Connecting Instruments, Data, and AI in the Orchestrated Lab

Hello, everyone. Welcome to the Dotmatics webinar today. My name is Ryan Bernhardt, and I am the general manager of Versidian and also one of the leadership members of the Dotmatics Luma team. Today, we will be discussing Luma lab orchestration and how we're enabling the connected digital lab. If if we just start historically at one of the challenges that that we all face in in this space, it's really about the journey of getting innovation to market. And more specifically, the journey from molecule to market can take over ten years and on average two point two billion dollars. We oftentimes use this statistic as it relates to medicine or drugs, but really getting any innovation from from concept or design all the way to the market can takes a significant amount of time and cost in doing so. We started at the at the very early stages with thousands of different drug candidates. And then as we narrow those down and and begin to think about scaling up and getting it into a, manufacturable state, there's a lot of different challenges that we experience along the way. If we look at this process, we we've seen that over the course of many years, manufacturing has has very much gone digital, but r and d remains hindered by outdated systems, disparate data that slow the pace of innovation. This this creates challenges at real as it pertains to fragmented systems and siloed data. We have trial and error in in approaching the the solutions with very labor intensive workflows. There's been limited adoption in automation and artificial intelligence. And overall, there's inefficient knowledge and and sharing as it relates to technology transfer, which causes compliant risks and and and and regulatory, hurdles. If we really want to be able to accelerate the path from molecule to market, we see, as part of the Siemens vision, this idea of being able to create end to end digitalization where we're able to create a digital thread from the design of the molecule all the way through manufacturing that at scale such that it can reach patients around the world. We see this digital thread being able to create data traceability and knowledge sharing along the way, which allows us to really streamline moving that molecule through the different phases and and and ultimately, getting it to market where we're enabling a seamless tech technology transfer as part of that process. If we take a more specific look at the research and development space where Dotmatics is is highly engaged, you may resonate with this image on the right hand side of the screen. And this may look like your laboratory where you have many different systems and instrumentation, but they're rather disconnected. They rely on the scientists or technicians or humans to stitch together. We have ELNs and LEMS systems, but we oftentimes have data that's, coming out on an Excel sheet or macros, where we, have to to to make those connections. We have an instrument sitting next to each other that don't know anything about the data or the samples that the other one is is is analyzing. And, overall, this creates an inefficient process where we have opportunities for a loss of experimental intent, challenges in being able to reproduce explain those results. And so if we're serious about streamlining scientific in innovation, we need a way to manage every aspect of the lab or or better yet, orchestrate it. And so when we refer to orchestration at dot matics, we're really referring to connecting scientific intent, data, and decision making by being able to bring together the physical and the digital and logical aspects of the laboratory, into a harmonized experimental process. And on the left hand side here, you can see many of those aspects. So the physical being the people, the instruments, the the tools that we're using, the materials, and the labware, the the digital being the the data systems, the sample information, and the logical really in referring to the SOPs and the experimental processes and and the batching capabilities. And those those aspects all start as inputs. They come together as part of the execution of the experiment, and and and out of that is the generation of results. And those results may come in the form of a a blockbuster molecule, but those those results may come in the form of being able to scale innovation or reduce cost because we can miniaturize a reaction and and use less of the material. It may come, in the form of being able to seamlessly reproduce that experiment, at different places around the world through that technology, transfer, transfer. Or it could be that we've been able to generate data in a structured and harmonized way where now we can apply AI such that we can we can learn and optimize as we as we go to the next iteration of that scientific process. And this brings me to the Dotmatics ecosystem. Over the last decade, Dotmatics has acquired a number of different industry leading scientific capabilities, that are used in labs all over the world. From GraphPad Prism, which is the most ubiquitous used statistical analysis software by scientists in the laboratory today, or the dotmatics, electronic notebook or a capability such as Versidian, which is a automated chromatogram analysis for small molecules that provides actionable intelligence. Many labs around the world have come to to rely and trust in these capabilities. But over time, we looked at our portfolio