if you could turn for a moment to Slide 4 in your deck, just a quick intro and just to level set and get everybody page. So as you appreciate, we're a precision medicine oncology company focusing on the development of tumor agnostic therapies. And we do this through the following means. We take a look and examine population level DNA sequencing information coming directly from patients. And so I'm focusing on bullet point 2. We analyze that to identify rare and recurrent driver mutations. Once we've identified these mutations using our MAT platform, we seek to aggregate them into like families, and these are usually based on pharmacological similarities. And we assume, and we have data to back this assumption, that these pharmacological similarities are based on similar conformational states of these mutant proteins. This allows us to do 1 important thing, and that's to develop spectrum selective master key inhibitors that target the entire family of mutant oncogenes. And that's unique. So again, population level data that is essentially identified rare and recurrent driver mutations are identified through a machine learning algorithm, we then test this in in vitro and in vivo models and then develop spectrum selective inhibitors. Our lead candidate is BDTX-one hundred and 89. It's an oral irreversible small molecule inhibitor. Again, it's a master key compound inhibiting upwards of 48 allosteric lesions in both HER2 and EGFR, we're in the midst of a Phase 1, Phase 2 clinical trial. We initiated the trial not even a year ago, and I'll give you an update on our status with that particular trial. Last year, we also gave guidance that we'd be bringing forward a compound that would target allosteric mutations in EGFR alone. These are tropic for GBM. And we've developed BDX-fifteen 35 as a development compound nominated, and we've started IND enabling studies for its use in CNS as a CNS penetrant molecule to treat glioblastoma. Now I want to take you to Slide 5 very quickly and just to contrast what we're doing with what has been done before us. So if you focus on the left, the classic approach is essentially a single set of mutations or a single mutation in a single tumor type that's led to some fantastic improvements in patient outcome, and we're really proud of those accomplishments. I and several founders of the company cut our teeth at OSI Pharmaceuticals, where we developed and worked on 1 of the early iterations of EGFR, Tyrosine Kinase inhibitors in the form of erlotinib. Over the last 5 to 10 years, there's been an expansion of genetic profiling, particularly of human tumors. And so, we now know that there are upwards of a 1000000 patients worth of data, either through subscription or in the public domain databases. And importantly, this is millions of unique mutations, not 100, and certainly not just focal to the active site. And the critical observation is most of these patients and most of the mutations are essentially unannotated and most of the patients derive essentially no actionable benefit. And so the idea of identifying rare and recurrent driver mutations, aggregating them into families, and then developing a precision medicine that can target that family, I think is hugely differentiating, and we've built the entire company and platform around this concept. I'd like to just quickly turn to Slide 6. This is an update on the pipeline. Obviously, hundred and 89 will give an update. We'll finish the Phase I in the first half of this year, and we'll provide a clinical update at ASCO on the status of the Phase I study. We've brought forth BTX-fifteen 35 again for the treatment of allosteric mutations, focal to EGFR and tropic to glioblastoma. And we'll be in the clinic at the first half of 20 22, and the intention is to file the IND as quickly as we can. It's a complicated molecule, and we're going to need the full year to get it into development. Okay. So I'd like people to turn to slide 8. We're going to spend 4 slides just describing the map platform and then we'll get into the clinical program on BDX-one hundred and 89. Again, we focus on population level data, analyzing it to identify rare and recurrent mutations. The allosteroid portion allows us to discern whether these mutations are actually oncogenic driver mutations or silent mutations, because all of these mutations, not just most, all these mutations are outside the canonical activating active site of just virtually every target we look at, they're allosteric in nature, they drive confirmational change, and they change the pharmacology and that's shown on the far right, such that 1st generation therapies just simply are not designed to target these mutations, they're significant commercially, they're important to patients, and there's really essentially a true unmet need. So how does this work? Shown on slide 9 is a depiction of the first step of identifying rare and recurrent mutations and annotating them with essentially a propensity for optogenicity. And so on the far left, you see what is really the true picture of population level data for a single oncogene, 100, if not 1000, of point mutations spread throughout the entire oncogene. We use computational sciences, both at the protein structure level as well as at the genomic level to rank order the oncogenic potential of each and every 1 of these mutations in each and every oncogene that we've studied. We provide a MAP score for every