Merck & Co., Inc. (MRK) Earnings Call Transcript & Summary

Hello, and welcome to today's webinar. I'm Ryan with the digital chemistry team, and I will be moderating today's presentation entitled Accelerating Synthesis Design and Modeling using Collaborative Mixed Reality. [Operator Instructions] We've provided resources related to the topics discussed in the webinar, which you can access using the resource list found in the platform. [Operator Instructions] Our team will follow up by e-mail regarding any questions we are unable to address during the live event. Please note that the information and advice contained herein does not relieve our customers of their own responsibility for checking the suitability of our products for their designed purpose. Also note that this webinar will be recorded and available for you to watch on demand and that you will receive a link to the video in an e-mail by tomorrow. During today's presentation, we'll hear from Dr. Emma Gardener. Dr. Gardener received her undergraduate degree from Trinity College and worked in the biotech industry prior to grad school. In 2017, she completed a PhD in organic chemistry at Brown University as an NSF research fellow, where she worked on developing new methodology for the synthesis of antibacterial peptide natural products. In 2018, she joined the Chem Informatics Technologies Department as a technical application scientist and is responsible for the commercial licensing of the retrosynthetic design software, SYNTHIA. We'll also hear from Dr. Maciej Wójcikowski. Dr. Wójcikowski holds a PhD in Biophysics from the Institute of Biochemistry and Biophysics Polish Academy of Sciences in Warsaw, Poland. His area of research was in-silico drug discovery methods with machine learning-based approaches being a major part of his thesis. Maciej joined the SYNTHIA R&D group in 2018 and has been working on improvements to SYNTHIA cheminformatics predictions ever since. He now serves as a cheminformatics lead within the digital center of excellence. Our special guest speaker is Co-Founder and Chief Experience Officer of Nanome, Edgardo Leija. After pursuing studies in computer science, neuroscience and cognitive science, he obtained his degrees from the University of California, San Diego. Edgardo previously worked at Google and Hewlett-Packard, where he focused on human-centered product design and implementing systems to improve the overall customer experience. At Nanome, he works on expanding Nanome's reach in the pharmaceutical and biotech industries to help them leverage the best computational chemistry tools through an immersive virtual environment that facilitates effective collaboration amongst scientists. Before we begin, we would like to take a moment to learn more about our audience today and encourage you to participate in a quick poll question. Which best describes your current role? We'll get the audience just a few more seconds to make their selection. And we have almost 50% of attendees, submitting a response. Okay. Thank you to everyone that participated in the poll. We see that 23% of respondents identify as a Process R&D chemist, 10% as a Discovery Medicinal Chemist, 16%, almost 17% in the IT cheminformatics or computational chemistry space, and then we have lesser roles of lab managers at 5% and a little over 1% biologists or bioinformaticians. Thank you again to everyone who responded. With that, I will now hand things over to our first speaker, Dr. Emma Gardener. Emma?