and said, how do we be how how can we bring these different capabilities together in a way where one plus one equals three, where we can get more out of this by being able to have that connectivity? And out of this was born our Luma scientific intelligence platform, which allows us to connect the different capabilities within the the dotmatics portfolio, together in a harmonized way, through APIs. But it also gives us the ability to make connections beyond our own portfolio of scientific capabilities to the data systems and and the the tools that your lab may be relying on to facilitate innovate innovation and execute experiments. And so the the Luma scientific intelligence platform is is really the heart of orchestration. You can think of it as a connectivity hub that brings together these disparate and and siloed systems and scientific capabilities. But Luma is also built on Databricks and and allows for the data harmonization and storage acting as a data lake as part of our AWS infrastructure. In addition to being able to connect and harmonize data from a variety of different dotmatics scientific cape capabilities, now as part of Siemens, we also have the ability to connect to capabilities within the larger Siemens portfolio such as GPROMs, which is a process modeling, capability that allows us to create digital twins and bring together the wet lab, dry lab paradigm. But maybe more importantly than our own, the connectivity of our own portfolio of capabilities is the plus at seen here at the bottom of this screen. Luma is built with an open and extensible open and extensible platform that allows connectivity not only to our own portfolio, but also to the data systems that your lab is relying on. And in addition to being a, connectivity hub, you can think of Luma also as a as an operating system that not only allows us to connect to other applications and data systems and tools, but allows us to create apps to make those connections, even with your own, home built, capabilities as well. And like any operating system, Luma has core processing capabilities. So I just mentioned that it's a it's an open extensible ecosystem that allows us to to create apps and make connections to different systems. But at the foundation of Luma is is artificial intelligence that is that is powering that at the foundation. In addition to that, we have the ability to to create visualizations for statistical analysis as well as dashboarding capabilities that allow for graphical user interfaces. We also have material registration capabilities as part of Luma. You have the ability to embed, business rules and data management rules, and we can create, adaptive scientific workflows that are digitalized for your laboratory environment. In addition to being able to connect to different data systems and and tools, our LumaLab Connect application allows us to make those connections, even further down into, stand alone instruments, integrated work cells, transportation devices, and it also allows us con to connect to proprietary file types where potentially had data that's been locked away for many years. We're we're able to go in and unlock that data, transform and harmonize that data in a way that's now connected with more modern technologies as well. One example of this is, is at a large pharma, customer where we're using LabConnect to make connections to over three thousand alone instruments across the research campus. On a daily basis, we're we're capturing over five terabytes, of data per day. Over five hundred thousand files are captured, per day from those instruments, which equates to billions of scientific data points. And we're able to do this by being built on on a Databricks enterprise infrastructure. In addition to being able to make those connections to stand alone instruments, we also can seamlessly integrate one or many different scheduling softwares and instrument controls, together simultaneously through our REST APIs. And beyond the connection to instruments and and transportation devices, integrated work cells, we also can make connections through three common data ingress patterns. We can make those connections to file types where we use our our file parsers to be able to go and and and really parse and unlock data from those files, transform that, and harmonize that in a structured way. We also have the ability to make connections to other systems through API calls, and we can also make connections directly through databases such as SQL Server or Oracle. And for those of you that already are using platforms built on Databricks, we can also leverage the seamless Databricks date Delta sharing for being able to pass data bidirectional bidirectionally in a seamless way. So one way that we are able to do this, if we dig into the unlocking data from files, we use our LumaLab Connect to be able to make those connections. We can we can capture the files. And then using, file parsers, we're able to go in and actually, parse and unlock data from those files, and and be able to then transform it into and and harmonize it into Luma. Now we have a variety of different file parsers that can be used, to to, capture data from hundreds of instrumentation, whether that's a a common file type or a proprietary file type where we can use proprietary file based parsers to go in and unlock that data in certain formats. Once that has been transformed and and harmonized, within Luma, it's now available for being able to to store long