Lab Automation and Digitalization: Redefining How Labs Operate

LAB AUTOMATION AND DIGITALIZATION: Redefining How Labs Operate How Is AI Speeding Pharma’s Journey Toward Precision Medicine? Innovations in LIMS for Enhanced Laboratory Efficiency Audit Trail Requirements for a Digitalized Regulated Laboratory Credit: iStock/gorodenkoff CONTENTS 4 Audit Trail Requirements for a Digitalized Regulated Laboratory 10 Will Labs Wake Up to the Environmental Cost of Data Crunching? 13 How Will Automation and Digital Technology Shape the Lab of the Future? 19 How Is AI Speeding Pharma’s Journey Toward Precision Medicine? 23 Cutting-Edge Advances Driving Developments in Robotics 26 Innovations in LIMS for Enhanced Laboratory Efficiency 33 Revolutionizing Laboratory Information Management Systems With IoT and Smart Technology 37 Advancing Mass Spectrometry Data Analysis Through Artificial Intelligence and Machine Learning LAB AUTOMATION & DIGITALIZATION 3 TECHNOLOGYNETWORKS.COM FOREWORD The rapid evolution of scientific research demands laboratories that are not only more efficient but also more intelligent, connected and adaptive. Automation and digitalization have become essential in the modernization of laboratory environments, offering unprecedented opportunities to enhance precision, reproducibility and productivity. This eBook presents a comprehensive overview of the technological innovations redefining laboratory operations. It explores the integration of artificial intelligence in areas such as drug discovery and data analysis, the deployment of advanced robotics to automate complex workflows and the transformation of laboratory information management systems (LIMS) into interconnected, IoT-enabled platforms. This eBook also examines the energy implications of data-intensive research and highlights emerging best practices in sustainable computational science. Through expert analysis and real-world examples, this compilation provides valuable insights into the tools and strategies that are shaping the laboratory of the future. The Technology Networks editorial team 4 LAB AUTOMATION & DIGITALIZATION Audit Trail Requirements for a Digitalized Regulated Laboratory Mahboubeh Lotfinia and Bob McDowall, PhD As regulated GXP laboratories undergo digital transformation from paper to electronic records, regardless of the media (electronic, paper), it is vital to ensure attribution and traceability of actions performed by both users and system. In a paper environment, creation or changes to a record is evidenced on the paper by initials of the analyst and verified by an independent reviewer. Transition to electronic records means that trustworthiness and reliability of activities must be ensured through generating time-stamped audit trail(s) (AT) where all entries must meet ALCOA++ criteria. Therefore, audit trail design is critical for any digitalized laboratory to reconstruct history of the course of any GXP activity. There are nine main audit trail(s) design elements and functions that suppliers must incorporate into their application for effective use in any digitalized regulated laboratory, see Figure 1. Credit: iStock/ismagilov TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 5 1. Audit trail cannot be turned off The first design requirement is a technical control for any audit trail to work from installation of the software and cannot be turned off. This is to eliminate any possibility of users or system administrators turning an audit trail off to hide data falsifications or deletions. The ability to turn audit trails off and then back on again is cited in warning letters and 483 citations. 2. Database or flat file? Implementing an application where data files are stored in directories in the operating system is not a recommended option for an effective audit trail. Where this has occurred, AT entries were stored within the data file itself. The problem is that these files were vulnerable via the operating system and, if deleted, the audit trail did not record the deletion on the file’s journey to the recycle bin. Therefore, the only way to design an effective audit trail is to use a database that monitors the whole system. Also, any application should be designed so that there is no backdoor access to data to prevent falsification via administrators and avoid audit trail entries. 3. Single audit trail or separate system and data audit trails? It is important to point out that the lifetime of e-records and data usually exceeds the lifetime of the computerized system that generates them. From a system design perspective, there are two alternatives for audit trail design: A single AT covers all activities within a system such as system configuration, user account management, user log on and off, instrument connections and any incidents, plus data acquisition, modification and, if allowed deletion. This is good at giving an overall big picture, but unless there are effective search routines for second person review, quality oversight and data integrity audits this apparently, simple approach has problem of separating the wood from the trees. Another consideration is the analytical instrument acquiring data. A polarimeter is an instrument where data collection is relatively simple with little data manipulation. Compare this with a complex instrument such as an LC-MS-MS which will generate much more data which can be subject to interpretation of the data by a user. FIGURE 1: KEY AUDIT TRAIL DESIGN ELEMENTS AND FUNCTIONS. CREDIT: BOB MCDOWALL 1. Audit Trail Cannot Be Turned Off 3. Single AT or Separate System and Data ATs? 5. Time Stamp Detail Audit Trail Design Elements & Functions 7. Predefined Reasons for Change 2. Database or Flat File? 4. Contemporaneous Recording of Changes 6. Date Stamp Detail 8. Documenting an AT Review 9. Archive and Restore TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 6 A separation of entries between system and data audit trails is a far better approach that permits more effective audit trail review as well as efficient archive and restore. Figure 2 shows the audit trail coverage of a system and data audit trail. The data audit trail focuses on the data life cycle and will be subject to second person review and data integrity audits. The system audit trail records the events of system configuration, user account management, instrument connections and operational status and is subject to data integrity audits and periodic review. Both audit trails should have functions to record on-line reviews. 4. Contemporaneous recording of changes One of the ALCOA++ criteria is contemporaneous recording of data changes coupled with the corresponding entries in the audit trail. Q12 of FDA’s Data Integrity and Compliance with CGMP guidance says: Draft issued in 2016: … it is not acceptable to store data electronically in temporary memory, in a manner that allows for manipulation, before creating a permanent record. Electronic data that are automatically saved into temporary memory do not meet CGMP documentation or retention requirements. FIGURE 2: DESIGN FOR SEPARATE SYSTEM AND DATA AUDIT TRAILS. CREDIT: BOB MCDOWALL Application Audit Trails System Audit Trail Data Audit Trail • User log on and off etc (mainly for DI investigation) • Identification of user or system • User account management • System configuration events with old & new settings versus change requests • (Option for reason for change) • Instrument identity • Instrument connection status and problem • Date and time stamp • Search functions for key events or changes • Record of system AT review Statement that changes are OK Resolution of problems • Identification of user or system action • Covering complete data from creation, modification including peak reintegration, results, e-signings & (deletion, if permitted) • Reason for change • Old & new values Instrument identity and qualification / calibration status • Date and time stamp • Search functions for key events • Highlight modified data • Record of second person data AT review Statement that any data changes are OK Resolution of problems FOCUS • Periodic reviews of computerised systems: Demonstrating systems are under control • Quality oversight • Regulatory inspections • Data Integrity audits and investigations FOCUS • Second person review of batch records • Quality oversight • Data integrity audits and investigations • Regulatory inspections TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 7 This was modified in the final guidance issued in 2018: ... For example, chromatographic data should be saved to durable media upon completion of each step or injection (e.g., peak integration or processing steps; finished, incomplete, or aborted injections) instead of at the end of an injection set, and changes to the chromatographic data or injection sequence should be documented in an audit trail. Aborted or incomplete injections should be captured in audit trails and should be investigated and justified ... Both quotes are valuable for understanding how data changes should be identified and recorded as they occur in an audit trail. Although focused on chromatography, it is applicable to any laboratory data system. 5. Time stamp detail Again, an ALCOA++ criterion is contemporaneous. We have split the time and date stamp discussion into two. An accurate time stamp is vital in determining the sequence of events in a computerized system. Time stamp accuracy was addressed in the FDA’s withdrawn guidance on time stamps as within a minute, which can be interpreted as ±30 or 60 seconds. Time zone is also important in global systems and an additional time stamp of UTC / GMT (Coordinated Universal Time / Greenwich Mean Time) can be used to verify consistent and sequential actions in a global digitalized workflow. However, there is no regulation or guidance document that states or suggests the detail of the time stamp itself. There are three possible options: 1. HH:MM 2. HH:MM:SS 3. HH:MM:SS.X(X) The time stamp can be either a 12 or 24-hour clock but the former requires AM or PM to be added, however the latter option is unequivocal. Option 1 is useless as many activities can occur in a minute. Option 2 is a possible option in a standalone system but if several activities occur in a second, it is only the order of AT entries that can infer the order of activities. Option 3, where time is recorded to 1/10 or 1/100 second, is better for multi-user systems. An issue arises in regions where there are summer / wintertime changes. FDA’s 2007 clinical guidance notes: There is no expectation to document time changes that systems make automatically to adjust to daylight savings time conventions. If used, ensure these automatic adjustments are in your specifications. The sequence of time stamping activities in any system must be understood so that a clear explanation can be given in audits and inspections. 6. Date stamp detail Completing the time stamp is the date format; there are several different date formats that could be used: ∙ DD-MM-YY(YY) ∙ MM-DD-YY(YY) ∙ YYYY-MM-DD ∙ DD-MMM-YY(YY) Adding the day of the week is an option in some systems. Regardless of the format selected it must be communicated to all, understood and be consistent throughout an organization, especially global ones. Controls should be established to ensure that the system's date and time are correct. The ability to change the date or time should be limited to authorized personnel, and such personnel should be notified if a system date or time discrepancy is detected. Any changes to date or time should always be documented. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 8 For an accurate time and date stamp, a network has a timeserver from which all the active equipment on the network is synchronized. In turn, the timeserver is synchronized with a time source typically a national observatory, a network time protocol (NTP) server or global positioning satellite (GPS). One last discussion point about combined date and time stamps is the recorded time and the presented time. The system may record the time as UTC and the operating system may present that in local time. An alternative in some systems is to record two time stamps: local and UTC. You should understand how any system records date and time before purchase as this might have a bearing on validation and routine operation. 