COTAN: Co-expression Table Analysis for scRNA-seq data

Abstract

Estimating co-expression of cell identity factors in single-cell transcriptomes is crucial to decode new mechanisms of cell state transition. Due to the intrinsic low efficiency of single-cell mRNA profiling, novel computational approaches are required to accurately infer gene co-expression in a cell population. We introduce COTAN, a statistical and computational method to analyze the co-expression of gene pairs at single cell level, providing the foundation for single-cell gene interactome analysis.

A eutherian-specific microRNA controls the translation of Satb2 in a model of cortical differentiation

“…To get insights into the pattern of CITF transcriptional activation in specific cell subsets, we analyzed CITF co-expression in single cells by co-expression table analysis (COTAN) ( Galfrè et al, 2020 ). COTAN can assess the co-expression of gene pairs in a cell and, by extending this analysis to all gene pairs in the whole transcriptome, can indicate the tendency of a gene to be constitutively expressed or expressed in a subset of differentiating/differentiated cells.…”

“…COTAN Gene Differentiation Index (GDI) discerns between constitutive and non-constitutive genes by globally integrating COEX values ( Galfrè et al, 2020 ) ( Figure 2 A). We used GDI analysis to infer the propensity of CITFs to be expressed in restricted cell subsets during corticogenesis.…”

“…We then compared miR-92a/b and miR-541 targets with those of three recently described miRNAs of corticogenesis, namely let7, miR-9, and miR-128 ( Shu et al., 2019 ). For this, we selected a subset of 395 genes associated with an embryonic cortical marker signature ( Galfrè et al, 2020 ). Among the six miRNAs analyzed, let-7 and miR-541 showed in silico affinity with more than half of the signature genes ( Figure 7 B and Data S6 ), suggesting a more relevant role for them in corticogenesis.…”

“…

Figure 7 In silico analysis of miRNA/mRNA interactions (A) In silico comparison of the affinity of mouse miRNAome (gray dots), miR-92a/b, and miR-541 (colored dots) with Satb2 3′UTR (Ensembl Mus musculus Satb2 -201 cDNA 3′UTR), in relation to the average miRNA expression levels during corticogenesis. (B) In silico affinity of cortical miRNAs to the 3′UTR of an embryonic cortical layer gene signature (395 genes) ( Galfrè et al, 2020 ). (C) GO enrichment of the mir-541 gene targets with high in silico affinity to Satb2 3′UTR (cumulative score higher than 400, n = 48) ( Enright et al., 2003 ) with respect to the layer gene signature employed in (B).

…”

“… (D) List of the eight genes common to all the GO terms shown in (C). (E) Plot showing E/I read counts developmental increase (x axis) with respect to miR-541/mRNA affinity score (y axis) to genes of the embryonic cortical marker signature ( Galfrè et al, 2020 ). Colored dots indicate genes listed in (D).…”