term, to be able to visualize as part of the, that user interface or statistical analysis, or we can take that data and actually, you utilize it as part of scientific workflows. And then beyond that, we now have that data available for being able to apply AI to the entire record. We can use use it to analyze and decide what to do next in in either a manual or automated format through the use of business rules. And it's available for audit trails as well as more of that long term storage and and being able to answer questions over time. So what does this look like when we have a connected e ecosystem where we're orchestrating all of this together as part of the Luma platform? So this is an example of our Luma Cynthia integration, which is really enabling a sin automated synthesis design, make, test, analyze cycle. As you can see there, we've we've drawn a compound of interest. We're looking at how could we, what are the pathways that we could use to to to, synthesize that compound. We've made the submission, into, the process, and and now we've we've, been able to go out and look at possible pathways. You can see there's about ten different pathways that have been brought back to us and utilizing business rules of being able to look at the availability of building blocks through reagent providers, the the time and the cost of getting those reagents, as well as the the number of, potential steps that would, would be required, within that pathway and the likelihood of success, it's the it's given a path score, and then we can drill down into those different synthetic routes, and look at those in more detail. We can even dig down further and look at each step synthetic step in in that multi step synthesis where and and once we're we are satisfied with that being the route, we can then generate a written protocol that's available now for, chemists to to follow. But it's also, available in natural language, so we can now import this directly into our ELN. And then we can go one step further, and we can actually take that natural language protocol and and generate a, a digitalized protocol from that, which is now ready to be passed down to your laboratory environment for, automation or semi automated, execution of that workflow. I wanna take a minute to to drill down a little bit further into, the artificial intelligence strategy that's that's really been built, at the foundation of our Luma platform. Luma is is one of the the only, platforms today that's really been built with AI and automation at the forefront of, of the offering. And so our AI in Luma, it really acts as a as an assistant or a coscientist, if you will, where it has been built for scientific scrutiny, where we are ensuring that it's traceable, explainable, and verifiable. Being that this is used for sign scientist, it is very important that the, AI has a deep learning model where we can we can actually drill in to seeing, how did it come up with those those, answers. We can optimize or refine that. And and also that allows us to, minimize the hallucination potential, and and increases the reproducibility, that's needed in the scientific, arena. But before we dive into, AI paradigms and strategies, I wanna just start by, going back to saying, really, in order to get the most out of AI, it's all about the data. Artificial intelligence does a great job of of trying to make sense out of unstructured data, but it does an even better job of making sense out of structured data. And so, that is really a key principle of our of what we do within our, Luma scientific intelligence platform. We're able to go out and harmonize that data from disparate sources and systems, and then transform that in a way where it's now, ready for, the applying AI. And that's really at the at the at the heart of, Luma is being able to capture, contextualize, and then transform data in a fair AI accessible format. So once we have our data harmonized and structured within Luma, we can now apply the we can now apply the different paradigms of of AI. So starting at really the the simplest form of of AI would be the use of assistive AI. And most of us would know this from the the way we use ChatGPT or or other AI on on the market. It would it's really assisting users with repetitive repetitive tasks such as building queries, or app configurations. So we could say, find all successful protein expressions with c d r three matching a particular amino acid string with their associated expression backbones in the last two weeks and ranked by liability type and count. And what you see on the the bottom right here is actually the the the response from from our Luma AI that's going back and and and looking at all of the the data that's available to to be able to answer that question. Now one thing I I highlighted as part of this assistive AI is not only can it can it ask and answer questions and and build queries, but you can actually use Luma AI to build app configurations for the connectivity to to the your own tools or commercially available systems. You can also use AI to to to build workflows and data models such that you're not having to to program that yourself within Luma. The the next, really ring of of AI strategies would be, to apply domain focused AI, which this is really about predicting highly specific scientific outcomes. And and I mentioned as part of the Dotmatics ecosystem that Dotmatics has acquired, many different, industry leading scientific capabilities over