7. Predefined or configurable reasons for change As seen in Table 1, regulations and regulatory guidance require a reason for change. EU GMP Chapter 4 requirement when making a change is a reason should be added as appropriate as this may be obvious if using paper, unlike GLP regulations. In contrast, Annex 11 focuses for computerized systems where a transcription error or change is not obvious; hence, the reason for change is mandatory. When working electronically a reason for any data change is critical for traceability, integrity and trustworthiness. TABLE 1: GXP REQUIREMENTS TO REASON FOR CHANGE Discipline Requirement GLP Any change in entries shall be made so as not to obscure the original entry, shall indicate the reason for such change … 8.3 5. … Reason for changes should be given. 6.6 … Reason for changes should be given and recorded GMP 9 … For change or deletion of GMP-relevant data the reason should be documented... 4.9 … Where appropriate, the reason for the alteration should be recorded GCP 6.2.1 … The audit trail should show … where applicable, why (reason for change) MHRA 6.13 … The reason for any change, should also be recorded PIC/S PI 041-1 9.6 … what action occurred, was changed, incl. old and new values; … why the action was taken (reason) WHO TRS 996 Annex 05 it should be possible to … and a reason for the change recorded where applicable TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 9 In our view, the only configuration required for any audit trail is if it is silent (no reason for change required e.g., typically activities at a system level or method development activities) or a user is forced to add a reason for changes and modifications to data. Software functionality should offer the ability for a laboratory to add predefined and context sensitive reasons for change. This has the advantage of speed and consistency of reasons for change, avoiding users typing the same reasons every time and if further input is required then a free text option could be used as well. 8. Documenting an AT review A digitalized laboratory must eliminate paper. To achieve this, any audit trail review must be documented within the computerized system by the reviewer. However, this is a major failing of audit trail system design as very few applications have this functionality, meaning that the review will be recorded on paper. 9. Archive and restore The design of audit trail has significant impact on the ability of archiving and restoring e-records including all associated metadata. Summary An audit trail is critical in a digitalized regulated laboratory to ensure trustworthiness, reliability and integrity of electronic records. We reviewed GxP regulations and guidance documents to identify important functions of audit trails and their design. Although the regulations are consistent, implementation of audit trail(s) in an individual system varies greatly leading to the requirement to have separate SOPs for the effective review. System design can impact the ability to archive and restore audit trails as part of record retention. Read the full version of our article for much more detail about audit trail requirements for a digitalized regulated laboratory. ACKNOWLEDGEMENTS We thank Monika Andraos, Akash Arya, Peter Baker, Markus Dathe, Eberhard Kwiatkowski, Yves Samson, Paul Smith, Christoph Tausch and Stefan Wurzer for their constructive review comments in preparing this listicle. 10 LAB AUTOMATION & DIGITALIZATION Will Labs Wake Up to the Environmental Cost of Data Crunching? RJ Mackenzie After years of sounding the siren, researchers concerned about lab sustainability are being heard at the highest levels. Many universities and private research companies now incorporate sustainable practice into their work and funders like the UK’s Wellcome Trust are basing their awards on evidence of environmental certification. Key to these changes have been the work of standards bodies that seek to benchmark labs’ green commitments. These include the non-profit My Green Lab and the grassroots LEAF standards. These schemes focus on tangible changes wet labs can make to their practice. Researchers who work with biological samples can reduce carbon emissions by turning their energy-guzzling ultralow temperature (ULT) freezers from -80 °C to -70 °C, with no loss in sample integrity. Those who use fume cupboards can install energy-saving alarms that signal when the hood has not been properly sealed after use. These are important and obvious changes. But these standards largely ignore a key and growing source of lab emissions: the energy-intensive business of computational science. Does going digital always mean going green? Digitalization of waste-intensive lab processes usually gets the green light in discussion about sustainable research, said Loïc Lannelongue, a researcher at the Credit: iStock/ Nikada TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 11 University of Cambridge who studies green computing. “We’ve always thought of digital as the green option. Therefore, we don't really think about the impact.” In 2020, Lannelongue was completing a PhD in health bioinformatics when he decided to work out what the carbon footprint of his computational work was. “I thought that would be a small project,” he said. The first sign that this side quest would become something far more substantial was Lannelongue’s discovery that there were no helpful calculators of carbon footprint available. The only tools he could find were used to weigh up the impact of deep learning analyses that bore little resemblance to much academic work. In response, Lannelongue cofounded the Green Algorithms Initiative, which incorporates a calculator that assesses the carbon footprint of a computation as well as advice for greener computational science. Things “snowballed” from there, said Lannelongue, who now heads up his own lab focusing on green computing. The cost of on-demand data crunching For decades, the computing heft needed to power scientific research has been largely sectioned off from the scientists using it. The buzz of a ULT freezer – which uses 16 to 22 kWh of energy every day – is hard to ignore in a modern wet lab. The scientists in that same lab won’t hear the hum of server racks in their university’s data center, which use 96 to 144 kWh per day. The data centers are important parts of universities and other research institutes, as they contain the servers necessary to crunch data-heavy calculations like genomic analysis or protein folding simulations. The rise of cloud computing has seen institutes with appropriately deep pockets move some compute resources off-campus. It’s important to note that the comparison between data center equipment and wet lab equipment isn’t straightforward – freezers must be turned on 24/7, and server usage will likely be shared among many different labs – but gives an idea of the scale of carbon emission that a wet lab-only view ignores. Choosing green software Lannelongue said that being separated from these compute resources doesn’t mean that researchers are powerless to reduce their energy use. An analysis of different computational lab techniques found that similar calculations could incur very different emissions depending on the software used.1 Genome-wide association studies (GWAS) are a powerful genomics tool. But large-scale GWAS are energy intensive. Lannelongue found that analysis using the BOLT-LMM statistical approach incurred a carbon footprint of 17.29 kgCO2e, equivalent to driving a car for 100 km.2 Importantly, the same analysis using an updated version of BOLT-LMM used only 4.7 kgCO2e – a carbon footprint reduction of 73%. Lannelongue pointed out that many researchers might be reluctant to change the version of the software they use, especially if they are tied up in long analysis pipelines, where altering one tool might necessitate updates to other programs. Lannelongue said updating software isn’t the only change that computational researchers can make to improve their sustainability. A 2023 paper he coauthored put together a set of best practice principles called GREENER.3 One of these recommendations was for cultural change in how researchers make use of computing heft. Lannelongue gives the example of researchers working with machine learning models. Many of these systems benefit from hyperparameter tuning, where the model tweaks relevant variables prior to analysis. This tuning process initially produces quick boosts to the model’s accuracy that then become more incremental improvements. A researcher working on a Friday afternoon might be tempted to leave the model running over the weekend to eke out minuscule performance gains. Lannelongue TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 12 explains that universities maintain competitiveness with private research by keeping compute resources at a very low cost. “But because there's no cost, there's no incentive not to waste it,” he adds. The culture change required, he said, will be to recognize that the hyperparameter tuning has an environmental cost, in the same way that a staining assay has a plastic cost in a pile of used pipette tips. A green culture change That change is happening. Slowly. While some universities have no plans to monitor the environmental impact of their computing resources, others are facing up to the problem. Sydney Kuczenski, a green labs outreach and engagement specialist at the University of Virginia, said her institution has been running an in-house certification program for researchers since 2019. This program does allow flexibility for computational labs – allowing them to focus more on appliance power usage rather than cold storage, for example – but until this year, data center power consumption had gone under the radar. That changed when Kuczenski found the Green Algorithms Initiative online. Now, sustainable IT and computing will form part of Virginia’s 2025 Decarbonization Academy, a summer fellowship that introduces students to the changes required for Virginia to meet its aim of being carbon neutral by 2030. Students will learn about how computing processes – from Google searches to data center number crunching – consume carbon. Kuczenski said this may just be the first step. She’s particularly interested in tools like CodeCarbon, which can be implemented into Python codebases to track CO2 emissions produced by computing resources. “I would love to do a campaign of promoting this to researchers to put this tool into their code, and have a year of data tracking of how much carbon emissions our codes on average use,” said Kuczenski. Back in the UK, Lannelongue is championing a new sustainability certification scheme called Green DiSC that is tailored to digital labs. Despite these moves towards greener computing, Lannelongue said he feels “quite negative” about the future of the field. Data centers and computing hardware have been getting more energy efficient for the last two decades, he explained. Despite this, the sector’s energy usage and carbon footprint has been on a relentless upwards trajectory. At the recent Artificial Intelligence Action Summit in Paris, Lannelongue said he talked to leading AI and computing corporations, who insist there is no real problem with AI’s energy use, because efficiency gains will eventually see energy use decline. “It’s completely delusional,” said Lannelongue. “That’s just not going to happen.” As more powerful computers are brought to bear in research, even scientists who don’t make use of AI tools will likely see their computing carbon footprint grow. Some of this change will cut both ways – a lab may use more energy if it adopts an electronic lab notebook system, but will use less paper. But this change shouldn’t be a rush to crunch as much data as possible in the shortest time. Instead, researchers utilizing compute-heavy resources may need to rethink the cost of using these services, “We need to change how we think about computing,” said Lannelongue. “We need to accept that there's an environmental cost that we may want to try to minimize.” ABOUT THE INTERVIEWEES: Dr. Loïc Lannelongue is a bioinformatics researcher at the University of Cambridge. He leads a research group studying the environmental impact of computing. Sydney Kuczenski is a green labs outreach and engagement specialist at the University of Virginia. REFERENCES: 1. Grealey J, Lannelongue L, Saw WY, et al. The carbon footprint of bioinformatics. Mol Biol Evol. 2022;39(3). doi:10.1093/molbev/ msac034 2. Loh PR, Tucker G, Bulik-Sullivan BK, et al. Efficient Bayesian mixed-model analysis increases association power in large cohorts. Nat Genet. 