Smruthi Karthikeyan, the Gordon and Carol Treweek Assistant Professor of Environmental Science and Engineering and a William H. Hurt Scholar, describes her research as "harnessing the power of microbes for human and environmental health." Karthikeyan takes an engineering approach to communities of microbes that live inside the human gut, in agricultural soil, and in the ocean, seeking not only to understand microbial behavior, but also to use microbes to improve outcomes for medicine, agriculture, and environmental sustainability.
Karthikeyan first came to Caltech in 2022 as a visiting associate, and she officially joined the Institute's faculty in 2023. While waiting for the completion of her research space, Karthikeyan has begun staffing her lab, mapping out her research agenda, initiating collaborations with other research groups on campus, and running computations for experiments being conducted in other labs.
Here, Karthikeyan shares her thoughts on her career to date and her future research goals.
How did you become interested in researching microbiomes?
I actually started out in the environmental sector. My undergraduate degree was in chemical engineering. I had no background in biology because the modeling I was doing in engineering was more math based. But when I started graduate school at the Georgia Institute of Technology, my principal investigator (PI) was a molecular microbiologist. When I first met with him to discuss possible synergies or possible applications for my engineering work, he said, "You know, you can do something completely different if you want." He then personally trained me in the wet lab. I had to do several minors just to keep up with the microbiology work I was doing during my PhD program.
One of the things I looked at in my graduate research was marine oil spills. I searched for microbes that could break down oil and asked how their presence altered the previously existing microbiome. I wanted to know what would happen to the oil-eating microbes after the oil was gone, and if the microbiome that existed before the spill would come back, or if it would be significantly altered.
What led you to explore the human microbiome?
I was interested in different applications for the work I had been doing in the marine microbiome. In environmental systems, microbes interact with each other and their environment extensively. I began to wonder if we could apply the principles we use in environmental ecology to gain a fuller understanding of microbial interactions in human gut ecosystems.
I had taken a lot of courses in computational biology and microbiology in graduate school, but I had never done anything medical. For my postdoctoral work, I decided I should go to a medical school where I could learn the conventional approach to the gut microbiome and see if there was space for alternatives. So, I applied to UC San Diego's medical school for a postdoctoral fellowship and was accepted.
What sort of research did you do during your postdoc?
That was driven mainly by circumstance. I got the job at UCSD in 2019, but I wasn't going to start until after my graduation from Georgia Tech in 2020. I moved to San Diego and, suddenly, everything was locking down because of the COVID-19 pandemic.
Since this was a medical school involved in clinical work, people turned their attention to this new virus. Everyone was doing wastewater research, measuring the COVID virus in wastewater using PCR [polymerase chain reaction] techniques. This is a very useful technique, but rather slow. I started to wonder if we could track mutations or see how the virus was spreading using other methods, and if it was possible to do this more quickly and efficiently.
I applied for federal funding but was rejected. They said the idea was too ambitious and that there wasn't enough precedent to show that this approach could really work. But my PI suggested we apply for funding through the chancellor, and we were given a grant to cover my work and that of an undergraduate assistant for one year.
This research was outside the lab's wheelhouse, but my PI told me I could do whatever I wanted, though he warned me that if my approach didn't work out, I should be OK with not having much data to show for six months of work.
It seemed like a fair trade off. We were in a pandemic; I wanted to contribute, and with the lockdown, I couldn't do my planned research anyway.
How did the wastewater project turn out?
We were very successful. We built robots that could sample wastewater, and we deployed them to many buildings at UCSD to help track COVID infection among the 10,000 students who were still living on campus and the 25,000 workers who were coming to campus regularly. As it became clear that the technique was working, the robots were also deployed to the Point Loma Wastewater Treatment Plant in San Diego and at more than a dozen public schools.
We were able to process more samples more quickly using a computational tool that tracks genetic markers. We screened for new variants that were at first only a very small percentage of the total COVID viral load. In fact, we detected the Delta and Omicron COVID variants nearly two weeks before they were found in clinical settings.
Did you want to continue with epidemiology through wastewater screening?
No. A lot of people felt I should. But I didn't want to just do research that was incremental. Having already built the system, I figured it would be easy enough for other people to build on that. I wanted to build something new.
This is what really attracted me to Caltech. People here didn't try to tell me I should keep working in the same area of research where I had been publishing and where it might be easiest for me to get grant funding. At Caltech, they don't try to define you. They just say, "Go ahead, do cool science."
Do you see a through line in your research that bridges different disciplines?
Yes, our overarching theme is to see if we can reimagine the microbiome landscape. Most studies of the gut microbiome, for example, use high-throughput sequencing on fecal samples, but this is only a fraction of the gut microbiome's composition. We need to understand more about how microbes function in their natural context. We need to measure their activity patterns, so that we can understand the whole ecosystem.
Up to 99 percent of environmental microbes cannot be cultured in the lab. Sometimes, this is because a particular microbe may be living off the metabolites produced by a different microbe, and it can't survive as a separate entity. And, a lot of times, when you look at microbes in isolation, or you engineer them, they can't survive in the world when you release them. In the lab, microbes don't have competition. It's not a natural environment.
We develop a mix of experimental—imaging and mass spectrometry—and computational pipelines to look at microbes on a larger scale and understand their underlying mechanisms in community structuring. To be able to effectively engineer solutions, it is crucial to look at the fundamental underpinnings that govern microbial community structure and in-situ evolution. We call this systems microbiology, because we're not looking at a model microorganism, but rather at the entire microbial ecosystem. We want to see why a microbe thrives in a particular environment. For example, is the host providing something it likes?
One of the things we're working on now is soil microbes. When you think about agricultural work aimed at increasing drought resistance in crops, the usual approach is to genetically engineer the plant. But we're looking at the agricultural system from the soil perspective, asking how we can leverage the soil microbiome to help plants be more drought resistant.
We're also looking at developing integrated computational and isotopically labeled microspectroscopy approaches for detecting metabolic determinants of antibiotic efficacy in the gut microbiome.
What are you teaching at Caltech?
I've developed a new course that I'll offer this winter. It's a 50/50 theoretical and computational course: Computational Tools for Decoding Microbial Ecosystems. Every week, we will start with a theory that explains the underlying algorithms. Then, later in the week, we will take real datasets and work on them in class. We'll be looking at how we can recognize patterns and sequence data for large-scale microbiomes and make interpretations.