Innovations in LIMS for Enhanced Laboratory Efficiency

1 Listicle 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. 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 Innovations in LIMS for Enhanced Laboratory Efficiency Bob McDowall, PhD INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 2 Listicle 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: ◦ Type (e.g., pipette, flask) ◦ Size ◦ Grade (A or B) ◦ 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 out-of-specification results. It also enables tracking of 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. INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 3 Listicle 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 INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 4 Listicle Given the vagueness of the definition above, IoT can mean almost anything to anybody, depending on the context. 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 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. INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 5 Listicle 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. • 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: ◦ A live version that operates without self-learning so that it remains under control. ◦ 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 INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 6 Listicle 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. Figure 1. Integration of innovations in a LIMS environment to enhance laboratory efficiency. Credit: Bob McDowall. Artical Intelligence/ Machine Learning: Trending Data Sample Registration Laboratory Information Management System Sample Receipt and Storage Locations Instrumental Analysis • Samples coded • Sample storage cabinets coded • Individual locations in cabinet coded • Check in of samples • Check out of samples • Monitoring of storage conditions (temperatute & RH as needed) • Return of samples for check in — storage location updated Interpretation/ Calculation of Results • 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 Collaion and Reporting Results • Results from all tests collated for the sample • Check for Out of Specification results • Resolve any issues or deviations • Issue electronic CoA • 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 Sample Preparation • 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 • Analytical data captured electronically or by voice input • Extracts are transferred to QR coded vials • Chain of custody for preparation work On-going Product/Annual Quality Reviews Electronic Analysis (CoA) INNOVATIONS IN LIMS FOR ENHANCED LABORATORY EFFICIENCY 7 Listicle 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. About the Author: Bob McDowall is an analytical chemist who has been involved with specifying laboratory informatics solutions for over 40 years and has nearly 35 years’ experience of computerized system validation in regulated GXP environments.