NEW YORK — Researchers have uncovered more than 200 genetic loci that influence face shape using a new approach to characterize facial morphology in siblings.
Various approaches for characterizing human face shape exist, but while simple linear measurements don't capture its full complexity, more complex approaches cannot easily be analyzed statistically.
With a new approach drawing on facial similarities shared by siblings, a team led by researchers from the Katholieke Universiteit Leuven in Belgium has defined face shape traits and linked them to genetic loci affecting how people look. This approach enabled the researchers to draw upon the advantages of simple facial measurements while also capturing the complexity of the human face. "Moreover, as sibling share on average 50 percent of the same DNA, the facial traits identified as such are likely to have a genetic basis," first author Hanne Hoskens from KU said in an email.
As they reported in PLOS Genetics on Thursday, Hoskens and her colleagues linked the facial features they identified to 218 genetic loci, many of which have previously been tied to embryonic facial development or the abnormal development of facial features.
For their analysis, the researchers first developed a set of facial phenotypes shared by siblings. Using a set of 273 sibling pairs — including 78 brother-brother, 79 sister-sister, and 116 brother-sister pairs — from 194 families of European ancestry, the researchers teased out a set of 1,048 shared facial traits.
Using these traits, the researchers then conducted a genome-wide association study of 8,246 individuals of European ancestry from the US and the UK. They uncovered 218 genetic loci that reached genome-wide significance with regard to these traits.
About half these loci either overlapped with or were located near other loci previously associated with facial phenotypes in other studies. These included signals near TBX15, PAX3, and RUNX2. Others, like signals near FOXE1 and EPHA3, had not previously been tied to normal-range facial phenotypes.
The researchers additionally searched for low-frequency variants associated with facial features using exome sequencing data, uncovering 53 genes with exome-wide significant results. Of these, five were linked to genes previously found to be involved in craniofacial phenotypes.
Through a gene ontology analysis, the researchers found the 218 genetic loci they uncovered to be enriched for involvement in craniofacial development and morphogenesis. A further ChIP-seq analysis found that the active regulatory mark H3K27ac signals near the lead SNPs were enriched in cranial neural crest cells and embryonic craniofacial tissues, underscoring their likely roles during early embryonic development.
According to Hoskens, she and her colleagues are now extending their analysis to a cohort of Tanzanian children. "We do intend to investigate the overlap with the current findings to try and better understand transethnic facial genetics," she said. "The current knowledge in this domain is largely incomplete and mainly hampered by the availability of datasets, which [are] now biased to those of European descent."
Eventually, having a better understanding of the genetic basis of facial features could have applications in the clinic as well as in anthropology and forensics to predict what people might have looked like, Hoskens said.