For Olukayode Sosina, working as a biostatistician often means playing the role of a prosecuting attorney. His team at the Regeneron Genetics Center is tasked with combing through terabytes of genomic data in search of needle-in-the-haystack mutations that might light the way to new medicines. Once they’ve found one, Sosina embarks on an iterative process of cross-checking all the ways it might just be statistical noise, accusing the data of perjury over and over until there are enough alibis to make it bulletproof.
“You validate and validate and validate,” Sosina said. “And then, with more and more confirmations of your validation, you reach a point where it becomes: Oh, wow, this is actually something that makes a ton of sense. You start getting this rising feeling of excitement.”
“What I love about statistics is there are scenarios where you might have certain fallibilities as a human,” Sosina said. “Maybe you have a bias, or maybe you have certain things you expect to see. And this is a way to clarify what is actually going on.”
That happened in 2021, when Sosina and his Regeneron colleagues used sequencing data from more than 600,000 people to find that rare mutations to a gene called GPR75 were strongly associated with protecting carriers from obesity. Earlier this year, the group looked at data on more than 500,000 people to conclude that mutations to the gene CIDEB substantially decreased the risk of liver disease. Each discovery survived scientific cross-examination, and each has since graduated to Regeneron’s drug development division, which will find out whether Sosina’s work can translate into new medicines.
The promise of isolating the truth from numerical chaos is what drew Sosina to biostatistics in the first place. After graduating from the University of Lagos with a degree in mathematics, he went to Harvard University for a master’s and eventually Johns Hopkins University for a Ph.D. in biostatistics. In 2013, when the Regeneron Genetics Center began, Sosina was its first hire.
“What I love about statistics is there are scenarios where you might have certain fallibilities as a human,” Sosina said. “Maybe you have a bias, or maybe you have certain things you expect to see. And this is a way to clarify what is actually going on.”
— Damian Garde
