Many genetic defects are medically a barrel without a bottom. Often the Doctors are as clueless as the patients, many of the Sick have Odysseys in the health system, and not a few of the patients in whom a genetic disorder is present, the for years, treated incorrectly. What kind of disease or medical blurred – is “syndrome” it often remains in the dark.

Joachim Müller-Jung

editor in the features section, responsible for the area of “nature and science”.

F. A. Z. Twitter

The complexity of the underlying molecular and diagnostic characteristics is a typical case for Big Data and machine, the knowledge from millions of individual cases, to learn, and in a fraction of a second, the individual data of a patient to be classified – in theory, but so far only. Because the available data are still scarce, genome analysis, anything other than Routine.

in addition: Who would like to confide in a vending machine? Accurate in the long term this will not be possible to avoid but probably can’t, because, as experience shows, grow digital expertise of the artificial medical intelligence from month to month. When it comes to hereditary diseases, can be well-trained machines with a view more than any single doctor. In “Nature Medicine” reported that a us-German Team of Bioinformaticians and Genmedizinern of clinical Tests with a KI-System, the Boston company FDNA, the aim is to allow clinicians to distinguish hundreds of different genetic syndromes, and only through a camera view of the face of the patient.

Smart wired

When the man with the machine intelligences in the laboratory


In the case of the artificial neural network is an algorithm called “deep shape”, which was derived from the classical facial recognition programs. Through the evaluation of more than a hundred areas on the face of each of the more than 17 of 100 patients were used for the Training of the learning software, wants to have recognized, the AI algorithm of the typical anatomical and morphometric patterns for 216 different Gensyndrome.

Behind it is a relatively old observation by geneticists: a genetic cause – the genotype – is not often caused only malfunctions in the organs, but manifests itself in many sufferers, even the outer appearance – the phenotype. In the case of the delimitation of the so-called Angelman syndrome from other genetic disorders on the portrait photos of the machine in the Test was much more successful than clinical Doctors. In more than ninety percent of the cases, the AI led the Doctors on the right track to what syndrome it was. This means, however, by no means, the machine according to the Frontal shot looking into the face of a number of healthy people in a “hidden” genetic defect (and the perfect diagnosis).