Modern medicine has become more and more personalized and is rapidly replacing the traditional medicine approach (one-drug-fits-all patients). While in the past the majority of medical treatments was designed for the “average patient”, nowadays, thanks to the enormous amount of available data including genomic, proteomic and electronically-stored health information, is it possible to obtain a bigger picture for each patient and design patient-tailored treatments.
Such approach goes under the name of precision medicine and consists in the customization of medical decisions and treatments according to patients’ genetic, molecular and medical information. Treatments in particular are nowadays designed for small groups of patients (or even for one), rather than large groups, and their development usually requires the analysis of the aforementioned big data.
The Human Genome Project has enchanted us by a visionary promise of solving the mystery of the majority of human diseases but the amount of data of unknown significance still exceed our analytical skills. As a result, scientists increasingly look at new innovative technologies to revive such promise.
Private companies, universities, hospitals and other research consortiums have started using artificial intelligence (AI) with the aim of improving medical diagnosis and developing precise treatments for particular diseases. The role of AI is increasingly important in health care because it is often more powerful than human mind in performing some types of analytical tasks, such as the investigation of huge volumes of data, and on the long run can be tremendously helpful to derive new knowledge from such data.
Facial analysis of human phenotypes using modern technologies aroused the interest of many scientific groups lately. Advanced computer vision and deep learning algorithms have shown promising results that allow medical geneticists and clinicians to differentiate similarities of hundreds of genetic diseases based on simple face images. Several international research teams have proposed the use of AI as a system to detect genetic disorders starting from facial analysis. In particular, two recent studies showed how such a technology may improve personalized healthcare and become a tool for the clinical routine. The recently published articles by van der Donk et al. 2018 and Diets et al. 2019 [1,2] discussed how the deep learning technology may help in the facial analysis of phenotypes of rare genetic syndromes.
These results are extremely important and promising for personalized care but we should not forget that for the diagnosis of a genetic syndromes with the computer-aided facial recognition system should always be used in tandem with genomic variant interpretation. The interpretation of relevant genomic variants, usually obtained through next-generation sequencing (NGS), is indeed a crucial step in the diagnosis of a genetic condition.
This stage of analysis could be performed with the expert Variant interpreter – eVai – our AI based solution for an accurate genomic variant interpretation. eVai prioritizes all the genomic variants of an individual and suggests a list of possible related genetic diagnosis by combining AI with ACMG/AMP (American College of Medical Genetics and Genomics and the Association for Molecular Pathology) guidelines. eVai significantly improves the accuracy of genomic variant interpretation thanks to the incorporation of genomic big data (well integrated omics resources), AI and ACMG/AMP guidelines.
The fusion of advanced facial recognition systems and accurate genomic variant interpreters like eVai could represent a perfect solution to speed up the daily work of clinical geneticists and improve the diagnosis of genetic disorders.
The most promising researches in personalized medicine are characterized by a strong collaboration across different disciplines (medicine, biology, statistics, and computer engineering) and medical institutions. A smooth collaboration across various specialists together with appropriate interpretation of results and AI could therefore create impressive diagnostic tools for precision medicine and personalized healthcare. We are doing our best to support such paradigm shift.
 R. van der Donk et al. Next-generation phenotyping using computer vision algorithms in rare genomic neurodevelopmental disorders, Genet. Med. 0 (2018). doi:10.1038/s41436-018-0404-y.
 I.J. Diets et al. De Novo and Inherited Pathogenic Variants in KDM3B Cause Intellectual Disability, Short Stature, and Facial Dysmorphism, Am. J. Hum. Genet. 104 (2019) 758–766. doi:10.1016/j.ajhg.2019.02.023.