Great. Thanks, Ryan, and thanks, everyone, for participating in that poll. I'm excited to talk to you today about SYNTHIA retrosynthesis software. And to get started, you saw we have a variety of sort of different roles in the audience today. Cynthia is really a tool that can help with a variety of different industries as well, really anyone that's doing organic chemistry research making organic small molecules can benefit from using to help reduce risk, support sustainability and generate new ideas in your synthetic planning. But today, we're really talking kind of focused on drug discovery. So I'm going to talk a little bit about how SYNTHIA fits into that workflow. So first of all, we start off with the biology research, where we're going to identify what biological targets that we want to modulate. So protein or an enzyme that we want to modify the activity of. Once those have been identified, we move into the what to make phase where we're trying to figure out what molecules are going to have the desired effect on that target. This typically involves high throughput and virtual screens and in silicon predictors to look at the vastness of the chemical space and narrow it down to a subset of compounds that are going to have the desired activity. Once we've identified that smaller subset that we actually want to try to make and test in lab, we move into this, how to make fees. This is where the chemistry route planning comes in to figure out how we can best make those molecules what starting materials can we use, what reactions can we use to actually get those targets. And then we move into the make phase where we're actually doing that synthesis and eventually scale up and hopefully, chemical production for the ideal target. And so we're going to be talking today about several tools that can help with various stages of this workflow. I'm going to start talking about SYNTHIA retrosynthesis software, our main web app version, which is really going to help with the how to make phase doing that route planning and continues to be useful sort of down the pipeline when you're doing the synthesis on scale up. And then I'll pass it off to my colleagues to talk a little bit more about some of the other SYNTHIA offerings as well as Nanome, and how these tools can work together to help you with the what to make phase. Great. Thanks, Ryan, and thanks, everyone, for participating in that poll. I'm excited to talk to you today about SYNTHIA retrosynthesis software. And to get started, you saw we have a variety of sort of different roles in the audience today. Cynthia is really a tool that can help with a variety of different industries as well, really anyone that's doing organic chemistry research making organic small molecules can benefit from using to help reduce risk, support sustainability and generate new ideas in your synthetic planning. But today, we're really talking kind of focused on drug discovery. So I'm going to talk a little bit about how SYNTHIA fits into that workflow. So first of all, we start off with the biology research, where we're going to identify what biological targets that we want to modulate. So protein or an enzyme that we want to modify the activity of. Once those have been identified, we move into the what to make phase where we're trying to figure out what molecules are going to have the desired effect on that target. This typically involves high throughput and virtual screens and in silicon predictors to look at the vastness of the chemical space and narrow it down to a subset of compounds that are going to have the desired activity. Once we've identified that smaller subset that we actually want to try to make and test in lab, we move into this, how to make fees. This is where the chemistry route planning comes in to figure out how we can best make those molecules what starting materials can we use, what reactions can we use to actually get those targets. And then we move into the make phase where we're actually doing that synthesis and eventually scale up and hopefully, chemical production for the ideal target. And so we're going to be talking today about several tools that can help with various stages of this workflow. I'm going to start talking about SYNTHIA retrosynthesis software, our main web app version, which is really going to help with the how to make phase doing that route planning and continues to be useful sort of down the pipeline when you're doing the synthesis on scale up. And then I'll pass it off to my colleagues to talk a little bit more about some of the other SYNTHIA offerings as well as Nanome, and how these tools can work together to help you with the what to make phase. So first of all, some of the challenges in developing a route for making a small molecule in that how to make phase, so one of them is that as human chemists, we spend a lot of time learning all of the chemistry reactions. But we typically are usually somewhat biased in the reactions that we sort of think of most frequently. We maybe can't remember every possible reaction that's ever been done or at least not quickly recall it. So we often are falling into a bias of a subset of reactions. And we might be missing something that could be really useful for a particular synthesis. So then, of course, we go and look to see what's been done before in the literature. A lot of times, I know we've all experienced an established route or published route that isn't reproducible or for making something brand new, there might not be anything that's really similar. And so we're stuck trying to adapt something that may or may not work. And this failure can happen at any step and sort of the later that, that happens, the more costly it is. So the more information we have upfront and the planning, the better off. And then finally, making sure that we can actually get the building blocks to start these routes. And again, most of us don't know off the top of our head, every possible molecule that you can buy and so having something that sort of gives you that altogether, we have to do separate searches to figure out can I buy this? Is it a reliable supplier, things like that? And so these are all challenges that