Data integration by fuzzy similarity-based hierarchical clustering

摘要

Nowadays, high throughput methods, in biological and biomedical fields, acquire a large number of molecular parameters by a single experiment [ ]. In particular, such measured parameters are collected in “omics” datasets (e.g., genomics, transcriptomics, methylomics). Among multiple measured parameters, DNA genome sequence, RNA expression and DNA methylation are representative instances. For individually analysing such data, several methodologies have been introduced in literature, even though, recently, a number of studies pointed out the best performance coming from the integration of multi-omics data. For instance, analysing each omic (or in the machine learning jargon), set separately, fundamental patterns can be detected from data, however some fine-tuned structures, such as cancer sub-types, can be highlighted by both gene expression and DNA methylation information, so that multi-omics analysis can reduce the effects of experimental and biological noise in data [ ]. From literature, three kinds of integration methodologies emerge: