Industry Perspectives

Does your lab need a data pond?

Labs generate more analytical data than ever, but real-time access remains elusive. A purpose-built data pond may be the missing link.
Written byTrish Meek
Brought to you byWaters
| 5 min read
Luminous streams of interconnected data lines and water droplets flow across a deep blue background, illustrating the movement of digital information through a networked system.

Flowing data streams with water-droplet particles illustrate the movement of analytical information between connected laboratory systems.

© iStock: Alllex

Register for free to listen to this article
Listen with Speechify
0:00
5:00

Steve Jobs famously predicted in a 1985 speech at Lunds University in Sweden that in the next 50 to 100 years, technology would evolve to the point where you could ask Aristotle a question. This vision, while still out of reach, does not feel as far-fetched as it once did.

Executives today are asking: can we use AI to query our data and get an answer? The short answer is, of course, yes. The bigger issue is whether you have the right underlying data to get the correct one. In all industries, IT organizations are aggregating data across applications, then refining and storing it in data lakes. Extensive time and effort are required to develop and maintain data pipelines, AI models, and ways to visualize the results.

Of course, none of this is real time. There is still a significant delay between data creation and its availability in the data lake. This raises a further question: how do we enable scientists to get their answers in real time? A smaller, purpose-built data lake for the lab, or data pond, may be the solution.

By aggregating and assimilating contextualized analytical data into a focused, smaller pond, scientists would have the information they need at their fingertips, and corporate data lakes would have a streamlined pipeline for scientific data. So, the final question is, does your lab need a data pond?

Why do labs need data ponds?

The concept of a data pond is not new. Prajesh Kumar used the term to describe a "series of context-centric, analytics-driven, purpose-built" data ponds that enable focused data science. Creating a single enterprise data lake that answers all data challenges and meets the needs of all aspects of an organization is incredibly difficult. Even if an organization is able to achieve this, maintaining the pipelines that keep the data relevant would be difficult and costly.

In reality, most organizations choose to have a broad data lake that holds critical information from all systems, but not all the information from any one system. They also refresh the data on a cadence somewhere between daily and weekly, and often a full database copy is required every time they refresh the data. The advantage of a data pond is that it can be incredibly deep, containing all of the contextualized data in near real time for a particular domain, such as the analytical laboratory.

The problems scientists are trying to solve are very specific and often require information that would not traditionally exist in a corporate data lake, as well as an understanding of chemistry, biology, or physics that data scientists may not have. Quality assurance/quality control (QA/QC) laboratories, for example, need tools to easily analyze their data without having to rely on their corporate data science team so they can make decisions quickly before samples degrade, or it is too late to save a production batch. Scientists don't need (and won't have access to) all organizational data, but they should have access to the relevant analytical data needed to find these answers, and a data pond makes this possible.

This is why Waters is prioritizing data access in the development of its informatics solutions: removing bottlenecks between data creation and use, improving interoperability across instruments, applications, and enterprise systems, and providing a scalable foundation for analysis across laboratories, sites, and data sources. By making contextualized data easier to access, connect, and analyze, these solutions can reduce time spent moving or preparing data and give scientists a more direct route from experimental results to informed decisions.

Connecting data ponds to intelligent agents

Data ponds are even more relevant when you consider the proliferation of agents being developed by organizations for their scientists, or by scientists directly. These agents assist scientists across a multitude of tasks, from research to routine workflows. Enterprise systems, such as Electronic Lab Notebooks (ELNs) and Laboratory Information Management Systems (LIMS), are building agentic frameworks into their platforms to simplify laboratory execution while maintaining regulatory and process control. These systems, like data lakes, have a high-level overview of all laboratory results and, crucially, the context of what that data means — for example, whether a drug candidate or formulation is more promising than others that have been tested, or whether a production batch passes or fails.