SYNTHIA can really help with in this route planning phase by uniting our databases of not only available chemicals, but also of chemical knowledge using modern high-power computing and some strategic chemical algorithms to be able to fairly quickly design full pathways for your target molecule, whether or not it's ever been made before, starting from commercially available starting materials. So how SYNTHIA does this. One of the key things is it needs to know what chemistry can be done. And SYNTHIA has a unique way of approaching this with what we call expert coated rules. So we've actually had a team of chemists, human chemists who have learned to code, who have written all of these reaction rules. And you can think of these like the reactions you would find in a textbook. They've been done many times, but they've been written in a way so that they are able to consider all the things a human chemists would consider when you're looking at reaction. So are there any groups here that might not be comparable with the reaction or that might need to be protected? Are there any stereo and regiochemistry things I need to consider? What are the potential steric and electronic effects. But they're also written in a general way, which is what allows them to be applied equally to novel molecules as to something that has been made before. So unlike just looking at a collection of published reactions, this is really sort of equally an unbiased approach to making your novel molecules. And I'd like to show on the slide here, just an example of what one of these rules looks like, and you can see there's a lot of code there to describe not only the reaction itself, but also all of those considerations that I mentioned. But what's also coated in here are some typical reaction conditions and some references. So that SYNTHIA can give you that information when it uses that rule. So these are all coded basically when we're writing the rules, we look at what are the typical conditions for this? And what are some key references so you get that information. So then with this database of all of the chemistry reaction rules, we're able to start to generate this very complex web of all of the possible ways that you could think about making molecule. And so in order to navigate this, we have some sophisticated algorithms that are able to fairly quickly find complete pathways. And how it does this is we start with the target we're looking to make and then look at all the retrosynthetic cuts going one step back. And from there, they're going to be scored so that we can prioritize the most promising options. And then it's going to look at is this molecule something that I can buy as a starting material. And if so, great, that's my whole pathway. And if not, it's going to look at how do we make those molecules. And do the same thing until it's found complete pathways starting from things that have been designated as starting materials, whether this is 1 step or 20 steps. And from there, it's going to rank these pathways so that it's giving you kind of the top options to consider. And this is where the user comes in to really guide the software because sometimes we really just want the shortest route to make something. But other times, we might have considerations like we're looking to avoid a certain reagent or we want to start from a certain building block. And so these are all customizations that the user can give to really guide the software so that you're getting a customized result set that works for you. And there's really essentially limitless ways to customize the software, but some of the things that SYNTHIA will consider, of course, the number of steps in the pathway. It can also consider the cost of starting materials and whether or not you want to use protecting groups, you can designate specific [ foreign ] disconnections that you either want to make or you want to make sure you don't make during the synthesis. And then really anything from starting materials, intermediates, reaction classes or certain reagents that you either do or don't want to use can all the information that software can use to not only change how it ranks the pathways, but really guide how it's exploring that graph of all of the possibilities. And then finally, defining what you wanted to consider starting materials, and we have our database of Sigma-Aldrich compounds as well as around 80 different partner vendors that are all vetted compounds that are real molecules you could combine not just something that's available on a virtual catalog. And you can also include a custom inventory as well, if you'd like. From there, SYNTHIA is then going to show you the results in some easy to navigate ways. My first picture in A is a screenshot of what our overall graph view looks like. If you want to look at all of your results together, just under that is showing our pathway view where you can see we have this node representation. So you can very quickly see the individual pathways and how many steps is it? All those pink circles are things that are commercially available, so you can kind of see that. And then for every reaction, you'll see the details that are shown here. So you'll see the reaction scheme itself as well as the name of the reaction and then those conditions and some references that are all coming from those rules. So you have an idea of where that reaction is coming from, why that's being proposed. And so this is everything that you would see or at least some of the things that you would see looking at SYNTHIA in our web app, which is really great when you've got that one target molecule or several molecules that you know you really want to make here and kind of explore the routes in detail. But we also have some API offering, so an application programming interface that allows SYNTHIA to be connected to your other cheminformatics tools so you can visualize SYNTHIA data along with some of that other data that you're using to inform your decisions about what molecules you want to make. And so for that, I'm actually going to pass it off to my colleague, Maciej, and he will be able to tell you a little bit more about our API offerings and how they work.

