A NETWORK-BASED COMPUTATIONALPIPELINE TO STUDY THE VARIABILITY OFTRANSCRIPTOME PROFILES FOR HUMANDISEASES

Abstract

Machine learning applications to high-throughput data in medicine – one of the
biggest resource for understanding complex disease – so far have been limited.
Here we present a computational approach for assessing the intrinsic variability
in the most prominent data type, transcriptomics data for disease cohorts. Our
study looks at situations, where multiple datasets for the same disease are
available. We leverage concepts of network medicine to assess, how the match
between a biological network and a set of differentially expressed genes varies
across different networks and experiments. Our results show that different
biological networks yield markedly different results. Also, the clustering of
diseases depends strongly on the choice of parameters contained in the data
analysis and network processing.