time. And so by being able to, have access to this this highly experienced scientific, capabilities, we now can leverage the AI within Luma, to to apply, really domain focused or scientific focused assistance. And so in this case, you're seeing at the bottom right, an example of where we're asking, we're asking, AI and Luma to auto gate, a a population of cells from a flow cytometry application. And then really the holy grail of of the use of AI is really in this composite AI mode. And it's referred to as composite a AI because it it can't use any one model of AI to answer a question. It it has to, compile or it's a composite of many different AI models that are being used in order to be able to answer a question. And so an example of this, would be to say in a lead optimization scenario, would would be able to to ask LUMA to optimize antibody efficacy and developability via machine learning trained, on correlated amino acid sequence properties, homology, and assay datasets. And Luma, AI and Luma would be able to go through, look at all of this different data that's been harmonized from, from various sources and different modalities and bring together this predictive outcome and and then be able to provide that deep learning so that we as scientists could understand, you know, how did how did we actually arrive at this at this conclusion or this prediction. And one thing I'll mention, in addition to Luma having an open extensible ecosystem for making connections to data systems and instrumentation and and scheduling softwares, We've we've also taken the same approach as it pertains to, AI. So, you can use the, our AI in Luma, which is is based upon a a Claude, foundation. But you can also point Luma to leverage your own AI paradigm or strategy, that is best for your organization. And instead of using using ours, it can it can leverage, your own a AI paradigms as as part of the Luma platform. So other things that that Luma AI can assist with as your coscientist is it is that it can also help build out your scientific workflows. So you can leverage AI to map out resources, and so this would be mapping out your, instrumentation, your integrated work cell capabilities, your sample reagents material as part of the inventory that you have access to, and also the data systems that you're connected to in terms of, you know, which information is is coming from from which of those data systems. Illumi AI can then assist with creating the data pipelines or or the the flow of data, as part of your process. So being able to stitch together, provide the necessary data and connections at each step in that experimental process. And then finally, Luma AI can can generate a written and a digital did digitalized workflow that is now ready for either manual or automated execution of that experiment. So what does this look like within our Luma workflow builder? So if if I just take a case a use case of something that we probably can all equate to regardless of the type of science we're doing, this would be one more from more what we do in in a kitchen. But if we if we just took an example of the the workflow of baking a cake, I could actually feed this a recipe in a natural language protocol on the left hand side here. And Luma can then take this protocol, mapping it against your laboratory environment to to generate a digital digitalized protocol or workflow that is now making those connections to, in this case, what ingredients you have access to, when those ingredients are needed as part of your workflow or recipe, and then and then the order of addition that that that will happen as well as the instrumentation or capabilities that would be necessary, in this case, an oven or a microwave, and what types of labware or pots and pans you would have access to. You can imagine now how this would translate to to a scientific application where it can map the the different data that that needs to be present at each step in the process, the instrumentation or or work cell that you'll be performing a task on, and and and the labware that that would be needed for that. Much like we see with use of AI in our own, personal, life, you you typically get it, you know, ninety five to ninety nine percent of the way there, in this case, the workflow. But you can certainly go into this workflow and say, actually, I'd like to rearrange the order of addition. I'd like to add vanilla before eggs. And you can make those changes from the the the, generated workflow without having to build this from scratch. If we take a use case and look at how this translates into that and and taking a common use case of kind of an analytical workflow where maybe we're going from an experimental design of some sort of reaction optimization or reaction screening. You you would start with experimental design. You'd then send the the appropriate orders or or protocols to instruments. There'd be instrumentation processing. You'd have that data that's being generated, off of, say, the LCMS going into maybe a CDS system. It's preparing files for analysis, and then your the scientists would be looking at that, making some decisions, and then being able to kind of determine what happens next to to those particular samples. But but then passing that data or entering that data into an ELN or some sort of data management system. And if you look at each step along the way, you you see kind of in the red bars here, the idea that each one of these steps has many different