residue. High MAP score predicts a likely oncogenic mutation. Low MAP score predicts silent or essentially inactive mutations. And so what we then do is we convert all of this information into empirical studies. This allows us to focus on the top mutations, and then to annotate them in house, we do that, and on slide 10 is really a nice summary of the importance of looking at these mutations in the way that we do. And so you see 4 different forms of EGFR HER2 in their dimeric state in maroon are the regions where these oncogenic mutations lie. You'll note that they're in the extracellular ligand binding domain. You'll note that they're in the transmembrane region, the juxtamembrane regions in cartoons 23, as well as in the kinase domain itself, but not necessarily in the ATP site. And so this allows us to assess each and every 1 of these mutations in cell based assays, convert them into in vivo models, and what you see on the bottom is essentially the spectrum of mutations against which BDTX-one hundred and 89 was developed. And as we've said in the past, this is upwards of 48 mutations. However, it's important to note as the data continues to accumulate, there are more and more mutations that are proximal to the anchor mutations that make up the core of this, and we anticipate that those patients may also derive benefit with a molecule like BDTX-one hundred and 89. So slide 11 then is a summary of a master key approach, where a single molecule has been developed that can target this family of activating mutations. We always get asked, how does this work, and the very simple answer is we don't care about the activated state per se. The activated state is in the kinase domain where the business end of the molecule is. What we care about is whether these mutations are activating mutations, and thus drive the kinase domain in an activated conformer that can be drugged by a novel molecule, and that's what we do. Okay, so Slide 12 then is a summary of all of the efforts over the last several years in essentially developing and perfecting the MAP platform. Obviously, we've got 2 lead programs targeting HER2 and EGFR that have derived themselves from this platform. We've essentially assessed more than 300 oncogenes and tumor suppressors focusing on the 92 kinases that are evident in all of these NGS platform panels. And we focused down on the 7 allosteric targets described here. We this year will update the community on our allosteric BRAF program as well as our allosteric FGFR program. We're developing quite a bit of know how in the ERB space, and we're going to continue to iterate off of the learnings that we've obtained from both the GBM program as well as BDTX 189. And so as we move forward in 2021 2022, you'll hear more about the IRB program in and of itself. Okay, so let's focus now on BDTX-one hundred and 89. Just stopping momentarily on Slide 14 again, it's a master key inhibitor, it's a tyrosine kinase inhibitor, it is targeting the ATP active site, it's irreversible, so it targets the exact same cysteine residue that is present in many receptor tyrosine kinase, the exact same cysteine residues, 797 or 795, depending on your nomenclature, that are the targets of neratinib, Tagrisso and so on and so forth. So I just want to reiterate on Slide 15 the depth of value that is possible with BDTX-one hundred and 89. Today, the assessment that we've made is that this can treat this compound can treat upwards of 30, 000 patients, not just in lung cancer, we're often class side by side with many of the other exon 20 inhibitors that are in development. We'll talk about that, but that represent the exon 20 space represents around 20% to 30% of the inherent value in the BDTX-one hundred and 89 molecule. Breast cancer driven by allosteric mutations is quite significant. Solid tumors with specific mutations in HER2, S310 are of interest. It's the most prevalent mutation in HER2 in all solid tumors, and we are the only company focusing on HER2 S310F. We have cohorts looking at all these mutations and other tumor types, hence the tumor agnostic strategy, And we are opportunistic, and we are looking for other avenues for development of BDTX-one hundred and 89, given the selectivity profile. And that's shown here on Slide 16, where we've now presented this in several conferences. This is the spectrum of mutations that are in the target product profile for BDTX-one hundred and 89 shown on the far left, annotated and color coded with respect to location on HER2 and EGFR, a cartoon on the lower left. The critical feature here is not only the broad spectrum activity of this compound, but the fact that it spares wild type in 2 gold standard, each wild type EGFR cell based assays. And that spectrum of selectivity, at least at the in vitro level, is anywhere from 7 to several 100 fold, as you can see here. And so the thought process in the company was to spare wild type, develop a spectrum specific agent, but not to do so at the expense of off target selectivity, and you can see that we've achieved that with the kinome screen shown on the far right. Really, what this molecule does is targets the ErbB family that's in the center, so that's EGFR HER2 through 34, off target activity for BLK as well as RipK2, not targets that are expected to give rise to adverse toxicities, and in effect, we'll talk about what we know about the toxicity of BTEX189 subsequently. It's important to point out that