Good. Thank you, Emma. I think we double click to the next slide. Thank you, Emma, for introducing the web app, and how SYNTHIA works to everyone. I will be now introducing the API offerings and the second part of this presentation might be somewhat technical, but I think this will be very informative to everyone, no matter how it was the technical background of the of people attending here. So I'm going to go back to the dry discovery workflow. So Emma said how SYNTHIA kind of excels at how to make a stage of the workflow. And as we mentioned, with API, we want to move SYNTHIA's method and SYNTHIA's solution to the left of the workflow basically to what's to make. And with retorsynthesis API being slightly closer to make, but with the synthetic accessibility score API, this is really tackling the high throughput screens and high-performance applications, which was unprecedented before in retrosynthetic space. So very briefly about the full retrosynthesis API, so basically, we have our web app with a full set of features and full retrosynthesis API mimics majority of those features, and it gives you very rich predictions for your target molecules, it returns multiple pathways with abundance of meta data about molecules, about the reactions, about the building blocks. So essentially, what you see in the web app is what you would be able to get from the API, including literature references that Emma mentioned and typical reaction conditions, prices of reaction of starting materials. All of this will be much relatable and are right in the answer from the API. And this gives you great versatility in how you can integrate that API in your internal systems or other solutions, but also what sort of analysis you can make because it will ask the web app return multiple pathways, not just a single one. So this will allow analyzing also the difference between pathways themselves for the single target. And the second part of the API is synthetic accessibility score, which is intended to serve a very different purpose. This is a machine learning-based predictor, which estimates how difficult it is to make synthesized molecules. It ranges or it predicts on a scale 1 to 10, where 1 is easy to synthesize and 10 is hard to synthesize. It roughly predicts how many steps it takes to synthesize a molecule, and it's sort of quick and dirty prediction. So it's -- it will generate you an approximate amount of steps. It was trained on SYNTHIA's full retrosynthesis predictions using a custom set of settings that worked best for these kind of applications. In particular, we didn't penalize on selectivity in the predictions. We use implicit protection strategies, which works really well in these applications. We focused on smaller building blocks. So basically, this problem of synthetically is very -- very specific to the set of building blocks you define. So we were a little conservative, and we [ started ] with smaller building blocks, more basic building blocks that were used for training of the models. And what user queries the synthetic accessibility score with is smiles of the molecule. So basically, you query your molecule and you will return the estimate score as an output. To have a few examples of how this can be integrated and we've -- what sort of analysis that this can be paired with, we show this funnel of high throughput experiments. And SYNTHIA SaaS is tailored to be used for large quantities of virtual molecules, for example, to pre-filter for experiments like molecular docking or more expensive computationally. And it takes roughly 20 milliseconds per molecule to get the prediction. So it's really fast. And on top of that, it can be paired with generative models like Edison from Merck, which is another software that Merck is offering, but also with your internal generative AIs that you have implemented yourself. To be more specific on how SaaS works, we used graph convolutional neural nets as our model of choice. So instead of having a coded representation of molecules like fingerprints, we are using this graph convolutional neural net to automatically generate the representation for the molecules that we use for training and for the specific implementation we choose [Indiscernible], which is an open source and widely used machine learning model. that is really known for superior performance on predictive properties applications. In terms of the molecules that we used to generate the predictions and train the models on, we used 2 data sets and roughly 33,000 of molecules for the training. We used GDB data set, which is a combinatorial database of molecules that are articularly generated and in closes just coming [Indiscernible] enumeration of all possible connections between the cell of atoms up to certain size, and that's roughly half of our data set and another half is more realistic set of small molecules -- small drug-like molecules, but also to get realistic, but also challenging in synthetic accessibility compounds we decided to use a subset of natural products out of [ Campbell ]. So these molecules and that model led us to achieving very good -- predictive performance of our SaaS API, which ended up to be squared of 0.73 roughly and the mean absolute error of 1.14, which is really good. And compared to our competitor products, this is very useful because not only it says if the molecule is synthetically accessible, but also you have this degree of how difficult it is to synthesize. So it's up to the user to set that cutoff for themselves, and they might decide to go with a little bit more challenging molecules, but also can be conservative and stay with very easy molecules to make as you decide. Just to highlight 2 examples of how SaaS estimates the synthetic accessibility. Usually, the course that estimates the accessibility are really correlated with the size of the molecule. So obviously, the protection chemistry might be challenging for them. And in this case, on the left, the protected I mean is a part of the synthesis of the molecule to the right. So the SaaS was managed to learn that the protection of a mean is necessary in the synthesis of that molecule and therefore, it has estimated the accessibility accordingly. On the right-hand side example that TBOC protected, I mean, isn't part of the synthesis. Therefore, the model was able to differentiate that and protected, I mean, is estimated to be less accessible than unprotected one. And yes, that concludes my part about synthetical accessibility scores. And with that, we have another set of questions that we would like to ask you, and I wanted to hand off to Edgardo, who will cover our applications of SYNTHIA in Nanome. Thank you. Edgardo?