kind of manual touch points or connections that need to be made. Experimental design, for example, may require the use of many different tools in order to to to generate that that experimental protocol. Anytime we have a human in in the loop or making making a decision, you have subjectivity in is part of that. And then you also are losing some of the context and and decision making kind of capabilities as part of that process. And so what would this look like then in a Luma orchestrated workflow for that same, that same process? And so in this case, what we're able to do is, streamline, this entire process in a in a orchestrated way where Luma is orchestrating this, and we're able to leverage other data systems and capabilities such as Versidian here to really create a completely automated and frictionless workflow where a scientist is part of it, but all of this is happening in a frictionless way where we're capturing, contextualizing, and harmonizing data and and decisions that are being made at each step in that in that overall workflow. And and so what does this look like as part of kind of that Luma orchestrated process? Well, if we take this the same workflow that we just walked through, and and look at okay. Now we have the generation of our chromatograms that are that are coming out. They've been captured, and and harmonized now, within Luma where we are you have visualization capabilities now within Luma of all of this data in a in a centralized location. We're able to click on the different compounds that have been identified in the peaks, and it shows us the peaks for each of those, the the quantification. It will show us information about the impurities. And we're we're now looking at all of this data in a single location that's that's coming from different systems. So coming off of LCMS, ELN, LIMS, data reduction tools such as Versidian, and we have this in a single location where we're able to visualize and make decisions as part of that seamless process. If I take a step back just to to orient and go back to really our definition of the orchestrated lab, I'd like to talk a little bit about the approach, the paradigm that we at Dotmatics are are making to to really enable this. And if we we go back to the definition of lab orchestration really being the the this this stitching together of the physical and the digital lab, we we start with a look at the physical domain of of the laboratory. And in this case, we're talking instruments. We're talking potentially, you know, mobile robots and barcode scanners and labware and integrated work cells with scheduling software. If we talk about the the digital domain of the laboratory, we're we're talking about things like ELN and LEMS, inventory management systems. We're talking about having a data mesh or or a a data storage such as AWS or Azure or cloud Google Cloud, data analytics, things that we're commonly being able to use with Power BI and Spotfire and Tableau, the ability to visualize this through graphical interfaces and and visualize our our lab processing as as experiments are in motion, having data processing capabilities, and then being able to have a layer that's really doing data capture and and integration. And as part of this, it's really in order to to to bring the physical and the digital, aspects of the lab together, it it's really about also bringing together the logical aspect of the lab, which is creating kind of the glue or the gel that applies scientific intent or experimental intent to the connections between the physical and and the digital domains. And so the logical layer allows us to to apply experimental intent, business rules, SOPs, and techniques, and and batching decision making. And and if we look at, you know, how this is being addressed by others in across the industry, we see that more of the classical lab automation players have really approached this by coming at orchestration from the ground up or the physical up where they've started at the instrumentation level and kinda moved into scheduling software and then a little bit higher into more of that that, larger workflow, layer. If we look at where others, that are more modern scientific data platform providers have approached this, they've really focused on certain aspects or pieces of that digital domain, whether that's at the ELN or data analytics or or getting into more of that data capture and integration layer. But there is a an advantage that we see of of of our position in this space, and that is being able to come at this challenge with a solution that really approaches it from the the digital side of the equation or or the digital domain within the laboratory and being able to also leverage many of the other scientific capabilities that would be be part or necessary to really enable true laboratory orchestration. And this is exactly where the the Luma, offering really, really focuses on is what used to require many different, scientific capabilities and and different providers, allow we're able to approach lab orchestration with a scientific intelligent platform in Luma that really allows us to, to accomplish many of these capabilities within a single offering. But by having an open extensible ecosystem, we can also make those connections to the, tools and and the capabilities that you're using today. And then we have the the the logical layer, is being applied through our workflow builder, that allows us to connect to down to the physical, whether