while many of these mutations are exon 20 focused, again, it's not the majority of mutations, but for those in the community that want to understand BDTX-one hundred and 89 relative to our nearest competitor compounds, be it the cloning compound, mobasertinib, we believe that based on our data and head to head comparison, side by side in our cell based assays, that we have what is probably best in class in terms of selectivity versus wild type. And hopefully, that selectivity translates into better tolerability and the ability to dose escalate above the range of IC50 values that you saw on the previous slide. It is important to point out that a critical aspect of BDTX-one hundred and 89 is its ability to act as a hit and run Tyrosine Kinase inhibitor. And this is really the first demonstration of a hit and run compound targeting solid tumors. There's obviously a history of this in the liquid tumor field, but for now let's just focus on BDTX-one hundred and 89. Lower left hand panel shows rapid target engagement in a cell based assay driven by the S310F mutation, and that inactivation is sustained out to 24 hours. It doesn't matter whether this inactivation occurs in the monomer or dimer form, both are essentially equally inactivated within minutes. This translates into a rapid inactivation in vivo PKPD model. And as many of you know, we've been doing this as essentially a way of correlating the activity with the PK in a number of models and we won't go through that today. In red is rapid shows rapid inactivation after an oral dose that sustained out to 24 hours. That uptick at 24 hours in red, we think relates to receptor recycling, certainly not due to resistance. This is a single dose experiment in mice. Importantly, you see rapid absorption of the compound in black dash in black dash lines, which is accompanied by rapid elimination, half life in the hour of 1 in the time frame of 1 to 3 hours and at 24 hours essentially fully eliminated, whereas the pharmacodynamic effects are sustained out beyond the period of time that the compound is present. This translates into dose dependent tumor growth inhibition on Slide 19, as well as tumor regression. And you can see how we're range bound in the ED50 values. We've used this to calculate and determine our human estimated clinical doses. All of this information will be laid out at the upcoming AACR meeting, how we've used PD PK modeling, physiological based PK modeling and these data sets to drive the conclusion that we have the potential to target each and every 1 of the mutations that were identified in the previous slides. And I'll just point out what we're looking at here are 2 different mutations, HER2S310F in an allograft model showing an ED50 of 30 milligrams per kg and the EGFR exon 20 mutation, otherwise known as the ASV mutation, where the ED50 is 10 mgs per kg. So there appears to be really no correlation between IC50 value and the ED50 value, and you can extract that data from the slides that we've provided you. All in all, this has led to a fairly benign tox profile. No ocular skin changes associated with wild type EGFR was seen in our non clinical studies. The only histopath finding that was target organ directed is in the GI tract. It was focal, as many of you know, and is reversible. And essentially, this is a compound with no QTC liabilities. And so we anticipate no cardiovascular risk or low cardiovascular risk for this compound in the clinic. Slide 21, I think is the bread and butter. Obviously, many of you are very interested in this particular slide we've completed, as given guidance. Dose escalation, both in the QD and the BID cohorts, and we've identified a maximum tolerated dose in both of these regimens. We're now examining food effect. I will save patient numbers for the question and answer period. We are on track for identifying a recommended Phase II dose. We'll confirm that in the Q1 of this year and then begin our safety expansion studies in Q2 of this year as well and well en route to initiating what we believe will focused Phase 2 programs targeting this family of mutations. And just to reiterate, I know this question will come up again, cohort 1, as designed, as shown on this slide, is focusing on non small cell lung cancer in the exon 20 groups, both EGFR and HER2 that will likely be divided into 2 separate cohorts. Cohort 3 then or excuse me, the breast cancer cohort will be focusing on allosteric mutations that are tropic for breast cancer. Cohort 3, as described here, is essentially the pure tumor agnostic cohort looking at S310F S310 mutations present in all tumor types. And everything else is then bundled into the last cohort tumor types and mutations that are not being studied in cohorts 1 through 3. So last few moments, just want to focus on BDTX-fifteen 35, again, our brain penetrant Tyrosine Kinase inhibitor that was developed against a family of mutations that are evident in GBM. And I'd like to turn your attention to Slide 24 just to remind you that the location of these mutations is focal to the extracellular ligand binding domain. There are no mutations in the kinase domain to speak of in EGFR. So all these mutations result in a constitutively active covalently dimerized form of EGFR. We've now published that in various poster presentations, a manuscript now submitted and under review describing all of this information. It is important to point out that the work that was