Sure. Hi, everyone. Yes. So thanks for joining. And as you can see, there's another poll question asking about your familiarity with the VR technology to date. So yes, let us know if you've tried it. Okay. Let's go ahead and see the results. Right. It looks like over 50% haven't tried VR yet and some have. So it's great. I'll give a little bit of an intro into how it's applicable to drug discovery. So my name is Edgardo, I'm one of the cofounders of Nanome and the Chief Experience Officer. And we've developed a collaborative molecular modeling tool that leverages advancements in spatial computing, software allows for anyone to interact with molecular structures in natural and immersive way. It has helped accelerate drug discovery and research efforts across pharmaceutical companies, academic researchers and educational institutions around the world. In this [Indiscernible], you'll see 2 of my colleagues here at the office, looking at the same protein while wearing an XR device. XR actually stands for extended reality, and it's an umbrella term for virtual reality, augmented reality and mixed reality. This particular example would be referred to as a mixed reality since both are able to see the real world and the virtual object in this case of protein structure. In Nanome, you have the flexibility to transition between VR and MR depending on the head of the issues. And yes, so we initially launched back in 2017, and we've come a long way since then. Let's back up and understand why we got into this into the first place. As we all know drug discovery is quite expensive, time-consuming endeavor. And so we looked into how we could potentially help scientists get key new insights faster. The one area that came to our attention was in structuring new drug discovery projects, where scientists were trying to solve 3 dimensional problems with 2-dimensional solutions. And so a good example of this is Professor [Indiscernible] from San Diego -- UC, San Diego Skaggs School of Pharmacy, where he talks about how even experienced crystallographers were able to gain new perspectives on structures that they've studied for years. You mentioned that the speed of comprehension was orders of magnitude quicker because when you grab a protein with your hands and enlarge the areas of interest, things become immediately obvious. So that's what we're all about helping teams make faster and more effective decisions. And we try to do this through an intuitive interface that you can pick up a protein just like you might pick up a water bottle to bring it closer to you and inspect it, and you can do this in a collaborative ecosystem with your colleagues in different parts of the world or even if they're physically next to you, while also integrating into workflows, internal workflows and leveraging computational tools or infrastructures that your team may already have in place. We've -- we understand the importance of integrating into many different workflows and build Nanome with a robust pipe on API that can communicate with internal computational infrastructures. If your team doesn't have one of these in place yet, Nanome also has its own computation infrastructure that currently supports several early to midsized biotech companies. We've been fortunate enough to have worked with many different organizations around the world, and they have published on different journals about their experience and results. If you want to learn more about those, feel free to visit our website, nanome.ai/publications. Another example I wanted to highlight was the Novartis case study, where -- they are actually one of our earliest customers back in 2017, and we're lucky to have found a visionary team at GNF, San Diego, that was willing to not only help our team understand the different use cases for early drug discovery teams, but also iterate on the product through close collaboration with the structural biology, medicinal chemistry and computational chemistry groups. Over the years, their use and the adoption of the application has evolved and spread across their multiple sites in North America and Europe. And one of the key factors for this was when they realized that a process that typically took 6 months was reduced to 4 months, which is a