(AGENPARL) – STANFORD (CA) lun 24 aprile 2023

When Purvesh Khatri was a high school student in India, he wasn’t shy about his disdain for biology. “I really hated it,” said Khatri, PhD, now an associate professor of medicine and biomedical data science at the Stanford School of Medicine.

The irony of his career arc — which currently has him sleuthing out genetic signatures of various biological or disease traits — isn’t lost on Khatri. He never envisioned himself digging deep into biological data. For a while, he was set on becoming an engineer — but not because he had great interest in pursuing that field, either.

When Khatri was a young adult in India in the 1990s, his mom told him there were only two acceptable career paths if he wanted an arranged marriage. “She said I could be a doctor or an engineer, or no one would want me to marry their daughter,” recalled Khatri. “And since I hated biology, I went the engineering route.”

Fortunately for Khatri, the lesser of those original evils led him to a life partner — he met his wife while tutoring her for an engineering class they shared. And though he got “bored” of engineering, it came in handy for his second career: The nexus of biology and computational science, where he applies data science to immunology and genetics.

Some 20 years after his first foray into biology, Khatri is now a leader in computational immunology, a field in which researchers use massive amounts of data to reveal and predict how the human immune system works and how it will act when confronted with disease or infection.

Among his research feats, he’s devised a genetic test that detects sepsis — bodywide inflammation caused by infection — early, an assessment that gauges a person’s immune response as signals of tuberculosis infection, and a test that helps doctors predict which patients with pulmonary fibrosis, an often lethal lung disease, are at most severe risk for lung failure. And it was all done by sifting through and gleaning insights from enormous amounts of data.

Someone now at the forefront of this important convergence of data and medical science once wanted little part of either — and doubted that he belonged among the leaders of the movement. An ability to adapt to new realities has taken Khatri on quite the journey.

Getting a master’s…er, PhD

Khatri moved around a lot when he was a kid because of his dad’s job, which involved fixing failing banks. But what some might see as an early life of instability, Khatri internalized as a means to hone agility and adaptability — key life skills that would serve him well.

After finishing his undergraduate degree in engineering at Saradar Patel university in India, Khatri harnessed that adaptable spirit and made the 8,000-mile move to the United States to earn his master’s degree in computer science at Detroit Wayne State University. The decision presented immediate obstacles. “I couldn’t afford the tuition. My dad took an early retirement and gave me what he could,” Khatri said. “But it wasn’t even enough for two semesters tuition.” He was committed, though: He worked jobs on campus while attending school.

Three months into his pursuit of a degree, Khatri hit a slump. “I couldn’t go the software route. It was just boring to me,” he said. Fortunately, his advisor was also looking to change course to pursue bioinformatics. He told Khatri he would keep mentoring him and funding his research. “And I told him, ‘If you’re funding my work, I’ll do whatever you want,'” Khatri said. So, Khatri became a budding bioinformaticist.

While working toward his master’s degree, Khatri was particularly prolific in publishing papers. His advisor suggested he keep on the research track and pursue a PhD. Game to level up, he completed his doctoral degree in 2006, establishing a data analytics platform that, unbeknownst to him, would clinch him a spot as a postdoctoral scholar at Stanford Medicine, another 2,400 miles across the map.

A Stanford Medicine professor had used a tool Khatri developed to analyze his data, and he had heard through the science grapevine that Khatri was looking for a position in research. That researcher, Atul Butte, now a professor of bioinformatics at UC San Francisco, offered Khatri a position in his lab. But Khatri had one small request before packing up and moving across the country.

“I wanted to put my skills in computational biology to the test,” Khatri said. “Whatever I did computationally, I wanted there to be a path to validation in some experimental system.” Khatri joined two labs — one that harnessed computational biology to analyze data, and one that focused on organ transplantation. Khatri’s first big research question centered on the role of the immune system in organ transplantation, how and why organ rejection happened in patients, and how new drugs could curb such rejection.

Computational immunology

With a foot in both camps — computational and biological — Khatri began to explore new uses of data to understand how the immune system worked. As he dug deeper into the niches of immunology, Khatri’s data science background kicked in. He recognized the immune system as a complex piece of hardware, and the differing immune cells as various sensors that picked up signals to turn the hardware on or off. That kind of search-and-alert system is called a distributed sensor in the engineering world, and it seemed to Khatri uncannily similar to how the immune system works. Immune cells patrol the body for signs of foreign invaders or microbial dangers, and when they detect a threat, they send signals to activate the immune system’s defenses.

“If there’s one thing I’m proud of, it’s that,” said Khatri. “Using my engineering background to help create a new way of thinking around the immune system.”

Khatri began to think about the immune system as its own diagnostic tool. Immune cells act up only if something’s amiss, and he found that certain genetic signatures associated with the immune system could act as a marker for different types of immune activity. Keen to apply this line of thinking to better understand how the immune system responds after a solid organ transplant, he analyzed gene activation post-transplant, discovering a set of 11 genes that, when activated in a specific way, flagged organ rejection with the same level of accuracy as a physician. And it could all be done through a simple blood test. That test was able to predict organ rejection before doctors were able to detect it through a biopsy.

As Khatri’s work as a postdoc progressed, others in his field took notice. Butte and others encouraged Khatri to pursue this marriage of computation and immunology and to grow the nascent field. Butte also told Khatri he should apply for a faculty position. “But I didn’t apply for it. I didn’t think I was worthy of the position,” Khatri said. “I didn’t think I had the track record to be a faculty member at Stanford.”

Then Khatri received a note informing him that the application reviewers were awaiting his materials. His adaptable spirit kicked in and he scrambled to put an application together overnight. “It was way past the deadline,” he recalled.

That was 10 years ago, and suffice to say, Khatri earned that faculty position. The vision he has for his lab — harnessing massive amounts of data to glean insights into the human body and predict how it will act and respond — has produced more than 120 papers and more than 30 patents. He’s uncovered genetic signatures that delineate viral infections from bacterial infections as well as a biomarker that predicts an individual’s flu susceptibility, and he has used data to show how drugs already approved by the Food and Drug Administration could be repurposed to fight other conditions.

“My lab doesn’t use beakers or microscopes; we don’t do experiments. We repurpose what’s out there,” Khatri said. “The whole idea is to leverage publicly available data to advance human immunology and improve patient care through the right diagnosis at the right time.”

Photo by Norbert von der Groeben

The post Unconventional Paths: Merging computation and biology appeared first on Scope.

Fonte/Source: https://scopeblog.stanford.edu/2023/04/24/unconventional-paths-merging-computation-and-biology/