done in the previous years with Tyrosine Kinase inhibitors did not appreciate the fact that there were additional mutations that occurred in the same patient population, even in the same patients carrying these mutations in the extracellular ligand binding domain. So it is critical that a molecule not only cross the blood brain barrier, but that it's able to inactivate each and every 1 of these mutations because any 1 patient can carry 1 or more of these in an individual tumor, and that data is shown in essentially a Venn diagram on the far right of Slide 24. So very quickly, Slide 25 just summarizes we think the differentiating properties of 1535, again, pan selectivity against the allosteric EGFR variance expressed in GBM, as shown in the middle panel for 1535, a slightly sparing wild type activity and certainly a completely differentiated profile in comparison to osimertinib. This has resulted in, on Slide 26, essentially substantial brain penetrance as well as activity in an orthotopic brain patient derived xenograft model and that data is shown on Slide 26. So we'll have a lot more to say about the GBM program as we move forward through our IND enabling studies. I just want to take this time to thank the team, to thank you for your patience and your interest in the company, and most importantly, to thank the clinicians and the patients who participated in our Phase 1, Phase 2 program. And with that, Eric, I'll turn it over to you for questions. Great, great. Thanks for that presentation and overview, David. I guess, yes, let's go to the highly anticipated or sought after question. Really, it's just a little more granularity on the patient numbers, numbers of treatment cohorts and really the mutational backgrounds that we might see comprising the Phase 1 data looking to ASCO later in the first half? Yes. So I think so we've it's gotten, I think, well, as you've heard. We have now treated up to or in the range of 40 to 50 patients in the Phase 1 program. And given the fact, and you've heard us make this statement in the past, but it's just absolutely true, given the fact that the heterogeneity of the target product profile upwards of 48, and in fact, we're including patients who could receive benefits. So those are HER2 amplified patients, patients that are carrying activating mutations, so called canonical mutations that didn't respond to prior treatment with an EGFR or HER2 directed agent. Those patients are also included in the trial. And so it's important to point out that many of these patients are also carrying CNS lesions into the study. So the patient demographics are pretty complicated and it is at this point in time I think impossible to draw a specific conclusion about the activity of the drug in a given tumor type or the given mutation. But that said, we think we're on track with essentially the initial deadline for ASCO submission to submit with 40 patients worth of data that will include the identification of MTD in both of these dosing regimens. But at the time of ASCO, we'll outline our food effects studies. We'll summarize our efforts in identifying the recommended Phase II dose, and we'll certainly summarize our conversations with the FDA with respect to how we're going to plan and organize these 4 or 5 Phase 2 programs that are diagrammed as Part B on the clinical development slide. So I think we're just going to have to hold our breath a little bit longer as we work through the data. Obviously, I think you should expect to see data that will allow you and the rest of the community to do a head to head comparison in the safety and tolerability of BDTX-one hundred and 89 in comparison to those competitors that we all know, particularly in the exon 20 space. So there will be sufficient data to compare the thesis that this means of identifying a wild type sparing molecule that is dosed with a short half life will in fact spare these EGFR mediated off target toxicities of the skin and the eye, and whether or not the gut toxicities that we see in the preclinical data actually are what are the dose limiting toxicities in the clinical studies. I think you'll have sufficient data to draw that conclusion. Certainly, we will put forth as much data with respect to biological activity of BDTX-one hundred and 89 in this diverse patient population. But I just want to remind you and everyone else, the intention of this study is safety, tolerability and dosing. And we think we have achieved the initial objective of being able to answer that in a very clear cut and concise manner for you and the community. I will say that we need to I want to make sure that everyone understands that the initial thesis that drove the company through its IPO was this hit and run strategy and the ability to use these kinds of molecules to infer a human clinical dose. And so we will have the data that will allow that direct comparison. It is important to understand that to do that, I think those who are deeply steeped in the assessment of this company and the compound prior to the IPO will appreciate the modeling that was done on this type of compound to get to an estimated human clinical dose. And again, we'll have all of that data available at AACR so that by the time we get to ASCO, a direct assessment of whether we could predict what we were going to get and did we get it both in terms of safety, tolerability and exposure. And of course, any attendant biological activity, we'll