Posted By: Sarah Ratzel, PhD, Science Editor, AJHG
Each month, the editors of The American Journal of Human Genetics interview an author of a recently published paper. This month, we check in with Jason Torres to discuss his paper, “A Multi-omic Integrative Scheme Characterizes Tissues of Action at Loci Associated with Type 2 Diabetes”.
AJHG: What caused you to start working on this project?
Jason: Funny enough, this project actually developed out of a necessity that came up while we were designing an experiment for a different project. We were planning a 3C-based experiment in a pancreatic beta cell line (to map regulatory elements at loci associated with diabetes) and we needed to prioritize ~100 variants to target out of a set of over 400 candidates. We wanted to focus on loci with the strongest islet signatures rather than those that may act through insulin-responsive peripheral tissues. Although there are methods that inform which tissues are most relevant to genetic susceptibility overall (i.e. through genome-wide enrichment or heritability partitioning), there wasn’t a suitable approach available that could help us discriminate tissues of action at specific loci.
AJHG: What about this paper most excites you?
Jason: I think the approach we devised for constructing tissue of action scores is really simple. The bigger challenge was trying to convince ourselves that it was actually useful. Had a clearcut set of “true positive” signals been available, I would have tried a supervised machine learning method (such as a convoluted neural network). However, our understanding of relevant tissues involved at specific loci associated with syndromic conditions is limited and often unclear. We therefore had to think more creatively to find ways to validate our tissue scores.
AJHG: Thinking about the bigger picture, what implications do you see from this work for the larger human genetics community?
Jason: There is a massive amount of interesting data being generated these days. On the genetics front, large-scale GWAS within national biobanks and trans-ethnic meta-analyses allow us to fine-map trait-associated loci to an unprecedented degree. Meanwhile, herculean efforts to map molecular epigenomic features (e.g. chromatin accessible regions, TF binding sites, enhancer elements) are uncovering tissue- and cell-specific regulatory programs that drive gene expression. But as these data mount, there is a growing need to devise principled and systematic approaches to integrate across diverse data types to resolve genetic mechanisms at specific loci. Our work offers a new piece to help solve this integrative “puzzle”.
AJHG: What advice do you have for trainees/young scientists?
Jason: Oftentimes, when we read a scientific paper, we get the sense that everything came together in a straightforward way. But that’s rarely the case! Experiments fail. Investigations may lead to dead ends, or veer off in completely different directions. Don’t get discouraged when this happens to you – it happens to all of us. Take a walk, get some rest, and then resume your work with a fresh mind.
AJHG: And for fun, tell us something about your life outside of the lab.
Jason: I’ve enjoyed playing guitar in punk bands since I was a teenager. During grad school, I played in two bands with fellow bioscience grad students (“The Amplicons” and “X-Rated Crystallography”), but I don’t jam much these days. However, since the start of lockdown, I’ve gotten really into tea and discovered that I am a “tea head” with a penchant for umami-rich matchas and aromatic oolongs.
Jason Torres, Ph.D., is a Senior Genetic Epidemiologist at the Nuffield Department of Population Health at the University of Oxford.