To pass the torch to the 2017 winner of the Centre de recherches mathématiques (CRM) and Statistical Society of Canada (SSC) Prize in Statistics, Professor Radu Craiu, the 2016 awardee, simply has to reach across the dinner table.
Craiu, a professor in the Department of Statistical Sciences, is married to this year’s prize winner, Professor Lei Sun, a statistician who is cross-appointed to the Dalla Lana School of Public Health.
They’re the only couple to receive Canada’s most prestigious mid-career statistics award, a prize that recognizes a statistical scientist’s professional accomplishments in research during the first 15 years after receiving a doctorate.
“We were both in contention last year and it made for a bit of tension, but I am glad Radu won,” said Sun with a smile. “I was still eligible this year — one of the benefits of having taken two maternity leaves. University Professor Nancy Reid, who nominated me, said it was good for domestic peace that I won this year.”
It’s clear that each is delighted for the other. Craiu, who is from Romania, and Sun, who is from China, have been together for 20 years, ever since they met in graduate school at the University of Chicago. They are celebrating their 15th wedding anniversary this year.
“It was love at second sight,” joked Sun. “We were so busy during the first quarter at graduate school that we barely noticed each other. In the second quarter, we were assigned to be teaching assistants for the same course and then it was love at first sight.”
As they finished their doctorates, the pair began searching for two tenure-track openings at the same university, which was a challenge, given that they work in the same field. Fortunately, while Craiu accepted a position with the Faculty of Arts & Science, Sun, whose research focus is statistical genetics, had an opportunity at U of T in biostatistics of public health.
“We didn’t plan for Canada, but we’re very happy we ended up here,” said Craiu. “We love Toronto and we like the atmosphere in our department; it’s very collegial.”
Craiu works largely with algorithms and with statistical modelling in instances when there is a complicated interaction between variables that are connected to, or dependent on each other. As an example, he uses a study that examines a group of women who exercise, have quit smoking and have lost weight.
“What is the dependence between the weight loss and quitting smoking?” he asked. “How does weight loss change with the number of hours the women spend in the gym? I try to create models that study such interrelation.”
His work with algorithms is increasingly important in the era of big data, since many statisticians are doing large-scale computations.
Sun’s research focuses solely on statistical genetics. She uses statistics to try to find the genetic variations responsible for influencing human traits, such as complications of Cystic Fibrosis and diabetes. She and a colleague created a general resampling-based solution to reduce selection bias, the so-called “winner’s curse”: an upward bias that inflates the estimated effect of a newly identified genetic variable on a specific disease.
“Once you say a certain variable influences the likelihood of disease, you tend to self-select and only report those instances that support your hypothesis; that creates a certain bias that is a key contributing factor to the irreproducibility issue,” Sun said. “The question is how to correct for that bias without collecting additional data.”
Sun, Craiu and other colleagues also collaborated to develop two concepts that have had an impact on whole-genome association studies: the stratified false discovery rate (sFDR) control and the non-discovery rate (NDR) for false negatives.
“Working on our first paper together was difficult,” said Craiu. “We learned that we needed to separate dinner table talk from professional talk.”
In fact, they don’t usually discuss their work at the dinner table, because their two children “are allergic to statistics,” Craiu said.
However, it’s not a challenge to arrange to have professional discussions: these two CRM-SRC Prize winners have offices about 10 metres apart. Collaborating both at work and at home has clearly been a success.