2015;47(3):284-290. doi:10.1038/ng.3190 3. Lannelongue L, Aronson HEG, Bateman A, et al. GREENER principles for environmentally sustainable computational science. Nat Comput Sci. 2023;3(6):514-521. doi:10.1038/s43588-023- 00461-y 13 LAB AUTOMATION & DIGITALIZATION How Will Automation and Digital Technology Shape the Lab of the Future? Blake Forman Advances in automation and digitization are driving the journey to the lab of the future, a vision of a lab where researchers are freed from manual repetitive tasks and efficiency and innovation are at the forefront. Some experts have estimated that 70% of lab workers' time can be wasted on administrative tasks, data analysis and reporting. In the lab of the future, a marriage between artificial intelligence (AI) and robotics will free up time for researchers to focus on more important tasks, such as interpreting data and answering scientific questions. The development of smart labs that utilize the Internet of Things (IoT) to improve data connectivity and autonomous robotic systems that can better navigate a lab environment are just some examples of the technologies restructuring how research is conducted. This article explores these advances further, discussing the roadblocks preventing labs from utilizing these transformative technologies and evaluating how digitization and automation will continue to shape the research landscape. AI-powered robots pave the way for lab automation Laboratory automation is designed to eliminate the most repetitive tasks in a lab, such as liquid handling. As technology has progressed more complex processes Credit: iStock/gorodenkoff TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 14 can now be automated, including entire workflows. Lab automation is usually classified as either partial laboratory automation (the automation of a single step in a lab process) or total laboratory automation (complete automation of a lab process or workflow). Driving lab automation is the progress made in robotics enabling labs to streamline more complex manual processes. Dr. Gabriella Pizzuto is a lecturer in robotics and chemistry automation at the University of Liverpool. Her research at the forefront of artificial intelligence and robotics is focused on improving current robotic systems in labs to make them smarter, more efficient and safer when working in human-centric environments. One of the projects Pizzuto has been involved in is advancing modular, multi-robot integration in laboratories.1 “There are many advantages to laboratory robots, for example, being able to work around the clock, carry out longer experiments and tackle reproducibility challenges,” Pizzuto told Technology Networks. A survey of 1,500 scientists by Nature found that more than 70% of researchers have tried and failed to reproduce another scientist's experiments.2 Many hope that lab robotics will help overcome the reproducibility crisis plaguing scientific research. “With a robot, you have more of a controlled experiment, you remove elements of human and repetitive error. Plus, there is a lot of data that you can record and gather from a robot,” said Pizzuto. Before the potential of laboratory robots can be fully realized, safety and cost concerns must be overcome, Pizzuto explained, “You would need to make sure the robot is as safe as possible. For example, if a robot working 24/7 drops something overnight, the researcher might be exposed to whatever the robot has dropped when they next enter the lab.” “Robots to this day are still quite expensive and so not accessible to everyone,” continued Pizzuto. “Historically, lab-based scientists won’t have experience using robotics so there are also training barriers that need to be overcome before robots can become commonplace in labs.” Traditional laboratory automation focused heavily on rigid machines specific to a single task. The next generation of lab robots should be flexible and capable of carrying out a variety of laboratory tasks. “Researchers are now more interested in intelligent robotics, or autonomous robotic systems. For example, if a robotic system has been developed and deployed for nuclear applications, there could be knowledge transfer to our laboratory environments,” said Pizzuto. Recent advances in robotics have also focused on incorporating AI and machine learning to improve current “robotic scientists”. A “self-driving” laboratory of robotic equipment powered by a simple machine learning model has successfully reengineered enzymes to be more tolerable to heat without human input save for hardware fixes.3 “In our lab, we use machine learning methods, including supervised learning, reinforcement learning and more recently generative AI methods to teach our robots different skills and tasks within laboratory experiments,” said Pizzuto. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 15 “In our lab, we use machine learning methods, including supervised learning, reinforcement learning and more recently generative AI methods to teach our robots different skills and tasks within laboratory experiments,” said Pizzuto. “This is to improve how the robots interact with laboratory environments.” Enter the digital lab As labs continue to incorporate digital hardware, it is more common to see researchers interacting with technology daily. This digital transformation has highlighted the importance of connectivity and security in laboratory practice. “If it's done well, there are a lot of pros both from an academic and an industrial perspective for increasing digitization in the lab,” Dr. Samantha PearmanKanza, senior enterprise fellow at the University of Southampton told Technology Networks. Pearman-Kanza investigates the social aspects of adopting digital technology to understand the barriers that stop labs from going digital. Her research involves applying technologies like electronic lab notebooks (ELNs) and IoT devices to improve the digitization and knowledge management of the scientific record. “Data that is preserved digitally is more secure. People don't put much stock in backing up physical paperwork. By putting something on a computer, you're suddenly affording it more importance,” explained Pearman-Kanza. “Additionally, integrating data on a single platform encourages collaboration within and between labs.” ELNs and laboratory information management systems (LIMS) are vital to ensuring lab data is FAIR (findable, accessible, interoperable and reusable).4 Both of these pieces of software are designed to improve the traceability of lab data and facilitate compliance and data integrity. Cloud-based solutions are available that promise built-in security and regulatory compliance. These third-party solutions reduce the need for traditional IT infrastructure, such as servers and storage devices. Building on data connectivity, smart labs have been created that utilize IoT technologies to improve instrument inter-connectedness. A demo smart lab has been set up at the University of Southampton called Talk2Lab to investigate how IoT devices can be integrated into the lab. One technology they have tested is voice command systems. “Integrating voice allows you to ask questions while you're completing other tasks in the lab. We have tried commands such as: can you show me the dashboard for this, what's the temperature of this,” said PearmanKanza. “In the future, it could be possible to integrate this with ELNs to take notes using your voice alone.” “For me, the grand vision of the smart lab would be to have everything interconnected with all the instruments ‘talking’ to each other,” said Pearman-Kanza. Another key aspect of a smart lab is the integration of smart cameras and sensors. The smart lab setup at the University of Southampton uses cameras that have alleviated the need for an additional staff member to be present during the calibration of certain instruments. The smart lab at the University of Southampton has also incorporated sensors that measure laser power. “Often our laser can short out and this can ruin experiments being set up,” said Pearman-Kanza. “Using sensors, we're looking at how we can take the data on laser power and make predictions on when we think that's going to happen, to save people time with their experiments.” Digital solutions of the future will need to address individual laboratories' pain points and needs. To maximize the benefits of digitization, labs need a clear vision of how to use technology in a way that maximizes their long-term “digital health”. Building trust in artificial intelligence Artificial intelligence is continuing to become more widespread in scientific research. A recent Elsevier survey of corporate R&D professionals found that TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 16 96% of researchers think AI will accelerate knowledge discovery, while 71% say the impact of AI in their area will be transformative or significant. Despite positive sentiments, the report also highlighted researchers' concerns about ethics, transparency and accuracy. There is particular concern over the use of AI in science publishing, with various examples of nondisclosed AI-generated text labels having already appeared in peer-reviewed research from different journals. As developers work through these issues, researchers are finding unique ways to utilize AI to help improve scientific data handling. These applications include automating data collection and curation and optimizing drug discovery and clinical trial design processes.5 The next frontier in bringing AI into labs will likely involve overcoming researchers' fears of the technology. “When there is a new technology that people don't completely understand, there will always be a level of fear,” said Pearman-Kanza. “If we can provide transparency and accountability, this will go a long way to increasing trust in AI technologies.” Overcoming digitization and automaton roadblocks Many labs today don’t lend themselves to the modern requirements of automation and digitization. Assessing whether a lab has the required hardware, space, power supplies and internet connection are just some initial factors to consider when onboarding digital and automation technologies. “You need dedicated lab hardware,” explained PearmanKanza. “One day in the future, there may exist selfcleaning laptops or ways to sterilize all the chemicals off hardware in a way that isn't damaging. But until we get to that, you don’t want to bring contaminated hardware in and out of the lab.” Cost and a lack of specialist knowledge also continue to hamper efforts to onboard these technologies in labs. Pizzuto stated, “Robots are a significant additional cost that labs would have to weigh up.” “We need to promote more teams where engineers, computer scientists and lab-based scientists such as chemists and biologists come together to address the current problems with automation in the lab to make these technologies more accessible,” said Pizzuto. An evolving landscape Digital technology and automation are becoming more prevalent in labs, Pizzuto imagines that within the next few “We need to promote more teams where engineers, computer scientists and labbased scientists such as chemists and biologists come together to address the current problems with automation in the lab to make these technologies more accessible,” said Pizzuto." TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 17 years “Most labs, especially the newer labs, would be using some form of AI within their research and experiment.” The adoption of robotics may be slower than AI due to cost constraints, but Pizzuto envisions the use of robotic arms, dual-arm robotic systems and mobile manipulators increasing to help with labor-intensive workflows. Mobile manipulators that can transport samples between different stations will play an important role in connecting workstations and different labs within the same building. “When we think about using more robotics and AI, data will become crucial,” said Pizzuto. “Having this data stored securely on a cloud-based software and having models on the cloud that can analyze these large volumes of data, will be vital.” Technologies used in other aspects of our day-today lives will continue to find unique uses within the lab. Virtual reality (VR) headsets are now widely used by gamers for a more immersive experience. These smart wearables are also being utilized in research, for example, to help navigate microscopy data.6 Northwestern University researchers have developed VR goggles for lab mice to simulate natural environments to more accurately and precisely study the neural circuitry that underlies behavior.7 “I expect to see the amount of digital hardware in the lab continuing to increase,” said Pearman-Kanza. “Many existing technologies we use in our daily life will find applications in the lab. Hardware you can interact with that also has computing power such as tablets will keep increasing in prevalence.” The future of lab automation and digitization will be a combined human and machine effort requiring new expertise and dedicated data stewards. Pearman-Kanza concluded, “I think it’s important to remember that one of the common goals with technology is to make data more reliable and reproducible. A successful digital future will involve understanding where we need people and where we can best use technology to optimize human efforts.” ABOUT THE INTERVIEWEES: Dr. Gabriella Pizzuto is a lecturer in robotics and chemistry automation and a Royal Academy of Engineering research fellow at the University of Liverpool, and a research area lead in chemical materials design at the Henry Royce Institute. Pizzuto obtained her PhD in computer science from the University of Manchester. Her research focuses on developing new methods for creating the next generation of robotic scientists. Dr. Samantha Pearman-Kanza is a senior enterprise fellow at the University of Southampton. Pearman-Kanza obtained her PhD in web science from the University of Southampton. Her research involves applying computer science techniques to the scientific domain, specifically through the use of semantic web technologies and artificial intelligence. REFERENCES 1. Lunt AM, Fakhruldeen H, Pizzuto G, et al. Modular, multi-robot integration of laboratories: an autonomous workflow for solidstate chemistry. Chem Sci. 2024;15(7):2456-2463. doi: 10.1039/ D3SC06206F 2. Baker M. 1,500 scientists lift the lid on reproducibility. Nature. 