In this work, we developed new approaches to significantly improve identification of molecular malfunction patterns leading to disease. We introduced a new feature, characterizing protein sequence positions by the distribution of variant effects on protein function. For application in improved effect prediction, we implemented a Machine Learning approach predicting this feature with high accuracy. We further explored the microbiome perspective and designed tools that enable the analyses of microbiome function profiles for disease states as well as microbial functional similarity. To account for Big Data bottlenecks, we published a load balancing software pooling compute resources which speeded up analyses drastically.
«
In this work, we developed new approaches to significantly improve identification of molecular malfunction patterns leading to disease. We introduced a new feature, characterizing protein sequence positions by the distribution of variant effects on protein function. For application in improved effect prediction, we implemented a Machine Learning approach predicting this feature with high accuracy. We further explored the microbiome perspective and designed tools that enable the analyses of micro...
»