huge time and cost savings that is not uncommon across our other customers. So what are the key areas that Nanome can impact? As you can see, there are several stages in a project/cycle where Nanome can be leveraged starting with the kickoff, target validation, [ head identification ], lead generation, optimization and last but not least, of course, project reviews. Let's get into each of these in more detail. So during a project kickoff, you can bring together the experts, each with their own unique avatars, the molecular structures and relevant data to get a shared understanding of the nuances of the projects from the very beginning. As I mentioned before, this can include team members that are working remotely across different parts of the world. During the target validation phase, anyone can start to analyze and internalize three dimensionality of what -- of the project. In this case, you can see a my colleague analyzing chemical interactions in the binding pocket. But you could also review molecular dynamic simulations to see how changes in the structure occur over time. During the [ head identification ] stage, Nanome can support our SAR analysis, perform docking, generic conformer or review results from a virtual screen to aid in your compound inspections. During lead optimization, you can design new molecules, apply computations to your design, such as the one in the animation. And you can calculate chemical properties, leverage integrations like RD kit to review things like [indiscernible] and solubility or even synthetic accessibility scores with SYNTHIA. We've worked with scientists to understand some of the things that would bring the most impact to them, and one of those is a customizable plug-in system. And so if they wanted to change something or connect with internal tools that would be helpful for them. Again, our robust Python API system is in place to help support that. Scientists also leverage existing tools like [indiscernible], Mo, PIMO and Nanome can support those session files. We also have an online repository so that you can save and upload workspaces, recordings, your structures and easily access them from different devices as long as you can log into your account. And of course, we have a built-in web browser, which is one of the key things that helped in one of our collaborations with Genentech, where we also integrated one of their internal tools. And it was a great collaboration there. Our typical users include medicinal chemists, computational chemist, protein engineering, structural biologists and each of these different workflows are supported for those different types of people. One more thing I wanted to mention about the integrated customized integration is that it facilitates integration with tools like SYNTHIA because you can keep the synthetic accessibility score values in mind as you go through the process of selecting your best candidates. And we'll be going more in detail on this in a video after this presentation. The -- one of the great things about this is that you can then record what happens in that session through something called spatial recording specific to Nanome. And it's coming into the -- presenting to the leadership or internal teams. It's a great environment to practice your presentations. And because it includes your audio your avatar's movements and any modifications to the structure that you perform over time. And the unique thing about this is that when you're going through the playback of the recording, you can actually pause it and switch from viewing mode to interact mode, which allows you to then work the recording into your workspace so that you can apply further changes from that point forward. Yes. So we've worked with many organizations, and there's time and time again, we see how it's helped them save both time and resources. And here are 2 quotes that support that. "In a nutshell, Nanome is an intuitive way to explore to [Indiscernible] molecules, save time, efficiently collaborate with your colleagues to help them gain new perspectives, customize the platform to integrate with internal infrastructure, and it fits across many different stages of the pipeline." And with that, I'd like to transition to the video. [Presentation]