2016;533(7604):452-454. doi: 10.1038/533452a 3. Rapp JT, Bremer BJ, Romero PA. Self-driving laboratories to autonomously navigate the protein fitness landscape. Nat Chem Eng. 2024;1(1):97-107. doi: 10.1038/s44286-023-00002-4 4. Wilkinson MD, Dumontier M, Aalbersberg IjJ, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016;3(1):160018. doi: 10.1038/sdata.2016.18 5. Wang H, Fu T, Du Y, et al. Scientific discovery in the age of artificial intelligence. Nature. 2023;620(7972):47-60. doi: 10.1038/s41586- 023-06221-2 6. Spark A, Kitching A, Esteban-Ferrer D, et al. vLUME: 3D virtual reality for single-molecule localization microscopy. Nat Methods. 2020;17(11):1097-1099. doi: 10.1038/s41592-020-0962-1 7. Xia M, Ma J, Wu M, et al. Generation of innervated cochlear organoid recapitulates early development of auditory unit. Stem Cell Rep. 2023;18(1):319-336. doi: 10.1016/j.stemcr.2022.11.024 DATA MANAGEMENT & ANALYSIS AI-DRIVEN TECHNOLOGIES IN THE LAB Lab data output has increased exponentially as experiments have become more complex and instruments more advanced. AI can help researchers analyze and draw conclusions from these large datasets, while machine learning can recognize patterns present in datasets and make predictions based on this. LABORATORY INFORMATION MANAGEMENT SYSTEMS (LIMS) A LIMS enables a lab to track data associated with samples, experiments, lab workflows and instruments. Incorporating AI into these systems allows lab managers to mine this data and make informed decisions based on prescriptive analytics. These insights range from resource management tips to maximize efficiency to providing predictive maintenance of laboratory equipment. REAL-TIME DATA MONITORING AI when used with sensors and LIMS systems can help monitor laboratory environments in real-time, for example, to detect unusual temperature variations within ultralow-temperature freezers. Predictive analytics approaches have found use in technologies such as HPLC, where researchers have used AI models to make predictions on the qualitative properties of an anti-Alzheimer agent, saving time and resources that would usually be spent in method development. OPEN-SOURCE AI SOFTWARE Open-source lab software solutions use source code that is freely available to download, modify and run. Open-source AI software has been developed for several lab applications. For example, a research paper published in Nature Catalysis introduced RENAISSANCE, an AI-based tool combining various types of cellular data to accurately depict metabolic states. The model’s code is publicly available on GitHub, a cloud-based code repository. CLICK HERE TO VIEW THE FULL INFOGRAPHIC 19 LAB AUTOMATION & DIGITALIZATION How Is AI Speeding Pharma’s Journey Toward Precision Medicine? Neil Versel While generative artificial intelligence (AI) burst into the public consciousness a little more than two years ago with the release of ChatGPT, AI and machine learning (ML) have been part of drug research and development for far longer. AI and ML are now being applied to every stage of biopharmaceutical R&D including discovery, compound development, trial design, regulatory submission, supplychain optimization and postmarket surveillance. In the last two years, several published, peer-reviewed articles have referred to precision medicine as “futuristic” – a state that has not yet been fully realized – but drug discovery is rapidly becoming more patientcentered with each new application of AI. Although ChatGPT and other generative AI engines built into smartphones, Internet search engines and social media platforms have captured public attention, generative AI is only one flavor of the technology. The pharma industry has been using predictive modeling for years. Investigators are turning to increasingly advanced algorithms to identify patient subgroups, forecast mechanisms of action, predict differential drug responses including toxicity and even adjust dosing strategies. Credit: iStock/Just_Super TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 20 Popular use cases and therapeutic areas In a 2023 report, the UK-based Wellcome Trust identified five key families of use cases for AI in drug discovery: ∙ Drug target identification and validation ∙ Small-molecule design and optimization by pinpointing “hit-like or lead-like small molecule compounds” ∙ Design and optimization of vaccines, particularly mRNA vaccines ∙ Design and optimization of antibody structures and properties ∙ Evaluation of safety and toxicity of promising compounds According to the report, more than 80% of published articles on AI-enabled drug discovery in the preceding five years were related to understanding disease, target discovery and optimization of small-molecule compounds. Of note, the organization said there has been a dearth of publicly available data on safety and toxicity to train AI models. Interviews with experts in the field included in the report mentioned the challenges of predicting safety and toxicity based on in silico data without sufficient supporting clinical validation. The Wellcome report also cited a 2021 Nature article describing how AlphaFold, an AI/ML algorithm from Google sister company DeepMind, predicted the three-dimensional structure of human proteins with “atomic” accuracy. About 70% of private-sector investments in AI for drug discovery between 2018 and 2022 were in the “commercially tractable” therapeutic areas of oncology, neurology and COVID-19, Wellcome added. As noted in the American Journal of Managed Care (AJMC), numerous drug companies have already incorporated AI into drug R&D processes, particularly to improve target identification, molecule discovery and patient recruitment for trials. “The integration of AI into the drug discovery process offers immense potential for accelerating drug development, reducing costs, and improving patient outcomes,” the October 2024 article concluded. “However, the successful implementation of AI requires addressing knowledge gaps, ensuring data quality, and navigating regulatory challenges.” At least one major drug company has said that it is looking at AI for developing precision therapeutics in oncology, immunology and neuroscience, all popular therapeutic areas for AI application among Big Pharma. However, the Wellcome report cited COVID-19 response as a shining example of the power of this evolving technology. In the early days of the pandemic in March 2020, AI was used to research monoclonal antibodies derived from convalescent plasma of COVID-19 patients. About 2,000 potential candidates were quickly narrowed down to 24, and the most promising compound, bamlanivimab, entered into clinical trials within three months. The US Food and Drug Administration (FDA) granted Emergency Use Authorization to bamlanivimab in November 2020, just eight months after research commenced. Though the FDA revoked the authorization in April 2021 after subsequent SARSCoV-2 variants proved more resistant to the therapy, this episode highlighted just how rapidly drug-makers could develop narrowly targeted treatments. Reality and promise AI analyzes large genomic and multiomic datasets faster than ever before to help target therapies. A study published in Science in September 2023 explained TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 21 how bioinformaticians have been able to build upon each other’s work to extend the capabilities of popular AI algorithms including AlphaFold to improve prediction of pathogenicity of missense variants in the human proteome. A May 2024 review article published in the journal Fundamental Research suggested that AI, in the form of unsupervised machine learning, is a more efficient way of generating patient clusters and phenotype candidates than human training of datasets in the development of gene therapies. AI is also useful for optimizing drug dosing based on individuals’ genetic profiles, the international team of authors wrote. A July 2023 article in Pharmaceutics noted that an AI technique called clustering – which groups similar datapoints to find subgroups of patient data, gene expression profiles, chemical structures and other pertinent information – is useful for identifying drug targets and stratifying patients. Other AI algorithms can sift through complex stores of data in search of anomalies. “AI is being utilized to advance precision medicine approaches. By analyzing patient data, including genomics, proteomics, and clinical records, AI algorithms can identify patient subgroups, predict treatment responses, and assist in personalized treatment decision-making. AI also contributes to the development of biomarkers for disease diagnosis and prognosis,” the authors, academic researchers from India and Northern Ireland, wrote. “AI might revolutionize the pharmaceutical industry in the future to accelerate drug discovery and drug development,” they continued, offering a clear caveat with their word choices. “AI-enabled precise medicine could categorize patients, predict therapy responses, and customize medicines by analyzing genomes, proteomes, and clinical records.” They saw promise in what they called “virtual screening” to accelerate identification of lead compounds by selecting therapeutic candidates with the necessary characteristics from massive chemical databases. “Scientists may create innovative compounds with target-binding characteristics using deep learning and generative models, improving medication effectiveness and lowering adverse effects,” the Pharmaceutics article said. “AI improves clinical trial design, patient selection, and recruitment. AI algorithms will use electronic health records, biomarkers, and genetic profiles to find appropriate patients, lower trial costs, and speed up approval.” But, as the AJMC article stated, AI performance is only as good as the data the algorithm is trained on — the old “garbage in, garbage out” maxim from computer science. One potential tripwire is data bias. Writing in NPJ Digital Medicine, investigators from the Berlin Institute of Health at Charité and Harvard Medical School noted that AI data is often trained on datasets that have underrepresent women, people of color, low-income groups and other historically disadvantaged demographics. “For example, an AI algorithm used for predicting future risk of breast cancer may suffer from a performance gap wherein black patients are more likely to be assigned as ‘low risk’ incorrectly,” they said. Regulation Regulation rarely keeps up with technological advancements, and AI is no different in this regard. While enabling legislation might be behind the times, regulators are at least attempting to stay ahead of the curve by offering guidance to constituents including drug developers. For example, the European Medicines Agency (EMA) recently spent more than a year developing a “reflection paper” offering guidance on using AI/ML in pharmaceutical discovery, development, approval submissions, manufacturing and outcomes assessment. This guidance document spells out how existing European Union and national laws cover AI/ML across pharma lifecycles, including when technologies might be treated as medical devices. