Fig. 5
From: CoMI: consensus mutual information for tissue-specific gene signatures
Overview of consensus mutual information. A Flowchart describing the main procedure. B We firstly evaluated the standard deviation (\(\sigma_{i}\)) of gene expression intensity for each gene in a given omics data. C We then computing the average standard deviation (\(\overline{\sigma }\)) from \(\sigma_{i}\) of all genes, as well as expression means of normal and disease samples (i.e., Ī¼N and Ī¼D), respectively, in each gene. For a given gene i in sample j, we assigned expression intensity into 7 integral symbols by considering the \(\overline{\sigma }\) and its Ī¼N and Ī¼D. D The gene expression values were converted to discretized integer symbol ranging from 0 to 6. The highly expressed genes were assigned to the highest symbol 6 and lowly expressed genes were assigned to the lowest symbol 0. (D) The calculation of consensus mutual information (CoMI) values of all genes