Fully automatic landmarking of 2D photographs identifies novel genetic loci influencing facial features

Qing Li,Jieyi Chen,Pierre Faux,Nicolas Navarro,Sijia Wang,Kaustubh Adhikari,Andres Ruiz-Linares
DOI: https://doi.org/10.21203/rs.3.rs-1479643/v1
2023-01-01
European Journal of Human Genetics
Abstract:Abstract We report a genome-wide association study for facial features in > 6,000 Latin Americans. We placed 106 landmarks on 2D frontal photographs using the cloud service platform Face++. After Procrustes superposition, genome-wide association testing was performed for 301 inter-landmark distances. We detected nominally significant association (P-value < 5×10− 8) for 42 genome regions. Of these, 9 regions have been previously reported in GWAS of facial features. In follow-up analyses, we replicated 26 of the 33 novel regions (in East Asians or Europeans). The replicated regions include 1q32.3, 3q21.1, 8p11.21, 10p11.1, and 22q12.1, all comprising strong candidate genes involved in craniofacial development. Furthermore, the 1q32.3 region shows evidence of introgression from archaic humans. These results provide novel biological insights into facial variation and establish that automatic landmarking of standard 2D photographs is a simple and informative approach for the genetic analysis of facial variation, suitable for the rapid analysis of large population samples.
What problem does this paper attempt to address?