that be standalone instruments or integrated work cells being controlled by scheduling software or transportation devices, we're able to seamlessly integrate those through a rich API and application structure. So this allows us one orchestrated system that's, for science enabling experiment insight. Luma offers a orchestration platform that is multimodal. So we support diverse application areas with foundational tools, that orchestrate discovery. Our Luma platform is truly an enterprise software, so it can be used by thousands of of scientists in multiple locations all over the globe, or it can be used by three scientists in a single lab in a research building. It also, allows us to harmonize data, in in a in a in a structured format built on the infrastructure of Databricks, which empowers the ability to utilize artificial intelligence and machine learning in a way, that is impactful to scientific innovation in the future of discovery, development, and manufacturing. And with this, we are realizing and enabling others the the ability to streamline delivering molecules to market quicker and more cost effectively than ever before with the emission with the mission to improve lives. And it's really about Luma creating the digital thread that allows us to go from design and discovery all the way through the scalability to manufacture in a way that meets the demand across the world. So with that, I wanna thank you for your time and attention and joining us today, and I'd like to open it up for q and a session. And feel free to put any questions in the chat. Looks like we have one question around AI models in Luma, the political sequence. Yeah. Great question. It around being able to capture if we're with the use of different modalities, being able to capture variations in datasets, biases. Yeah. Great question. And that is certainly something that we are we're looking to do as part of part of the AI strategy that we're implementing. We actually have a whole team that's dedicated to AI, not only at the dotmatics level, but also now as part of the larger Siemens portfolio. And so one one of the one of the key strategies of of AI at the foundation of Luma is that utilizing it in a scientific domain where, you know, it's really, really important as scientists. We want we want the the datasets. We wanna understand how things were were were decided and then be able to to kinda learn from that. So that's a big part of the strategy. And and and, certainly, there will be more and more models and different paradigms of AI. It's it's hard to even keep track of of the different AI solutions and technologies that are that are coming out on nearly a weekly basis at this point. Awesome. Hey, Ryan. This is Hailey. I'm jumping on to help to moderate this section as well. And we had a few other questions coming in too, and I thought I could read it out just so everybody is able to see it. And so or to hear it. And so the first one is thanks so much for that first question we already had. The second question is, does this layer onto our ELN? Is there professional services engagement, or is it more self-service? Does was that question specifically around AI or around the Luma platform? About Luma and, like, how Luma is interfacing with the ELN? Yeah. Great question. So Luma is is designed to integrate seamlessly with your ELN, whether that's a dotmatics ELN or another ELN tool such as signal signals or benchling or or something else, even a even a a home built LIM system. So it integrates seamlessly into that. Now you you may envision a future where the capabilities that have been relied upon historically for an ELN, those capabilities do exist within the Luma ecosystem as well. So I mentioned Luma has a material registration capability really as part of its core. You could envision a future where maybe Luma actually can can can suffice as that LIMS or ELN type of of system. Awesome. Okay. And then we have one more question in the queue. So if anybody else has any burning thoughts, feel free to get them in there. But, Ryan, here's the last one that I've got on the list for right now, which is we have a mixed mixed vendor environment with instruments of different ages and configurations. How do I know if my instruments are compatible? Is there a supported instrument list? Yeah. Great question. Yes. We have, we have a supported instrument list that we're happy to make available. We have hundreds of instruments already in our in in in our library. In addition to that, we are adding new instruments and on on a nearly daily or weekly basis. And, you know, one of the approaches to to the Luma strategy was about having that open extensible ecosystem, knowing that there's there's going to be a lot of different systems, technologies, instruments that will be coming out both now into the future, but also that it's certainly not one size fits all. Every lab has preferences and differences and and reasons and rationale for why they they use what they use, whether that's a brand or or a type, and we wanna make sure that all of those capabilities can be connected into this orchestrated lab through the Luma Scientific Intelligence platform. So there's been a ton of emphasis in the way this has been designed that it is a rich API, and and it's available so that you can leverage that. I also mentioned the app structure. So Luma does act as an operating system for science, and so you can think of it as, like, iOS, Android, Windows, where