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 22 The EMA pointed out the importance of paying attention to data quality and relevance in training algorithms. ML developers should take a “humancentric approach” to design be able to explain their methodologies to allow the agency to review and monitor “black box” models, the paper said. “This requires not only that active measures are taken during data collection and modelling … but also that both user and patient reported outcome and experience measures are included in the evaluation of AI/ML tools when they interface with an individual user or patient,” the EMA explained. 23 LAB AUTOMATION & DIGITALIZATION Cutting-Edge Advances Driving Developments in Robotics Kate Robinson Once limited to repetitive industrial tasks, robots can now learn to cooperate like human teammates, walk without electronics and even jump without legs. Advances in fabrication techniques and autonomous probes are accelerating progress across disciplines, from soft robotics to semiconductor discovery. Here, we highlight five recent developments in robotics that show where the field is heading. Robot Teams Learn Like Humans in 40 Minutes A team from Duke University and Columbia University has developed a novel framework that imbues robots with a primitive “Theory of Mind,” a human trait that allows people to empathize with one another and predict each other’s actions. Unlike traditional hive-mind behavior, the framework (called HUMAC) allows a single human coach to guide robot teams strategically. Through brief, high-impact interventions, robots learn to anticipate peers’ and opponents’ actions. The robots were tested in a game of hide and seek, where three seeker robots tried to catch three faster-moving hider robots. Non-collaborative seekers continued to chase the closest hiders and achieved a 36% success rate. After 40 minutes of guidance, the success rate of the seekers rose to 84% in simulations and 80% in physical ground vehicle tests. Credit: iStock/sankai TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 24 Rapid, efficient robot teaming could have applications in disaster response, where robots need to cooperate and collaborate under constraints, with hierarchical team structures, uncertainty of the environment and communication bandwidth limits. “AI is not just a tool for humans, it’s a teammate. The final form of super-intelligence will not be AI alone nor humans alone, it’s the collective intelligence from both humans and AI,” said Boyuan Chen, the Dickinson Family Assistant Professor of Mechanical Engineering and Materials Science, Electrical and Computer Engineering and Computer Science at Duke University. “Just as humans evolved to collaborate, AI will become more adaptive to work alongside with each other and with us. HUMAC is a step toward that future.” Artificial Muscle Brings Multidirectional Motion to Soft Robots Engineers from the Massachusetts Institute of Technology (MIT) have developed a method to produce artificial muscle tissue capable of twitching and flexing in multiple directions. In recent years, scientists have looked to muscles as potential actuators for “biohybrid” robots. Such biobots could reach spaces where traditional machines cannot, but researchers have only been able to fabricate artificial muscle that pulls in one direction, limiting any robot’s range of motion. To demonstrate the new process, the researchers created a structure similar to the human iris using hydrogel scaffolds stamped with microscopic grooves. The grooves were seeded with muscle cells, which grew along the grooves, forming fibers that contracted in multiple directions when stimulated with light. “Instead of using rigid actuators that are typical in underwater robots, if we can use soft biological robots, we can navigate and be much more energy-efficient, while also being completely biodegradable and sustainable,” said Ritu Raman, the Eugene Bell Career Development Professor of Tissue Engineering in MIT’s Department of Mechanical Engineering. “That’s what we hope to build toward.” Legless Soft Robot Jumps 10 Feet by Emulating Nematode Motion Georgia Tech engineers have created a legless soft robot that can jump up to 10 feet. The robot, a silicone rod with a carbon-fiber spine, was designed to mimic nematodes, which store energy by bending their bodies into tight kinks and then release it in a sudden leap or backflip. The researchers built soft robots to replicate the leaping worms’ behavior, later reinforcing them with carbon fiber to accelerate the jumps. This energy-efficient, biomimetic locomotion could inspire agile robots capable of terrain traversal where traditional wheels or legs fail such as rubble, sand or uneven surfaces. “A jumping robot was recently launched to the moon, and other leaping robots are being created to help with search and rescue missions, where they have to traverse unpredictable terrain and obstacles,” said Sunny Kumar, lead coauthor of the paper and a postdoctoral researcher in the School of Chemical and Biomolecular Engineering (ChBE). “Our lab continues to find interesting ways that creatures use their unique bodies to do interesting things, then build robots to mimic them.” Electronics-Free, 3D-Printed Robots Walk Immediately After Printing A group at UC San Diego’s Bioinspired Robotics Lab has crafted a six-legged, electronics-free robot that can walk right off a 3D printer – powered solely by compressed gas. These robots are made from simple 3D-printing filament and cost approximately $20 each to manufacture. When TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 25 tested, the researchers found that while connected to a source of gas under constant pressure, the robots could keep functioning non-stop for three days and were able to move across surfaces like sand, turf and even underwater. A pneumatic oscillating circuit delivers air pressure at the right time, alternating between two sets of three legs to allow the robot to walk in a straight line. Thanks to the lack of electronics, these robots could be utilized in high-radiation zones. “This is a completely different way of looking at building machines,” said Michael Tolley, a professor in the UC San Diego Department of Mechanical and Aerospace Engineering and the paper’s senior author. Autonomous Robotic Probe Accelerates Semiconductor Discovery MIT researchers have deployed a self-supervised robotic probe to autonomously measure photoconductance of semiconductor materials at unprecedented speeds. The pace of new semiconductor material discovery is bottlenecked by the speed at which researchers can manually measure important material properties. The newly developed system was able to perform over 125 unique photoconductance measurements per hour during a 24 hour test, totaling more than 3,000 measurements. In order to measure materials, the robotic system takes an image of a material and then cuts the image into segments. The images are fed into a neural network model, which uses domain-informed machine learning to determine the optimal points for the probe to contact. These contact points are fed into a path planner that finds the most efficient way for the probe to reach all points. The robot’s motors then manipulate the probe and take measurements at each contact point in rapid succession. “Being able to gather such rich data that can be captured at such fast rates, without the need for human guidance, starts to open up doors to be able to discover and develop new high-performance semiconductors, especially for sustainability applications like solar panels,” said Alexander Siemenn, postdoctoral associate at MIT and lead author of the paper. The future of robotics Whether through human-like teamwork, biofabricated muscles, legless leaps, electronics-free locomotion or autonomous discovery, each development shows how engineers are moving beyond conventional designs to create machines that are more adaptable, resilient and capable. As robots become more flexible and autonomous, their potential applications will extend well beyond the factory floor, supporting innovations in healthcare, sustainable energy, disaster response and exploration. 26 LAB AUTOMATION & DIGITALIZATION Innovations in LIMS for Enhanced Laboratory Efficiency Bob McDowall, PhD The term LIMS typically refers to a software application. However, a better term is LIMS environment. This broader concept not only encompasses the traditional LIMS functions such as sample management, data management, results calculation and reporting, but also integration of analytical instruments and other laboratory software systems, including instrument data systems, ELN (electronic lab notebooks) and LES (lab execution systems). The following listicle explores how innovative technology can increase process efficiency and support a wider laboratory digitalization strategy. The key to utilizing the technologies discussed in this listicle is to first understand and redesign lab processes to work effectively electronically prior to implementation. In addition, equally important is having a resilient, robust and fault-tolerant IT infrastructure. This includes cybersecurity, adequate data storage, backup and recovery and responsive help desk support. Whether hosted on-premises, in the cloud or a hybrid of both, this infrastructure is critical. This is essential because as laboratories become more digitalized, any IT issue can quickly escalate to major disruptive incidents. Credit: iStock/andresr TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 27 Caveat emptor! Or in plain English: buyer beware. When it comes to adopting new technology, the laboratory and organization are responsible for ensuring that the technology being purchased and implemented is truly fit for its intended use. It is easy to get excited about shiny new tools, especially with tech-savvy people in the lab or IT seeing something cool and thinking: how can we use this? But that is often a case of technology in search of a problem to solve. Take the alternative view. What is your business problem? How can it be solved efficiently? What are the options for implementation, at what cost and over what timeframe? How mature is your organization when it comes to implementing and delivering IT projects? Are they typically on time and on budget? Before diving into any of the innovations discussed here, you need a solid business case. Some solutions are relatively simple; others are more complex. Either way, you must ensure that the business case is driving the adoption of innovative technologies and not vice versa. Bear this in mind as we discuss the various innovations individually and how they can work to support your digital transformation. QR-coded volumetric glassware This is not the white heat of technology, but it is a simple and effective way to save time and reduce errors in manual sample preparation. While laboratory automation often focuses on informatics, wet chemistry and sample preparation remain a major element of analytical workflows. With smaller sample volumes, it can be difficult to find cost-effective automation or robotics to implement. The problem When preparing a sample for instrumental analysis, many analytical procedures involve dissolving and diluting a sample using volumetric glassware such as flasks and pipettes. Each item of grade A volumetric glassware has a unique serial number etched on it. Analysts are expected to manually record these numbers in the analytical batch record, which is a slow, tedious and error-prone process. And how do you know the right glassware is being used for the right analysis? The solution To automate and streamline the process, volumetric glassware is now available with individual QR (quick response) codes etched directly on each item. There are some options to affix QR-coded labels; however, these must be durable and robust enough to resist laboratory use and washing/drying cycles for the life of the flask. Etching, on the other hand, is permanent and wear resistant. Here's how it works: ∙ Each item of glassware is registered in the LIMS application with key attributes such as: ₀ Type (e.g., pipette, flask) ₀ Size ₀ Grade (A or B) ₀ QR code identifier ∙ Sample preparation workflows can be set up and validated in the LIMS, ELN or LES, specifying the required glassware. ∙ When a specific workflow is executed, each item of glassware used is scanned. The system checks the item’s identity against the LIMS database and if incorrect, the analyst is prompted to select a correct item. As only grade A