SwRI novel software tool wins R&D 100 award
GCxGC-MS instruments detect and identify chemical substances within a test sample. Floodlight is a game changer for processing non-targeted analysis (NTA) data. In NTA, all substances detected by the instrument are evaluated, producing vast amounts of data. That's in contrast to traditional, targeted analyses in which only specified chemicals are sought for evaluation, producing more limited results. SwRI invested internal funding to develop the Floodlight technology.
"Known chemicals in a particular sample are relatively easy to find and quantify," said Dr. Kristin Favela, an analytical chemist specializing in chemical forensics and one of the leaders of the development team. "NTA is a different story. Through an extensive, multiyear NTA program of consumer products we discovered only about 20% of identified chemicals were listed on product labels or datasheets. The remaining 80% were previously unidentified in these products. We don't know the possible health effects of exposure to chemicals if we don't know they are in products."
Without artificial intelligence algorithms, the vast amount of NTA data collected requires careful, manual examination by a chemist to identify and exclude data artifacts, flaws in the data caused by equipment, techniques or conditions. Automating this data quality review with Floodlight was necessary to develop a viable analytic process.
"The key to the Floodlight software is artificial intelligence and machine learning algorithms that enable advanced analysis of chemistry big data," said Dr. Keith Pickens, one of the leaders of the Artificial Intelligence for Mass Spectrometry (AIMS) group at SwRI. "The Floodlight software is a sophisticated software tool that can make sense of the vast amounts of data NTA generates."
SwRI is well positioned to attack this issue, with wide ranging expertise in analytic chemistry, machine learning, data science and engineering. Key to the Floodlight technology is SwRI's experience in labeling ground-truth data for chemicals in a wide array of consumer products, ranging from food and medicine to packaging and toys. SwRI's machine learning team, led by Michael Hartnett and David Vickers, leveraged this data to train and develop new algorithms.
"Using this automated signal quality review method, analytical chemists spend less time on a tedious, yet demanding task, while maintaining accuracies comparable to human experts," Hartnett said. "This frees chemists to better utilize their expertise to characterize unknown compounds and draw conclusions about the chemical samples."
Chemical data analysis is used in analytical chemistry, environmental testing and manufacturing. It provides the most valuable chemical information, helping guide the decision-making process within the chemical context. SwRI's chemical data analysis solutions range from chemical testing to data analysis and software development.
"We are committed to solving difficult technology challenges with innovative approaches," said SwRI President and CEO Adam L. Hamilton, P.E. "I am honored that SwRI remains on the forefront of technological advancement. It is an incredible honor to be recognized for having three of the top 100 most significant innovations of the year."
In addition to Floodlight, SwRI-developed technologies Catalyzed Diesel Exhaust Fluid, or Cat-DEF, and Eco-Mobility with Connected Powertrains were recognized at this year's innovation awards competition. The R&D 100 Awards are among the most prestigious innovation awards programs, honoring the top 100 revolutionary technologies each year since 1963. Recipients hail from research institutions, academic and government laboratories, Fortune 500 companies and smaller organizations. Since 1971, SwRI has won 50 R&D 100 Awards.
Floodlight was developed by SwRI's Artificial Intelligence Mass Spectrometry (AIMS) group, which uses data science and analytical chemistry to advance mass spectrometry data analysis.
For more information, visit https://www.swri.org/industry/chemical-analysis-services/mass-spec-data-analysis.
Provided by Southwest Research Institute