Machine learning for multimodality genomic signal processing

Multiple modalities can be generated in various ways. One possiblity is via sensor diversity, and the other is feature diversity. In terms of sensor diversity, both the motif and gene expression modalities can be considered. Motifs are short sequences of DNA responsible for regulating gene networks and the expression of genes, whereas gene expression is the processing of producing proteins from information coded in genes. To further facilitate multi-modality fusion, a diversity of features may be extracted from each sensor by computational means. This is called feature diversity.