If you're reading this, your body probably contains forever chemicals.
Forever chemicals, also known as PFAS (short for per- and polyfluoroalkyl substances) can be found in more than 90% of Americans, who are exposed to the chemicals in consumer products, food, and drinking water.
In the last two decades PFAS have been linked to multiple health problems, including some cancers, developmental delays and behavior problems in children, decreased fertility, weaker immune responses, and obesity.
The biggest source of PFAS exposure for many people comes out of their kitchen tap, with nearly half of U.S. municipal water supplies known to contain PFAS.
As states act to reduce PFAS in drinking supplies, a team of Columbia and City College scientists, led by Virginia Cornish, professor of chemistry and systems biology of Columbia University, are working to make the process easier and more efficient.
The team is building a "biohybrid" device-using engineered yeast, AI-designed proteins, and complementary metal-oxide-semiconductor (CMOS) integrated-circuit sensor chips-that will help municipal water systems detect the levels of forever chemicals and monitor their removal.
Why is a new PFAS detector needed?
Under the new EPA rules, municipal water suppliers are required to test for PFAS frequently, but current PFAS detection methods are costly. Most suppliers will need to send samples to commercial labs that use expensive mass spectrometry machines to measure PFAS concentrations.
Delays in receiving results-commercial testing sometimes takes weeks to send reports-will also be a problem for municipal systems that must remove PFAS. These systems will need real-time results to closely monitor the removal equipment and determine when components need replacement.
"There are approximately 150,000 public water systems in the United States, and PFAS monitoring is becoming a huge cost for them," says Alex Rosenthal, professor of civil engineering at City College. "We've talked to several dozen water industry professionals around the country, and there is unanimous excitement about how our new technology can optimize operations."
Why can yeast and artificial intelligence do better?
Biology can provide a cheaper, faster way to detect chemicals.
Over the years, the Cornish lab has perfected a way to turn Saccharomyces cerevisiae, the same yeast species used by bread makers and beer brewers, into single-celled versions of the sniffer dogs at airports that can detect explosives or illegal drugs hidden in luggage. "The dogs detect chemicals from these substances because their noses have odor receptors that bind to those chemicals," Cornish explains. "With genetic engineering and synthetic biology, we can design the same type of chemical-detecting receptors and implant them into yeast."
In the Cornish lab, a protein that detects a specific chemical is first designed and then inserted into yeast cells. These synthetic receptors are connected to a fluorescent protein, also inserted into the cells, so that when the synthetic receptors bind to the chemical of interest, the yeast glow green. The intensity of the glow is related to the chemical's concentration.
Before artificial intelligence, it took the Cornish lab a few years to design the synthetic receptors. Now, with the help of an AI tool developed by Mohammed AlQuraishi's lab in the Department of Systems Biology, they are working towards a design-build-test cycle of only weeks.
With AI, we should be able to detect any chemical we want, which we couldn't do before with chemistry alone," Cornish says. "It will be a huge breakthrough for the field of sensing and could lead to a new industry of yeast-based sensors in health care."
Computer-brain interface chips report results
The Cornish lab is currently developing an AI design-build-test cycle to generate yeast that report on PFOA and PFOS compounds.
Once the yeast are engineered with the sensor and fluorescent proteins, they will be freeze-dried and embedded in a material resembling pH paper.
CMOS sensor chips-originally developed as wireless computer-brain interface devices by Ken Shepard's lab in the School of Engineering and Applied Science-read the fluorescence and convert it to a concentration. These new optoelectronic sensor chips make it possible to create a handheld device that can report PFAS concentrations in real time.
To measure PFAS, users will simply dip a disposable "stick" containing the yeast and sensor chip into a water sample and get an instant readout of the PFAS concentration range. The stick is read out wirelessly by a small reader unit.
Cornish estimates that the readout devices and disposable sticks will cost orders of magnitude less than current methods and could save hundreds of thousands of dollars for each municipal water system.
"These devices are example of biohybrid systems, systems in which engineered biological systems are combined with semiconductor sensors and actuators to create a system with a function beyond what is possible with solid-state or biological system alone," says Shepard.