glassware is permitted for pharmaceutical analysis, any use of grade B glassware can be easily prevented. This approach speeds up data entry, reduces errors and helps eliminate one potential source of outof-specification results. It also enables tracking of TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 28 individual items of glassware across analyses, providing a clear usage history. The device used to scan the volumetric glassware can also be used to input other information about the sample preparation, such as sample identities and contemporaneous documentation of any issues. We’ll return to this topic later in this listicle to see how it fits into a fully integrated LIMS environment. Voice input (Alexa for the lab) We are all aware of voice commands on smart speakers and mobile phones to handle some everyday tasks. But what can be achieved in a laboratory? The problem Imagine you’re working in a fume cupboard and need to document what you’re doing. You’d have to stop, remove your gloves, write down your observation, re-glove and carry on working? What a drag! Trying to remember everything until you’ve finished or writing them on your lab coat sleeve or disposable glove are not reliable or acceptable record-keeping options. The solution What you need is voice input, a way to capture observations in real time, hands-free. There are now apps that support this and can be integrated with your LIMS. Here’s how it works: ∙ An electronic workflow is established in the LIMS (or other informatics systems) to enforce the analytical procedure. ∙ Voice and other input methods applicable for the procedure are linked to the workflow. ∙ The mobile devices for capturing inputs can be phones or tablets, but they should be company property, not personal ones. Voice input is not just used when encumbered with protective equipment but can also be used to record manual tasks requiring observations that are currently documented on paper. And mobile devices can do more than just capture voice, they can also take photos, scan barcodes or QR codes for identification of samples, reference standards, instrument identities, QR-coded volumetric glassware, etc. For example, tests based on observation or appearance could be automated by voice input along with a picture of the sample. This not only speeds up initial testing but also makes second-person review faster and more objective, thanks to the recorded evidence to confirm the observation. However, voice input is not as easy as using a smart speaker at home, where any instruction from anyone will be carried out. The system needs a vocabulary of laboratory terms so that it can understand the context and meaning of the input. It also needs to be trained to recognize each user so that entries can be properly attributed. This is where artificial intelligence (AI) comes in to help; it is used to analyze and train the system and generate a voice profile for each user. Internet of Things (IoT) IoT is defined by the Food and Drug Administration (FDA) as: A type of cyber-physical system comprising interconnected computing devices, sensors, instruments, and equipment integrated online into a cohesive network of devices that contain the hardware, software, firmware, and actuators which allow the devices to connect, interact, and exchange data and information. IoT devices include sensors, controllers, and mechanical equipment.1 Given the vagueness of the definition above, IoT can mean almost anything to anybody, depending on the context. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 29 The problem Analytical processes are often complex, involving many subtasks and data points. Take a sample, for example. You need to know: ∙ Storage conditions of each storage location (e.g., ambient, refrigerated, frozen or deep frozen) ∙ Whether those conditions are maintained correctly and are within limits ∙ How often a sample is taken out of storage and replaced (especially important for thermolabile samples) ∙ Knowing the position of the sample in its storage location, essential to aid quick retrieval and replacement However, this is only one part of the analytical process; all tasks must be linked together for each analytical procedure as a key step in the digitalization journey. That is where IoT helps. The solution Translating IoT for a laboratory environment, this involves a miscellany of analytical instruments, instrument trays, displacement pipettes, volumetric glassware, samples, reference standards, buffers, standard solutions, microtiter plates, vials for instrumental analysis and storage locations that are all uniquely identified, usually by QR codes or barcodes. Even voice input can be included as part of the LIMS environment. The goal of IoT is to connect tasks across the analytical workflow to enable traceability, and it can be used for: 1. Process verification, ensuring each step has been executed correctly and consistently 2. Faster analysis and second-person review, reducing delays and manual checks 3. Improved quality oversight with verified or validated workflow and traceable electronic records prior to release A few words of caution: 1. Cybersecurity can be a critical issue with some IoT devices. Before purchasing any device, assess how security is set up, managed and maintained. Can default administrator credentials be changed to protect the device and the laboratory data? Are communications protocols secure? Some devices may be hardcoded with this information and cannot be changed – this is a red flag. 2. As mentioned earlier, is IoT a technology in search of a problem to solve? Or is there a business need to meet? Ensure you have a strong business case. If devices and instruments are used to acquire and transfer data, they need to be read and understandable. One way to do this is analytical information markup language, a hypertext method of storing analytical data using XML adapted for laboratories that includes compliance features. This is also important when we consider the use of AI. Artificial intelligence and machine learning (AI / ML) If not handled carefully, AI can become another innovation driven by hype and a search for a problem to solve. AI has real potential in a LIMS environment. It’s important to distinguish between the two key types of AI: ∙ Generative AI: This type of AI creates content based on training data. You must train the tool using YOUR data rather than trawling the internet. This ensures you can set the boundaries of learning and use. You will need two datasets: one for learning and another for independent testing. The learning is only as good as the dataset used and directly affects the AI's performance. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 30 ∙ Adaptive AI: Once taught, this system continues to learn after development. While powerful, it presents a challenge in regulated environments. How do you ensure it remains under control and does not generate unreliable outputs or hallucinations? Our recommended approach is to maintain two versions of the released and tested system: ₀ A live version that operates without selflearning so that it remains under control. ₀ A shadow version that is not operational but can self-learn from the additional data inputs to the first version. The shadow version can be tested using the updated test dataset and released for operational use if it performs well. The first operative version would be retired, and the cycle repeated. To manage this responsibility, organizations need AI governance to oversee how generative AI is used.2 The problem: Where can AI deliver business value? One major area is data collation and trending. The current focus is on testing versus specification release. At the end of a year, an annual product review/product quality review must be conducted in pharma.3, 4 Typically, these involve the collation of test results from all batches, usually by generating spreadsheets and manually performing trend evaluation. This is a slow, tedious and time-consuming task. The solution: AI can help laboratories change from a reactive to a proactive approach. Instead of waiting until year-end to review trends and respond to issues, AI can be trained to continuously monitor batch data as it’s released. Indeed, why limit the review to just a year? An ongoing review could be performed over a longer period if required. The benefits would include: ∙ Predictive analytics that spot trends early ∙ Faster, more consistent detection of potential issues ∙ Quicker identification of problems before they escalate The FDA has recently published a draft guidance on using AI for regulatory decision-making. It emphasizes the credibility of the AI, not just validation and introduces the concept of context of use to define how AI should be applied.5 To make AI work effectively, you will need: ∙ High-quality, well-structured data for training and testing ∙ Clearly defined use cases ∙ Sufficient computing power for the AI models to work effectively Bringing it all together In this listicle, we have explored four innovations that can enhance a LIMS environment. Each one offers a targeted solution to a specific business challenge. Think of them as pieces of a laboratory jigsaw puzzle, each of which can contribute to the broader goal of automation and digitalization of a laboratory. Figure 1 illustrates how they could all be integrated to leverage major process improvements in a laboratory. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 31 FIGURE 1. INTEGRATION OF INNOVATIONS IN A LIMS ENVIRONMENT TO ENHANCE LABORATORY EFFICIENCY. CREDIT: BOB MCDOWALL. • Samples coded • Sample storage cabinets coded • Individual locations in cabinet coded • Check in of samples • Check out of samples • Monitoring of storage conditions (temperature & RH as needed) • Return of samples for check in - storage location updated Sample Receipt and Storage Locations Sample Preparation Instrumental Analysis Interpretation / Calculation of Results Collation and Reporting Results • Identify instrument to use and check status • Is it the right instrument? • Unique user identity to set up instrument • Point of use check / System Suitability Test • Transfer vials to autosampler • Bottom of vial uniquely numbered and linked to injection sequence • Each vial scanned when injected and linked to sample identity: no vial mix ups possible • Results from all tests collated for the sample • Check for Out of Specification results • Resolve any issues or deviations • Issue electronic CoA • Colour and observation tests captured electronically (results sent to LIMS from here) • Standards, solutions, reagents and buffers coded • Solutions prepared fresh coded • QR coded glassware • Balances and pH meters on line to LIMS or Instrument data system • Instruments coded • Analytical data captured electronically or by voice input • Extracts are transferred to QR coded vials • Chain of custody for preparation work • SST samples interpreted and checked versus acceptance criteria • Standards and QCs interpreted and checked • Samples interpreted and checked • Results of aliquots calculated • Calculate reportable result • Transfer of the results to LIMS for collation Artifical Intelligence / Machine Learning: Trending Data Laboratory Information Management System On-going Product/ Annual Quality Reviews Electronic Certificate of Analysis (CoA) Sample Registration TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 32 Organizational maturity and staff training We can discuss many kinds of innovations to boost laboratory productivity, but how ready is your organization to implement them? An organization’s technological maturity plays a crucial role, and it should consider the following: ∙ Are automation projects delivered on time and within budget, or are they hopelessly late and become a money pit? ∙ How are staff involved in these projects? Are they engaged and motivated, with a sense of ownership? ∙ Do analysts have the skills, incentive and experience to implement and use new technologies effectively? ∙ How well are analysts trained to use the new ways of working? There needs to be a clear pathway from the current to the new ways of working. Roles will evolve as these systems do not come with a simple “ON” button; they still require staff to set up, operate and manage them. You might have cutting-edge innovative technology, but you also need the trained and motivated analytical staff to use it to your advantage and unlock its full potential. ACKNOWLEDGEMENTS I thank Gemma Harben and Joost Van Kempen for their help in reviewing and preparing this listicle. REFERENCES 1. Food and Drug Administration. Artificial Intelligence in Drug Manufacturing. Silver Spring, MD: Food and Drug Administration; 2023. 