Deciphering Human Infection Resistance with Machine Learning: A Way to Novel Antibiotics?

Biomedical data are difficult to analyze by current Machine Learning approaches due to their small sample size (< 500 patients) and many features (> 1 Million). From there, the multiple comparison problem in statistics occurs: If many data series are compared, similarly convincing but coincidental data may be obtained. In contrast to standard biomarker discovery technologies that test correlations of single features,’s machine learning platform is designed to find complex patterns of interactions in high-dimensional biomedical data. This results in next-generation biomarkers with outstanding accuracy and sensitivity prediction properties. They are highly needed for drug development and healthcare decision-making.

Recently, we analyzed data from genome-wide association studies (GWAS) to investigate gene-gene-interactions for a better understanding of infection resistance. Here, we unravel human infection resistance for a better understanding of disease mechanisms in order to find novel treatments.