Researchers have identified 24 single-nucleotide polymorphisms in 20 genes that are significantly associated with facial features and used this data to develop a program that can predict 3D facial structure on the basis of these markers. High-resolution facial images of ~600 participants from 3 West African–European admixed populations were used to form detailed 3D images. Claes et al. then jointly modelled face shape, sex and genomic ancestry with genetic markers in candidate craniofacial genes to analyse the independent effects of genetic variants on facial features. Although more research is needed before this method might be applicable — for example, in forensics — this proof-of-principle study indicates the plausibility of such an approach.