Excellent. Thank you so much to all of our presenters. Dr. Gardener, Dr. Wójcikowski and Dr. Leija, we really appreciate it. What a fascinating presentation, the technology is really advancing wonderfully here. So I'd like to open it up to some questions from the audience, we have a number of questions that have come in. So thank you so much. We have quite a bit of time left for questions, but probably will still take just about 5 or 6 in consideration of everyone's time.

So the first question looks like it could be directed towards Emma Gardener, is coding and learning programming language needed to operate SYNTHIA?

Yes, that's a great question. I'm glad someone asked that. So the short answer is no, especially with the web app, that's really designed for a synthetic chemist to use. So we've kind of done all of that coding and computer science on the back end, but made it pretty easy to use just through a web app. So you just log in with a user name and password on your browser, and then as you saw in the video, you're able to just basically take the structure of your molecule, hit start and then let SYNTHIA show you all of the pathways and different ways to make it and all that information there. So it's really meant to be navigated by the bench chemist, not necessarily a computer scientist. And really no knowledge of any of that kind of back-end stuff is required to use it.

All right. Great. Next question we have, it looks like it might be in reference to the synthetic accessibility score, Maciej, when it estimated the steps to synthesize a molecule are the steps completely reaction based or is any processing information embedded in the "Level of difficulty?"

So in the prediction that SYNTHIA makes in general, and this is not only specific to the API or SaaS, the steps that we see are reaction based only the steps like purifications and technical steps are not really accounted for here. Usually, they are also including so what SYNTHIA predicts includes protections and the protection steps that are required, but these within our definition are reaction steps. But to answer your question shortly, no, the technicals that are not included here.

Okay. Great. Does SYNTHIA provide reaction conditions such as solvent temperature, reaction time, et cetera. Now Emma I know you touched on some of these during the presentation, but maybe you could elaborate a little further.

Yes, absolutely. So SYNTHIA will always provide you with some general reaction conditions for the reaction that's proposed. So depending on this reaction, this can include information like a suggestion for solvent or if it's something that really needs to be heated, it will say that or if you need to do it at low temperature, it may say that. Now these are, like I said, the general conditions for that rule. So it's not necessarily optimized for those specific substrates. That's where you're going into the lab and trying different things, it really comes in there. But it at least gives you a place to start, and then what you can also see with those similar reactions is now some other published examples of what people who have done similar reactions have used. So again, giving you kind of additional information, some additional examples. And those may have -- obviously, those papers will have more detailed conditions there for those substrates. So it doesn't give you exactly these are the perfect conditions for your reaction necessarily, but it's certainly giving you some conditions that you have an idea of is this something that's got to be heated or cooled or an idea of a place to start.

Excellent. Okay. We actually have a couple of questions here for Nanome as well. So here is one, Edgardo, what is the level of technical expertise required to utilize Nanome?

So we tried our best to make it as intuitive as possible. And so barely just getting familiarity with the VR controllers so that you can grab the structures, like I mentioned in the presentation, grabbing anything like a water bottle. It's as simple as that to get started. And then when you want to get into some of the more advanced functionalities of the application, we do have a training program that can help all of our customers.

And actually, if I could add maybe just an anecdote there, you saw, we were in the video making a video and then kind of doing this collaboration with Nanome was my first time using any kind of VR. So it's pretty intuitive, pretty quick to be able to get in and do that. So just my perspective as sort of a nontypical VR user getting in and starting to do that.

Great. And that actually addresses another question, whether there are specific training or support resources available. So please do reach out to the Nanome team, visit nanome.ai as well. And then finally, there's a question, how to download this software? And I would assume that, that might be specific for SYNTHIA. I'll let Emma address that. But Edgardo, maybe you could also kind of help us understand that process as well for Nanome.

Sure. So for SYNTHIA, as I mentioned, this is a web-based software, so you actually don't have to download it. We do have a free trial version if you want to kind of play around with it. We've got our website on the screen. If you go to www.synthiaonline.com. You're able to make a free account there. If you want to kind of take a look, but yes, because it's all web-based, you actually don't have to download it at all.

Yes. And on the Nanome [indiscernible], if you visit nanome.ai/setup, that's where you can download the software.

All right. Excellent. So again, I just want to thank all of our speakers, that's all the time we're going to take today. So thank you to all of our participants as well for being a great audience and answering our poll questions. Our team will follow up with the unanswered questions individually at a later date. In the meantime, please visit our website synthiaonline.com to learn more as well as nanome.ai to learn more about Nanome or reach out directly. You can e-mail us at [email protected]. Thanks again, everyone, and have a great day. Goodbye.