Loading
Genes that rise and fall together across samples form co-expression modules; this correlation structure is what expression models mostly learn.
A gene co-expression network is a graph in which genes are nodes and an edge joins two genes whose expression is significantly correlated across samples (Wikipedia). Modules in such networks correspond to cell types, proliferation, immune infiltration and tissue of origin, so a masked-gene model that reconstructs hidden genes from visible ones is learning this structure. It also explains why a few hundred well chosen genes carry most of the information in twenty thousand.
Shares Masked autoencoders and masked gene modelling, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Tissue-of-origin signal in tumour data, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Tissue-of-origin signal in tumour data, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Masked autoencoders and masked gene modelling, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Masked autoencoders and masked gene modelling, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Tissue-of-origin signal in tumour data, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Gene set enrichment analysis (GSEA and ssGSEA), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Tissue-of-origin signal in tumour data, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.