report on as well. So we're pretty happy with where we've gotten to. Did all this through the COVID crisis, as many of you did. But again, I just really want to thank the teams for doing this under what really is an incredible year. So I'm happy to answer any more questions on that. Yes, yes. When it comes to biological activity, right, it might be tough to get sort of a clear signal just given the heterogeneity of the population. Is there an opportunity to assess sort of pharmacodynamic activity in a tumor, I guess, maybe that would require biopsies because that might be challenging in this space? So we do have so biopsies were optional. And we do have a few samples. I think it's a little early right now. We're trying to assess whether we're going to have that data for the we'll have that data. It'll be looking at downstream ERK signaling, as we've done in our preclinical assessment. So certainly, the goal in this year is to have that information. You know, it's obviously, you know, I think the team is working to get that. We'll see how far we get. I think I just want to come back because the other question that I know you're going to ask is, will there be enough depth in the dataset to provide an ORR? Absolutely not. I just want to really make sure that we level set. Biological activity assessment is going to be inherent in what we present. But given the fact that we're going to have, we've essentially said that we've tested this compound in 40 to 50 patients in this Phase 1 portion. The majority of those patients, more than half essentially, had to have received sub efficacious exposures. The other half may have received efficacious exposures. That's the analysis that we're now doing. That's the analysis that we'll provide at ASCO. Even with that, the range of people, of mutations that have come in on this study is just too great to provide a meaningful ORR today. We will, of course, take a look at the patients that continue to come in before the final data cutoff, and we'll provide an update as to how we're doing. But as it stands now, I think we should just focus on, on the PK, PD, the activity data, attended with that in ancillary groups of patients. And then we'll move forward from there with the recommended Phase II dose and the expansion cohorts moving forward into the Phase II. I think this could relate to either the Phase 1 or the expansion Phase 2 cohorts. I guess, are patients I guess, what are the eligibility criteria with respect to either 1st or second generation EGFR inhibitors or other exon 20 inhibitors are new? They could if they had an exon 20 mutation or a HER2 amplification, you know, or a canonical mutation in EGFR, they were eligible as long as they didn't respond to that therapy. If they couldn't tolerate the therapy or if they just progressed on that therapy and never responded, then they can come on to the trial. That said, so it's a that I think is the reason why we were able to get so rapidly through the dose expansion, dose escalation cohorts within a year. I mean, we really haven't even been into the study for a full year. But that said, we're going to tighten up those criteria now going into the safety expansion studies. I think there's discussion on the team about whether we will eliminate patients or prevent patients from entering the study who have CNS lesions. We wanted to ask that question. I think at this point in time, that's probably I think it's safe to say that BDTX-one hundred and 89 is not blood brain barrier penetrant. It wasn't designed to do so. It's a Type 2 inhibitor. And so that leads to the question of are we developing agents that can cross the blood brain barrier and target some or all of these lesions, absolutely, the answer is yes. That's not part of this study. I think what we want to do is focus BDTX-one hundred and 89 on the patients that can, in fact, respond under recommended Phase II dose conditions and get an answer as to whether we have path forward in all of these cohorts. So the intent is to study that extensively through this year and to have more robust activity information for you and the community by the end of 2021 and into 2022. Okay. I guess, many incremental preclinical work that you've done since the IPO with respect to the, I guess, the durability of effect in some of the more rare, allosteric mutating subsets of those Cohort 4 subsets given the incremental sensing ability to sort of more durably, you know, get tumor regressions or tumor growth inhibition mutation backgrounds? Yeah, I think, so from a preclinical model, we're starting, from a preclinical assessment, we actually are starting those studies now. I think that there's a lot of focus really on the I think there's a lot of benefit to focus on what we're learning from summit study and the continued publication from Puma and others around the use of neratinib in this setting. You know, there's a that says to us that these mutations, you know, lead to clinically active oncogene driven tumors essentially provide an agent that can be dosed over a long period and tolerated over a long period so that we can actually assess tolerability and durability in a clinical setting. And you can see the impact of that when you look at TAK-seven 88 with a 7 month durability, 43% response rate in a population that doesn't really respond well to any therapy. And so that I think that sets a benchmark. Obviously, more to think about with all these