2. Mintanciyan A, Budihandojo R, English J, Lopez O, Matos JE, McDowall R, Artificial Intelligence Governance in GXP Environments. Pharm Eng. 2024;44(4):62-67. 3. Food and Drug Administration. 21 CFR 211 Current Good Manufacturing Practice for Finished Pharmaceutical Products. Silver Spring, MD: Food and Drug Administration; 2008. 4. European Commission. EudraLex - Volume 4 Good Manufacturing Practice (GMP) Guidelines, Chapter 1 Pharmaceutical Quality System. Brussels: European Commission; 2013. 5. Food and Drug Administration. FDA Draft Guidance for Industry Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products. Silver Spring, MD: Food and Drug Administration; 2025. 33 LAB AUTOMATION & DIGITALIZATION Revolutionizing Laboratory Information Management Systems With IoT and Smart Technology Phil Williams, PhD Laboratories face growing demands for efficiency, accuracy and compliance in managing complex data and workflows. To meet these challenges, many are adopting Internet of Things (IoT)-integrated laboratory information management systems (LIMS). This combination enables real-time data exchange, task automation and enhanced process visibility, supporting precise sample tracking and regulatory compliance. This article explores the benefits, key applications, challenges and future directions of IoT-enabled LIMS. Enabling a fully connected lab By incorporating an IoT-enabled LIMS, laboratories can foster an interconnected environment where instruments and analytics tools communicate more efficiently. This enhances operational efficiency, streamlines workflows, reduces costs and ensures precise data collection, even in high-throughput settings. One example of the benefits of such interconnectedness is in smarter sample management. Sample management is critical, especially in labs that handle thousands of samples daily. Credit: iStock/poba TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 34 With integrated IoT sensors able to monitor vital conditions such as temperature and humidity, IoTenabled LIMS helps labs maintain sample integrity. These sensors provide continuous feedback and trigger alerts for quick responses to storage changes, preventing degradation. This proactive approach preserves sample quality, enhances traceability, helps ensure compliance with relevant regulations and reduces the risk of costly sample losses.1 Automated documentation and predictive maintenance transform laboratories Professor Karnik Tarverdi, Director of Extrusion Technology at the Wolfson Centre for Materials Processing, Brunel University of London, states, “Compliance is vital for clinical and pharmaceutical labs, which require accountability and traceability for regulations such as ISO 9001 and 17025. Relying on manual documentation risks compliance issues developing through the introduction of errors and inefficiencies. IoTenabled LIMS with cloud integration avoids these risks, ensuring accuracy, efficiency and compliance.” IoT sensors, for example, are crucial for continuously monitoring environmental parameters such as humidity and temperature in stability chambers. They may send out real-time notifications in the event of abnormalities. Automated documentation is distinguished by this degree of accuracy and dependability, which guarantees full traceability and helps labs meet compliance requirements. Cloud integration is also beneficial because it improves productivity by offering safe data storage, quick reporting and simple access to past documents. Labs can retrieve years of data in minutes, significantly reducing the burden of preparation during audits. As Tarverdi notes, “Automated documentation systems are essential for maintaining compliance, boosting productivity, and staying competitive in the ever-evolving laboratory environment as regulatory requirements continue to shift”. The integration of IoT systems in the lab can also enable predictive maintenance by monitoring equipment health to flag failures before they happen. Sensors track performance metrics, identifying early signs of wear and allowing proactive maintenance to be scheduled. This reduces downtime, extends equipment lifespan and ensures efficiency. For instance, IoT-enabled LIMS can alert technicians when a centrifuge requires servicing based on its past usage, minimizing costs and avoiding disruptions to critical workflows. As Tarverdi notes, “Automated documentation systems are essential for maintaining compliance, boosting productivity, and staying competitive in the everevolving laboratory environment as regulatory requirements continue to shift”. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 35 AI and edge computing: Transforming the data landscape As IoT becomes integral to lab operations, advanced technologies such as artificial intelligence (AI) and edge computing can offer additional layers of capability. AI algorithms can assist with the processing of the extensive datasets generated by IoT sensors, revealing insights that drive data-driven decisions and operational improvements. For instance, predictive modelling powered by AI can help labs anticipate trends, optimize workflows and enhance resource utilization by uncovering patterns that might not be immediately obvious. Edge computing is a networking philosophy that moves the processing of data closer to where it is generated, thereby reducing latency and also bandwidth use. In labs, edge computing enables faster data analysis, which is critical for high-stakes research, such as personalized medicine or diagnostics experiments. By minimizing dependence on centralized cloud systems, edge computing enhances data privacy - a valuable feature in clinical environments where data security is crucial. Integrating robotics for automated sample preparation Integrating robotics with IoT-enabled LIMS has revolutionized sample preparation, a traditionally labour-intensive task. By connecting LIMS with robotic systems, labs can automate processes such as sample aliquoting and labelling and sorting, reducing human error and enhancing reproducibility. In a high-throughput environment, IoT-enabled robots efficiently handle routine sample management tasks, allowing lab personnel to concentrate on more complex analytical work. Additionally, robotic automation ensures consistent sample handling, which reduces the variability often introduced by manual processing. This level of precision not only increases throughput but also aligns with the rigorous standards required in regulated industries such as pharmaceuticals and biotechnology. Data analytics, security and realtime decision-making IoT-enabled LIMS transforms labs by enabling real-time data collection and analytics for immediate, data-driven decisions. IoT devices monitor lab parameters such as environmental conditions and equipment performance, allowing managers to analyse trends, optimize workflows and respond proactively. For example, IoT systems can alert staff if temperature or humidity measurements exceed safe ranges, protecting samples and experiments while ensuring reliable results. Data security and privacy are critical for IoT-enabled LIMS. Labs must implement encryption and multifactor authentication and comply with regulations like GDPR and HIPAA.2 Cloud-based LIMS enhances security with audit trails, backups and restricted access, ensuring sensitive data is protected. IoT devices also monitor for breaches, maintaining high standards of data integrity and privacy. Towards a global network of labs The future of IoT in laboratories promises greater connectivity and efficiency. With 5G technology, IoT devices will enable faster, more reliable communication for real-time applications, including remote equipment monitoring and virtual labs.3 AI advancements will enhance IoT-enabled LIMS with predictive analytics and automated decision-making for research and diagnostics. IoT shows potential in multi-omics research and personalized medicine by facilitating data collection and analysis. Labs can collaborate with healthcare providers using IoT-enabled diagnostic tools like smart blood analysers, networked imaging devices, and wearable health monitors to transmit patient data in real-time. This connectivity allows labs to receive immediate updates on patient health metrics, such as glucose levels, heart rate or oxygen saturation, directly from wearable devices. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 36 “Subsequent benefits of this technology include earlier detection of health problems, which would decrease the potential resource burden on healthcare,” said Sheri Scott, a senior lecturer and biomedical scientist at Nottingham Trent University. “Faster intervention would decrease the likelihood of hospital admissions or trigger earlier interventions, thus reducing overall costs of treatment and decreasing the overall carbon costs of healthcare,” Scott continued. “A further benefit could see a reduction in waiting times for other patients and result in potential workload decreases for healthcare professionals.” IoT can integrate with lab equipment such as automated handlers and sequencing machines to centralize data seamlessly. This hub analyses trends and shares insights with healthcare providers securely. For example, a detected change in biomarker levels can trigger instant alerts, enabling faster healthcare decisions and personalized treatments. “Although IoT-enabled LIMS promises many benefits - including faster diagnosis and quicker subsequent interventions - professionals’ exhibit concerns over the standardization, costs, security and potential biases resulting from the training data used for the tech,” Scott cautioned. “These are valid concerns and they would need to be addressed before wide-spread adoption can occur.” Despite these concerns, with healthcare seeking new environmentally sustainable ways of working, IoT is a clear potential tool to support healthcare’s sustainability agenda. Challenges and the future outlook for IoT-LIMS Integration Despite its benefits, IoT-enabled LIMS presents challenges. Labs must be aware of potential issues with network compatibility, ensuring interoperability and integration with existing systems. Security is a key concern, requiring robust data protection and continuous monitoring. Additionally, evolving regulations necessitate systems that can adapt to new compliance standards. High upfront costs may also deter smaller labs, making cost-benefit analysis and scalable solutions critical for IoT adoption. The integration of IoT with LIMS has the potential to completely revolutionize laboratory management, delivering connectivity, automation and data-driven insights beyond what has been seen before. It can enhance workflows, from sample tracking to predictive maintenance and real-time analytics, creating more efficient and reliable labs. As IoT evolves with AI, 5G and edge computing, labs will gain even greater precision and agility. For cutting-edge research, IoT-enabled LIMS is essential to streamlining processes, boosting capabilities and accelerating discovery, with future innovations promising to redefine lab operations. ABOUT THE INTERVIEWEES: Professor Karnik Tarverdi is the Director of Extrusion Technology at the Wolfson Centre for Materials Processing, Brunel University London. He has many patents and has published over 80 papers. Sheri Scott is a senior lecturer and biomedical scientist at Nottingham Trent University and a Fellow of the Institute of Biomedical Science. She has over 21 years of clinical biochemistry experience in NHS laboratories. REFERENCES: 1. Srivastava AK, Das P. Chapter: Smart laboratories and IoT transformation. In: Biotech and IoT. Apress; 2024:37-73. doi: 10.1007/979-8-8688-0527-1_3 2. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) [2016] OJ L 119, 04.05.2016; cor. OJ L 127, 23.5.2018. http://data.europa.eu/eli/ reg/2016/679/2016-05-04. Accessed February 2025. 3. Ahmed SF, Alam MdSB, Afrin S, et al. Toward a secure 5G-enabled Internet of Things: A survey on requirements, privacy, security, challenges, and opportunities. IEEE Access. 