The psychological and economic burden of intellectual disability (ID) on families is huge, with 1.5-2% of all children having an IQ < 70 and 0.3-0.5% of children having moderate-severe ID with an IQ <50. A craniofacial anomaly is described in 30-50% of the known genetic causes of ID.

Advanced genomic sequencing can now identify variations in the DNA code of all ≈1500 known developmental disorder genes in one test. In 30-50% of cases, a diagnosis will be made by showing that a DNA variant has affected the function of a known gene.

Unfortunately there are a large number of genes of unknown function. When variants are detected in these genes or in non-coding DNA, it can be difficult to determine if they are normal variations or the cause of the syndrome. Therefore, the question remains how best to proceed when a child with syndromic intellectual disability remains undiagnosed despite testing all the known developmental disorder genes.

Comparing the DNA variants of individuals with similar features can lead to the discovery of new ID genes. However, locating at minimum a second individual or a cluster of individuals with similar facial features currently relies on doctors comparing photographs of undiagnosed children at medical meetings or recording written descriptions into international phenotyping and genetic variant databases.

One novel approach to this issue is the use of computer face-matching technology. In a paper published in BMC Biotechnology, Dudding et al. report an automated approach to matching the faces of non-identical individuals with one of ten genetic syndrome subgroups within a database of 3,145 images.

Computer face-matching

The core computer face-matching technology (FMT) used in this study was initially developed to match facial images of individuals for the primary purpose of recognizing blurry faces in CCTV for policing and counterterrorism. The software is being used to detect persons of interest in large crowd gatherings through various police and other security agencies.

Since the software was developed for CCTV, high quality professional photographs are not required and even poor quality historical photographs can readily be used.

Since the software was developed for CCTV, high quality professional photographs are not required and even poor quality historical photographs can readily be used. This technology is unique as it uses low-resolution structural and frequency domain features rather than high resolution features. It is based on spatial textures and statistical models and is simultaneously insensitive to pose, illumination, expression, obscuration, blurring, decoding artifacts, and low-resolution images.

Using the leave- one-out method (removing an individual image from the database and letting the software list the top 10 closest matches when the removed image is used as the test case), the computer face matching technology correctly identifies a top match, at least one correct match in the top five and at least one in the top 10 more than expected by chance. The accuracy of the software scored favorably compared with three senior clinical geneticists. There was low agreement between the technology and clinicians, with higher accuracy of the technology when results were discordant for all syndromes except Kabuki syndrome.

Although the accuracy of the computer face-matching technology was tested on images of individuals with known syndromic forms of intellectual disability, the results of this study illustrate the potential utility of face-matching technology within deep phenotyping platforms to facilitate the interpretation of DNA sequencing data for individuals who remain undiagnosed despite testing the known developmental disorder genes.

Pharmacological therapy is only available for the minority of individuals with intellectual disability due to rare metabolic diseases; however, a definitive molecular diagnosis can inform prognosis, guide management, reduce the need for extensive investigations and restore reproductive confidence for parents planning further children. Understanding the genetic basis of ID is the first step towards understanding interacting biological pathways and possible targeted genetic therapy.