Imagine the power of connecting these systems to an analytical data pond and agents that could query both the enterprise system and the data pond to provide answers that are informed by all the scientific information about a compound or product. This is even more powerful as systems adopt Anthropic's open standard, Model Context Protocol (MCP). MCP enables connectivity between systems to move beyond fixed, customized interfaces toward intelligent communication and the ability for users to interact with data and make decisions based on it. A scientist running a mass spectrometer can now decide the next step of the experiment and communicate with the instrument software while the run is happening. Scientists have never had as many tools to deliver high-quality, data-driven results.

Almost everyone has heard mathematician Clive Humby's phrase "Data is the new oil," coined in 2006. It is often used to describe how data can enable new commercial approaches. What is often left out is data scientist Michael Palmer's elaboration on that phrase: "Like oil, data is valuable, but if unrefined it cannot really be used. It has to be changed into gas, plastic, chemicals, etc. to create a valuable entity that drives profitable activity, so, must data be broken down, analyzed for it to have value."

This second half of the quote is even more relevant today than it was in 2006. Accessing fully contextualized data remains the biggest barrier to using AI to its full potential. While many organizations have attempted to create open standards, the Allotrope Foundation's Allotrope Simple Model (ASM) is gaining traction across the industry. ASM enables access to a broad range of analytical data that can be taken out of its proprietary format and transformed into a standard that is human- and machine-readable and can be optimized for data queries.

Conversion of file types into an open standard has historically been a barrier, but this is becoming easier with AI-assisted and agentic development. By adding ASM data to your data pond, you can support data from many of the common analytical techniques using a single ontology. Through focused, deep data ponds, you can't speak to Aristotle, but you can ask your scientific data questions and be confident in the response.

Waters™ Data Intelligence Software brings this idea to life by helping laboratories refine Empower™ Software data into information that can be understood and acted upon. Purpose-built analytics and near real-time dashboards make it easier to visualize trends across projects, sites, and databases, monitor data integrity and system utilization, and uncover insights that might otherwise remain hidden in disconnected records.

By turning complex analytical data into consistent, accessible views, the software helps teams identify risks, prepare for audits, improve laboratory operations, and make confident decisions more quickly. In other words, the value does not come from storing more data; it comes from giving that data the context, visibility, and analytical tools needed to produce insight. With plans to expand beyond just Empower Software data, Waters is simplifying the creation and management of the data pond.

Building a more connected laboratory

Ultimately, the value of a data pond lies in making the laboratory's most relevant data accessible, contextualized, and ready to use. By removing barriers between instruments, systems, and scientists, organizations can support scalable analysis, enable more effective use of AI, and turn analytical results into confident decisions faster.

The laboratories that prioritize connected data today will be better prepared to adapt, innovate, and create value from every experiment tomorrow.

Add Drug Discovery News as a preferred source on Google

Add Drug Discovery News as a preferred Google source to see more of our trusted coverage.

About the Author

  • Trish Meek Waters Corporation

    Trish Meek is a Sr. Director and Portfolio Owner in the Informatics Business Segment at Waters Corporation. In her current role, she leads the product management, development, cybersecurity and cloud infrastructure for Waters next generation cloud applications. Trish has over 25 years of experience in the laboratory informatics industry.

    View Full Profile

Here are some related topics that may interest you:

Loading Next Article...
Loading Next Article...
Subscribe to Newsletter

Subscribe to our eNewsletters

Stay connected with all of the latest from Drug Discovery News.

Subscribe

Sponsored

3D illustration of Y-shaped antibody molecules floating against a soft blue background, with a detailed antibody structure in the foreground.
From molecular developability to functional activity, integrated characterization strategies help inform lead selection and optimization in biologics discovery.
3D illustration of Y-shaped antibody molecules floating against a soft blue background, with a detailed antibody structure in the foreground.
From molecular developability to functional activity, integrated characterization strategies help inform lead selection and optimization in biologics discovery.
3D illustration of a multicolored protein structure composed of folded ribbon-like strands against a blue and teal background with scattered spherical particles
Learn how advanced analytical methods support the characterization of antibody–C1q interactions and the qualification of analytical binding assays.