2024;12:13125-13145. doi: 10.1109/ACCESS.2024.3352508 37 LAB AUTOMATION & DIGITALIZATION Advancing Mass Spectrometry Data Analysis Through Artificial Intelligence and Machine Learning Isabel Ely, PhD Since mass spectrographs and spectrometers were introduced in the early 1900s,1 mass spectrometry (MS) has undergone tremendous technological improvements. Once a methodology primarily used by chemists, MS is now an incredibly versatile analytical technique with several applications in research including structural biology, clinical diagnostics, environmental analysis, forensics, food and beverage analysis, omics and beyond. MS produces a vast quantity of data that needs to be analyzed. Managing, processing and interpreting these large data outputs is computationally intensive and often prone to errors, particularly when manual or semi-automated processes are used. Consequently, artificial intelligence (AI) and machine learning (ML) have become immensely popular for processing MSgenerated data and statistical analysis as they can be applied to various biological disciplines,2 limit errors and enhance data analysis. This article delves into what MS data analysis entails and its associated challenges, how AI/ML can aid analyses and exciting potential future developments in the field, with specific applications to proteomics and metabolomics research. Credit: iStock/NicoElNino TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 38 What does MS data analysis entail? “Data analysis in proteomics and metabolomics is a complex, multi-step process that begins with the collection of biological samples and culminates in the extraction of meaningful biological insights,” Dr. Wout Bittremieux, assistant professor in the Adrem Data Laboratory at the University of Antwerp, said. After the laborious sample preparation of extracting proteins, peptides or metabolites of interest,3 they are ionized and introduced into a mass spectrometer where they are detected based on their mass-to-charge (m/z) ratio, producing a mass spectrum. The coupling of MS with other analytical tools, such as gas chromatography and liquid chromatography, allows for the further separation and identification of such analytes. “One of the key challenges in MS is the accurate annotation of MS spectra to their corresponding molecules,” Bittremieux said. “In proteomics, the dominant method for this task is sequence database searching. This relies on comparing experimental to theoretical spectra simulated from peptides assumed to be present. However, these theoretical spectra are often oversimplified and do not capture detailed fragment ion intensity information, which can lead to significant ambiguities and false identifications.” Once data are quantified, either relatively or absolutely, statistical analyses can take place to facilitate biological interpretation. “To contextualize the results, pathway analysis tools can be used to map the identified proteins or metabolites onto known biological pathways to help in understanding the functional implications of the changes observed in the data. Alternatively, biomarker candidates can be identified based on their ability to distinguish between different biological conditions or groups,” Bittremieux explained. Applying AI/ML to MS data analysis Although some scientists still have concerns about large-scale AI implementation, AI and ML have become indispensable tools for MS data analysis; aiding clinical decisions, guiding metabolic engineering and stimulating fundamental biological discoveries. Applying AI/ML in MS research attempts to minimize errors associated with data analysis –including high noise levels, batch effects during measurements and missing values4 – enhance usability and maximize data outputs.5 Further, training ML models on large datasets of empirical MS spectra allows the generation of highly accurate predicted spectra that closely match the experimental data.6 This overcomes the limitations of traditional sequence database searching, which relies on crude, theoretical spectra. Developments in AI/ML have led to more accurate, efficient and comprehensive interpretations of biological data, including de novo peptide sequencing.7 “De novo peptide sequencing, which involves determining the peptide sequence directly from tandem (MS/MS) spectra without relying on a reference database, is a challenging problem. ML approaches are starting to impact this area significantly by learning patterns from known spectra and using them to predict peptide sequences from unknown spectra, making it feasible to analyze complex proteomes without relying solely on existing protein databases,” Bittremieux said. Another area AI/ML has been applied to in MS data analysis is repository-scale data analysis.8 Public data repositories have continued to expand, now containing millions to billions of MS spectra. Despite existing data providing ample opportunity to extract new biological insights, the sheer volume of data presents significant challenges in terms of data processing and analysis. “We have developed AI algorithms capable of performing large-scale analyses across these repositories, identifying patterns across experiments and detecting novel peptides and proteins that were TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 39 previously missed. This has led to discoveries that would have been impossible with manual or traditional computational methods.” Recent developments in AI/ML Although advancements in AI have been fruitful, applying these technological developments to MS data is challenging due to its unique nature, making a direct translation of AI advancements to MS data non-trivial. “One of the most significant recent advancements in AI relevant to MS data analysis is the development of more sophisticated deep learning models capable of handling high-dimensional data and extracting intricate patterns,” said Bittremieux. “For example, transformer neural networks, which were originally developed for natural language processing, are now effectively used to ‘translate’ between sequences of peaks in tandem MS spectra to sequences of amino acids during de novo peptide sequencing. These models can learn from vast amounts of empirical MS data, identifying subtle features that traditional methods might overlook.” “Despite such advancements, the successful application of AI to MS data still requires deep expertise in both AI and MS. This multidisciplinary skill set remains relatively rare, which has slowed the broader adoption of AI in the field. However, as more researchers receive training in both areas and as AI tools become more accessible, we are beginning to see a new generation of scientists capable of bridging this gap.” Looking towards the future Although significant advancements in AI and ML have aided the continual development of MS data analysis, there is still room for improvement. “One of the key areas where I believe future developments should be focused is on the generation and curation of high-quality, large-scale datasets. While advancements in AI model architectures have been impressive, these models are only as good as the data they are trained on,” discussed Bittremieux. Greater availability of diverse MS data sets would ultimately enable the development of AI tools suitable for use across multiple experimental conditions in differing biological topics.9 “These datasets should include comprehensive annotations, such as accurate peptide and metabolite identifications, quantification data and metadata related to sample preparation and instrument settings. This diversity will enable AI models to learn more generalizable patterns, improving their performance across different applications.” “One of the most significant recent advancements in AI relevant to MS data analysis is the development of more sophisticated deep learning models capable of handling high-dimensional data and extracting intricate patterns,” said Bittremieux. TECHNOLOGYNETWORKS.COM LAB AUTOMATION & DIGITALIZATION 40 Researchers are sometimes at fault for testing their models on cherry-picked datasets. This contributes to a lack of standardization evaluations assessing the performance of different models. Bittremieux detailed that “the development of benchmarking suites would allow for a fair comparison of different algorithms, fostering transparency and driving genuine progress in the field.” “As AI tools become more accessible and interpretable, we will likely see a surge in innovative applications, from personalized medicine to environmental monitoring.” REFERENCES 1. Wilkinson DJ. Historical and contemporary stable isotope tracer approaches to studying mammalian protein metabolism. Mass Spectrom. Rev. 2018;37(1):57-80. doi:10.1002/mas.21507 2. Neagu AN, Jayathirtha M, Baxter E, Donnelly M, Petre BA, Darie CC. Applications of tandem mass spectrometry (MS/MS) in protein analysis for biomedical research. Molecules. 2022;27(8):2411. doi:10.3390/molecules27082411 3. Luque-Garcia JL, Neubert TA. Sample preparation for serum/plasma profiling and biomarker identification by mass spectrometry. J. Chromatogr. A. 2007;1153(1):259-276. doi:10.1016/j.chroma.2006.11.054 4. Liebal UW, Phan ANT, Sudhakar M, Raman K, Blank LM. Machine learning applications for mass spectrometry-based metabolomics. Metabolites. 2020;10(6):243. doi:10.3390/ metabo10060243 5. Beck AG, Muhoberac M, Randolph CE, et al. Recent developments in machine learning for mass spectrometry. ACS Meas Sci Au. 2024;4(3):233-246. doi:10.1021/acsmeasuresciau.3c00060 6. Adams C, Gabriel W, Laukens K, et al. Fragment ion intensity prediction improves the identification rate of non-tryptic peptides in timsTOF. Nat Commun. 2024;15(1):3956. doi:10.1038/s41467-024- 48322-0 7. Yilmaz M, Fondrie WE, Bittremieux W, et al. Sequence-tosequence translation from mass spectra to peptides with a transformer model. Nat Commun. 2024;15(1):6427. doi:10.1038/ s41467-024-49731-x 8. Bittremieux W, May DH, Bilmes J, Noble WS. A learned embedding for efficient joint analysis of millions of mass spectra. Nat Methods. 2022;19(6):675-678. doi:10.1038/s41592-022-01496-1 9. Dens C, Adams C, Laukens K, Bittremieux W. Machine learning strategies to tackle data challenges in mass spectrometry-based proteomics. J Am Soc Mass Spectrom. 2024;35(9):2143-2155. doi:10.1021/jasms.4c00180 ABOUT THE INTERVIEWEE Dr. Wout Bittremieux is an assistant professor in the Adrem Data Lab at the University of Antwerp, Belgium and a worldleading expert in computational mass spectrometry. He leads a research team that develops advanced artificial intelligence and bioinformatics tools to analyze mass spectrometry-based proteomics and metabolomics data. LAB AUTOMATION & DIGITALIZATION 41 TECHNOLOGYNETWORKS.COM CONTRIBUTORS Blake Forman Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He holds an honors degree in chemistry from the University of Surrey and an MSc in chemistry from the University of Southampton. Bob McDowall, PhD Bob is an analytical chemist with over 50 years’ experience and has been involved with process redesign, specifying and implementing laboratory informatics solutions for over 40 years and has nearly 35 years’ experience of computerized system validation in regulated GXP environments. Isabel Ely, PhD Isabel is a Science Writer and Editor at Technology Networks. She holds a BSc in exercise and sport science from the University of Exeter, a MRes in medicine and health and a PhD in medicine from the University of Nottingham. Her doctoral research explored the role of dietary protein and exercise in optimizing muscle health as we age. Kate Robinson Kate Robinson is a science editor for Technology Networks. She joined the team in 2021 after obtaining a bachelor's degree in biomedical sciences. In her role as science editor, she supports the publication’s in-house writers, produces scientific content across all communities and works closely with the managing editor to help coordinate commissioned pieces. Mahboubeh Lotfinia Mahboubeh Lotfinia works as a Qualified Person and Quality Partner at F. Hoffmann-La Roche and is trained in GMP/GDP audit execution and CSV (Computerized System Validation). Neil Versel Neil Versel is a healthcare and life sciences journalist, specializing in bioinformatics, information technology, genomics, patient safety, healthcare quality and health policy. Versel has been covering healthcare since 2000, across a wide range of publications in the US, Canada and Europe.  Phil Williams, PhD Phil Williams holds a BA Hons and PhD degree in chemistry and is an independent LIMS consultant with over 40 years of experience in LIMS, lab automation and thermal analysis. On LinkedIn, his group “LIMS4U” has achieved over 5,200 members, with Williams himself amassing over 30,000 followers. RJ Mackenzie RJ is a freelance science writer based in Glasgow. He covers biological and biomedical science, with a focus on the complexities and curiosities of the brain and emerging AI technologies. RJ was a science writer at Technology Networks for six years, where he also worked on the site’s SEO and editorial AI strategies.