Data-driven search for 3D facial traits determined by major gene effects in health and disease

In this project, we challenge the current paradigm shift in the post-genomics era that led to the transition from investigating monogenetic to polygenetic traits, and thus from rare major gene effects to common variants with weak effect. We hypothesise that many major gene effects among patients with complex phenotypes remain undiagnosed and that major gene effects in the so-called healthy population result in subclinical phenotypes that escape the attention of the clinician. Therefore, we embark on the search for a phenotypic signature for major gene effects through the complex phenotype of the human face. In addition, we explore the intriguing but understudied interplay between clinically abnormal and normal facial variation. To this end, we aim to critically improve data-driven phenotyping from large-scale image data through an unconventional integration of three different disciplines (clinical genetics, orthognatic surgery and image-based morphometrics and deep learning), each with a different conceptualisation of the human face.

Keywords:Facial Genetics, 3D Shape Analysis, Geometric Deep Learning, Subclinical Phenotyping, Dysmorphology, Oral and maxillofacial surgery, Imaging Genetics

Disciplines:Computer vision, Clinical genetics and molecular diagnostics, Quantitative genetics, Oral and maxillofacial surgery, Data mining, Morphological sciences not elsewhere classified