How has your job description changed over the years? Thirty years ago, odds are the responsibilities of someone in your role looked a lot different than yours do today. Add in the monumental advancements of technology in this generation and the unprecedented global pandemic, and the picture changes even more.
Sridhar Iyengar, CEO & founder of Elemental Machines, a company that manufacturers sensors for real-time data monitoring, thinks the role of laboratory manager is one of those job positions with shifting modern responsibilities.
As a serial entrepreneur, Iyengar has extensive experience with data and devices, founding Misfit, a company that makes wearable products, as well as AgaMatrix, a blood glucose monitoring company that made the world’s first medical device to directly connect to an iPhone. The more time Iyengar spends in the life sciences, the more he sees parallels to the high-tech industry.
Editor-in-Chief Michelle Taylor recently spoke to Iyengar about the changes he’s seeing in life science and pharmaceutical laboratories—both expected and unexpected.
Q: In your experience over the past year or so, how has the COVID-19 pandemic altered the laboratory equipment marketplace?
A: COVID-19 has affected not only the laboratory equipment market, but the laboratory operations market, as well. If you can work off a laptop behind a desk, then you can work from home. But if you are in a lab, you have biologics, reagents, chemicals, assets—you need to be physically there. We’ve seen a few things shift. Anything that is superfluous to the lab’s mission is being sidelined. Labs are being opened up only to prioritize those critical workflows.
We recently did a study that looked at instrument utilization through our platform. We saw pretty consistent instrument usage from 9 to 5 pre-pandemic, but as soon as the pandemic happened, that spread out up to almost 24 hours. You would see people doing work remotely at all hours of the day to makes sure they have eyes and ears inside the facility because they knew they couldn’t be there for an extended period of time. Utilization and asset management were immediately affected, and the impact of that is calibration and maintenance schedules have to be readjusted.
Q: Lab managers play an important role in pretty much everything you just mentioned. How has their role evolved in the last 20 years or so?
A: I think the term lab manager is going to be expanded into lab operations. In the world of software and IT, we had a very important shift over the course of 10 to 15 years, and that’s when system administrators became known as developer operations, or dev ops. What was considered administration became operations. That’s interesting because I am seeing parallels with what we’re experiencing in life science. What this new position entails is a diverse set of skills and knowledge of integrations.
Lab managers schedule maintenance, order consumables, manage calibration, etc., but today the beneficiaries and users of labs are not just the bench scientists. It’s also data scientists, who need information access, and facility managers, as well as QA/QC and regulatory personnel. So, you have a confluence of many different stakeholders. That’s why we are seeing today’s lab manager as more of a laboratory operations engineer.
Q: For biopharmaceutical labs specifically, what shifts have you seen in the last decade in terms of how they are approaching drug discovery?
A: There has been a very noticeable shift in the last 5 to 10 years, and it has to do with the use of machine learning and AI techniques. If you look at the biopharma companies of yesteryear, the people who were founders, executives and chiefs were trained as classical scientists. They would then run a very hypothesis-driven organization, hiring data scientists, statisticians and engineers to crunch the numbers. Once you understood the science, you could move to the next step and the one after that.
What’s happened now is labs are employing machine learning and AI, which means the founders, executives and chiefs are now data scientists, computer scientists and data engineers who have a data-first approach. They prioritize the data side of it so the AI funds the data, and then they bring the scientists on board to gather that data. It’s the same set of people, but the leading roles have been reversed.
Q: Some of those changes can definitely be attributed to the rise of “The Smart Lab.” How can lab managers and staff adjust to a more data-driven environment?
A: The focus should be on obtaining structured data and having it be interoperable between different systems. When you look at a modern laboratory, there are hundreds of different vendors. Different companies make different machines and they don’t often talk to each other, if at all. One of the first questions lab operations people need to answer is, “how is the output of this asset going to be accessible to everyone?” You also have to think about a systems approach from end to end—"how are we doing to get data from a beaker or a pipette into our analytics tool?” Taking an overall approach allows purchasing and management decisions to be made based on total cost of ownership.
Q: What are common challenges of data management in the lab?
A: Information security is often top of the mind. It’s well and good to have data interoperability, but you also need to work with the relevant IT and security stakeholders to make sure that data is accessible and secured by the right type of vendor. I’ve walked into so many labs where there’s still a Windows 2000 machine hooked up and no one touches it because it still works. But what that does is stymies some of the rapid growth. The nature of software and information technology is always evolving, and that needs to be balanced with things like cloud security and information access, including bring your own device. The risk versus reward shifts whether you are an early-stage or late-stage company.
Q: How important are partnerships between hardware and software companies?
A: Alan Kay once said, “if you really care about the software, you need to build your own software.” In the world of laboratory science, there is an innate need for physical instruments and the software that runs them, so that’s already there. What is not there is the connectivity to the cloud. There are hundreds of millions of scientific instruments in the U.S. that are not connected to anything. You don’t swap out lab equipment every couple years like you do your phone. An HPLC made 20 years ago is still good, you just have to maintain it; the same with a balance.
For hardware and software manufacturers, it comes down to understanding that anything you bring into this ecosystem must work with systems that are 20 years old and will have to have a shelf-life of another 10 to 20 years itself. The way to do that is to do minimal amount of work on the hardware and push all processing onto the software, which can be updated on the cloud in the future.