other mutations, but that's sort of how we're thinking about it. I guess if we're thinking about the timelines related to the expansion, the Phase 2 expansion cohorts, I mean, there is some precedence that by the exon 20 lung cancer studies. How should we be thinking about the pace of enrollment in, say, the breast cancer cohort or the S310F based on sort of the frequency with which those mutations? So, yeah, so this is the group of patients that's just focusing on breast cancer alone in sort of cohort 2 as described here. These are non HER2 amplified patients, so these would be considered essentially HER2 negative in an IHC format and therefore would have been sequenced. And so this is a significant group of patients. If you go back to, you know, our essentially our summary slide on the patient numbers. I mean, this is a large group of patients. It's upwards of 8, 500 patients per year. So in percentages, they're low, but given the fact that there are almost 300, 000 women with breast cancer in the United States alone annually, You know, 1 can then determine the prevalence, but we're seeing these patients on the study. So these patients are being sequenced. We don't have hard numbers for the percentage of breast cancer patients that should have been sequenced are being sequenced, but there's enough that we've been able to fill the cohorts as we've gone through the dose expansion studies, dose escalation studies with virtually examples of all of these clusters of mutations. So we don't at this point in time, I don't have enough data to say 1 way or another, but I think we should expect, I think, significant uptick in the enrollment as we proceed through the Phase into the Phase 2. We'll certainly give an update on the number of sites that we're going to open, all of that, I would say, as we proceed through the spring. Certainly, the 1535 compound is, sort of the next priority moving into the GBM space, but I'm curious about sort of ongoings kind of further, and sort of more of the discovery programs, looking at allosteric BRAF mutations, FGFR, PI3 kinase, maybe you can just sort of tease us a little bit more as to, you know, where things stand on those programs, how far we might not be from sort of developing candidates being nominated there? Thanks for the question. So obviously, we've done we've made, I think, a huge push to get out from the umbrella of being a single asset company. And happy to report that we have viable chemical matter for both BRAF and the FGFR programs. We have just submitted abstracts to present a summary of the data without declaring the compound at the TAP meeting in March. So I would look to that. We will update the GBM program at AACR. But FGF so our allo FGF, our allo BRAF programs are on track for development compound nomination this year. And then initiation of the IND enabling studies and filing the INDs for those 2 programs next year. So really happy with the progress. I will reiterate that we believe we have the first example of an FGFR program that is truly sparing of wild type FGFR1. As I think the field knows, FGFR1 is the EGFR wild type equivalent of the dose limiting toxicity target. And so we have compounds that actually can spare wild type FGFR1 and target allosteric mutations and fusion proteins in FGFR2 and 3. So we think this is quite an achievement. And also, we've demonstrated that we can do that and pick up some of the gatekeeper mutations that have been identified in FGFR. These compounds are biologically active in preclinical models, pharmacodynamically active, and we have essentially shown us the same phenomenon. Mutations outside the canonical active site have a profound effect on pharmacology. And when you focus this way, I believe, and I think the team is now convinced that when you focus mutation first, you can design and discover molecules that spare unintended off target activities like FGFR1. And as you know, the field is littered with statements about FGFR1 selectivity, and that selectivity is essentially met with dose reduction and cessation of therapy. So I think clinically, I don't think it's those are statements in fact. And so we'll come forward with our molecule and we'll see whether we're able to succeed in the clinic with that kind of profile. The BRAF space, we understand the Class II, Class III arguments and we think that they're invalid. And so we are developing molecules that inhibit the BRAF dimers. It's important to point out that this concept of allosteric mutations that have originally been classed as class II and class III is absolutely the same thing. It's just that the platform that we've built allows us to assess all mutations in the physical cluster that would contribute to either a class 2 or a class 3 annotation. And so the sense that you get from the map platform, it's just a more holistic way of looking at all mutations that are present in the patient population, assessing them, aggregating them and then building the programs first around that. So this idea of screening against wild type and then retrofitting to mutant, it's just not something we're interested in doing because it doesn't lead to differentiated molecules. And I think, you'll see this year that this platform is truly enabling and truly differentiating. Great. Thanks so much again for your time, David. Thank you for the time. Thanks everybody for tuning in. Yes. Have a good morning. You too. You too. Take care. Bye bye.