New approaches, better insights
Dr. Kovats was particularly impressed by the methods of this study. “It was a natural study, the kind of thing we can usually do only in mice,” she said.
Oklahoma Medical Research Foundation
Dr. Susan Kovats
“One problem with studies on the effects of hormones in disease is that historically researchers have not paid that much attention to the actual hormone levels in the humans they studied. They might look at 100 women and 100 men, roughly between the ages of 20 and 50. We’re starting to see more, but there aren’t a lot of studies correlating numbers of cells in blood with actual hormone levels in the person. And as we know, just because someone’s a certain age doesn’t mean that they have a textbook hormone level. Early menopause, birth-control pills, many things can affect those levels.”
The researchers hope that these findings will shed light on the mechanisms that create sexual bias in autoimmune diseases, particularly lupus, as well as help researchers to better understand the innate and adaptive immunological differences between men and women. It could also be useful in the clinical setting, Dr. Robinson said. Because of the extreme sex bias in lupus, doctors see far more women with the illness than men. When they do see men with lupus, they need to be able to consider how the patient’s sex affects the development and course of the disease. “I think that people need to start looking at patients as clinically different, depending on their sex and gender,” he said. Information like that analyzed in this study could help with that. This could be especially important because as Dr. Kovats pointed out, although men get lupus far less often than women, when they do have it, they tend to have more severe disease.
Help from machines
This study was groundbreaking in another area as well. The researchers used machine learning to analyze the data. “We’ve started working a lot more with these analysis methods to try to answer as much as we can with these smaller data sets,” Dr. Robinson said. “Rather than the conventional analysis that we would typically perform, we’re able to use machine learning and artificial intelligence to try and learn from the data and increase the numbers that we’re working with by using a training data set. This allows us to interrogate the data with a lot more precision.”
The authors declared no competing interests.