you can have you can create Luma apps, and we we provide the architecture to create Luma apps, not only for commercially available products or systems, but also for maybe your own in house systems as well. You you can connect those that way as well. Okay. And then while you were talking, we had a few more questions pop in. So next one up is, what does a typical implementation look like? If you had to describe the sweet spot for Luma, what would that be, and where could it be a bit of overkill? And by implementation, I meant the project process rather than the end state. Great question. So the the you know, Luma as a platform, there is there's core functionality, and then there's the ability to add on a lot of different kind of modules, capabilities, even even going beyond into more of the the the larger dotmatics portfolio. But the implementation in terms of the Luma platform is is pretty seamless to get that stood up. And I meant I I showed as part of that digital physical divide, there's a lot of capabilities that come as a standard part of Luma. It it's not a requirement that you use our own, you know, Luma statistical analysis processing capabilities. You could use your own with a third party like dot or like Tableau or Power BI, but you certainly also have the ability to have to leverage the capabilities that are already part of Luma. So, we we have a base base platform, but we would also scope out your your specific, project as part of a scope a statement of work, and then we'd be able to deliver that. And that could be as simple as standing this up as really a, you know, data aggregation, harmonization, and visualization process in your laboratory, or it could go as far as being able to create end to end fully automated workflows that are that are really tying in every one of your data systems and your instrumentation of work cells. You know, that would that would be a much more detailed and and extensive project as opposed to just standing up Luma. The good part about Luma from, like, when is it overkill is it's totally modular. So you can start with a very small implementation, and over time, you can expand the the connections. You can expand the utilization, the use of that, not only with data systems, but also down to the physical of being able to make those instrument and and work cell connections or even process things in more of a semiautomated workflow at at with stand alone instruments at the LabBench. Okay. And then I just I wanna thank the audience so much for putting your great questions in here. There's two more that just popped in. So, Ryan, I'll put them to you. As long as we have time able to chat. So the first one up here is, is Cynthia the only option route for scouting with Luma, or can I integrate Reaxis and more? Great question. So we are certainly agnostic to the capabilities, and there's a lot of there's a lot of ongoing projects and and implementations and and partnerships happening right now. So Cynthia is not the only capability, but it's it's it's one that is is is, you know, being launched and and out there right now. So, yeah, happy to take any and all considerations and opportunities from the tools that you're already using. Again, that's a that's a key part of of this is making sure that we are applicable to the tools that you're relying on. Okay. And next one. Can Luma replace statistical analytical software like JMP or Prism? I don't understand where Luma ends and analytic tools begin. Great question. So we this the we're using Sigma as the statistical capabilities underneath. So you can leverage Luma for a significant part of your statistical analysis and being able to create charts and graphs and and all of that. Will it will it replace everything? Depending on the specificity of what you're doing. So, like, Prism has things like curve fits and different capabilities where the way we would leverage that in Luma is it would be a Luma Prism integration where we can leverage those capabilities. For JMP, you may be using it for experimental design or design of experiment where you're looking at factorials and very specific things that that that capability does really well. So we we can both make those connections with other systems that offer a specific capability, or we can leverage the the cape we can leverage and recreate those capabilities inside of Luma. And, you know, you may start today with utilizing something that you're already using, and over time, you you you find, hey. I can actually do that within Luma. You know, I can I can now kind of sunset that other product or tool that we've been using to now move over to to to leveraging Luma for for those capabilities? And anytime you can do that, obviously, there's, you know, the the ability to have lesser risk for integrations and and supporting multiple systems and things over time. So, you know, there's there's certainly advantages of having that possibility from not only a cost standpoint, but from a support ability standpoint as well. Okay. And with that, I haven't seen any new questions come in. So I just wanna say thank you so much, Ryan, for taking the time to present and share this knowledge with us and then also to answer these questions. And yeah. Thanks so much. You're welcome. Thank you so much for helping, and thank you all for making time to join. And I look forward to future conversations and and having the opportunity to help you in orchestrating your lab. Thanks, everybody. Have a great day.