Ali Shojaie shares how biostatistics is key to protecting human health

^{Biostatistics Professor Ali Shojaie in the Hans Rosling Center for Population Health. Photo by Elizar Mercado.}

A few years ago, Biostatistics Professor Ali Shojaie was approached by a team from Baylor College of Medicine who asked for help analyzing data to understand mechanisms of an aggressive form of prostate cancer. The team had data that they hoped contained clues for understanding drivers of this cancer and, ultimately, combating it.  

The data was a rich source of information about the tumors and nearby benign cells of cancer patients. It contained information on how genes were expressed as well as metabolomics, which are a snapshot of tiny molecules at the end of a metabolic process. The problem was that patient data were not completely aligned; data from some included gene expression, but didn’t have metabolomics data, and vice versa. 

Typically, researchers would discard mismatched data sets like these. But because the sample size was small, the team didn’t want to throw out any data. This posed a unique challenge for Shojaie and his team: Was there a statistical method they could develop that could analyze the data set, even with its mismatches? 

Shojaie, who works at the University of Washington School of Public Health (UW SPH), regularly uses statistics to find answers to these complex health problems. If a statistical method doesn’t exist for solving a problem, Shojaie and his team will try to create one.  

In this case, that’s exactly what they did. Using the data available, Shojaie and his fellow biostatisticians came up with a new method of analyzing the mismatched data to understand the drivers of this castration-resistant prostate cancer. Eventually, this helped the Baylor team develop a new treatment option for this cancer. When tested in mice, the treatment significantly increased their survival. 

“Biostatistics is an interesting combination of scientific research, statistical research, and methodological research, and I find that really rewarding,” Shojaie said. 

Shojaie has been a leader in developing biostatistical methods for helping us understand human health, from predicting the risk of cardiovascular disease, to understanding how dementia develops, to tracking how the environment impacts health. Through the use of statistics, machine learning and artificial intelligence, Shojaie collaborates with interdisciplinary teams to understand the body’s series of complex, interworking systems. 

“Biological systems are noisy and there’s less certainty with them,” Shojaie said. “We can bring rigorous statistical thinking and modeling into understanding these complex and connected systems which is important for understanding how our bodies function and what goes wrong in disease conditions.” 

Shojaie’s interest in systems began as a young engineering student studying in his home country of Iran. While his first job was in the auto industry, he soon wanted to study more complex systems.  

After he and his wife moved to the United States for graduate school, Shojaie, who had been considering a doctorate in engineering, took a few statistics classes. In a machine learning course, Shojaie was able to use a statistical model to solve a complex problem with more accuracy than some of the fancier machine learning approaches. He laughs at that memory, attributing his solution to a fluke in the data. But the lesson stuck with him:  

“Statistics is powerful in a lot of practical problems,” he said. 

Shojaie switched his Ph.D. to focus on statistics, and was able to take biology classes simultaneously thanks to a professor’s federal training grant. Studying genetics fascinated him, but he soon became curious about how other systems within the body interact with each other. How is someone’s health impacted, not only by their DNA, but also by their activity, diet, environment, and even their communities.  

Understanding these factors and interconnected systems is the focus of Shojaie’s research now. For example, he’s interested in how the digestive system is impacted not just by what type of food enters the body, but by how the gut microbiome interacts with it, how genetic factors unique to each individual process that food, how factors like stress or cardiovascular health impact the digestive system, how external factors like someone’s environment impacts the types of food they have access to, or how someone’s social network impacts the way that person feels about eating food. 

Statistical methods can be developed to understand these system interactions. This creates a cycle where these new methods lead to new research questions, which can be answered with either existing statistical methods or lead to additional research questions, unlocking answers to human health along the way.  

A recent example of this cycle is Shojaie’s work with a UW research team who studies longevity in fruit flies. It’s been well documented that calorie restriction impacts health and longevity in fruit flies and mice, but researchers wanted to understand the mechanisms within the body associated with this. Understanding these mechanisms could help support human longevity. Shojaie was able to take an existing method he’d already developed and use it to look at changes in metabolic networks between fruit flies on a restricted diet and fruit flies on a regular diet. They were able to identify changes in metabolites in these fruit flies. 

This discovery then led to another project also involving collaborators from the University of Michigan that looks at how fruit flies' social networks play a part in their longevity, which could then lead to new insight into how social networks impact human health. 

“We are at a very interesting juncture where we now have data from all these systems, from what's happening in individual cells to what's happening in societies, and my hope is that in my lifetime we see a full understanding of these connections,” Shojaie said. 

Student contributions are critical to this work, Shojaie said. Graduate students are able to lead research questions, and Shojaie serves as a sounding board for their ideas. He encourages all of the graduate students he works with to pick and lead their final research projects themselves.  

“It’s really rewarding for me to see that maturation in students as they go from starting graduate school to becoming independent researchers who are the voices and leaders of the future,” he said.  

Teaching is also integral to Shojaie’s work. He directs The Summer Institute for Statistics in Big Data, an annual program from the UW Department of Biostatistics that strengthens the statistical and data science proficiency of scholars. Several hundred students from all over the world join each year for sessions that introduce biologists, quantitative scientists, and statisticians to modern statistical techniques for the analysis of biological big data.  

“That's an important component of practice and the mission we invest in: to develop expertise in local communities in different parts of the world,” Shojaie said. “That's how we make sure we do better science and use resources better in a way that benefits the society.”