Bayesian correlation is a robust gene similarity measure for single-cell RNA-seq data

Sánchez-Taltavull, Daniel; Perkins, Theodore J.; Dommann, Noëlle; Melin, Nicolas; Keogh, Adrian; Candinas, Daniel; Stroka, Deborah; Beldi, Guido

doi:10.1093/nargab/lqaa002

Cited by 21 publications

(19 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, TIL isolated directly from patient tumors can be segregated into both CISH high PD1 low 41BB low TIM3 low and CISH low PD1 high 41BB high TIM3 high populations. In subsequent analysis using a cluster-based method 17 , consistent with previous reports 13,18,19 we observed a signi cant positive correlation between PD1 (PDCD1) with activation/exhaustion markers TOX 20 and CD39 (ENTPD1) and a negative correlation with the memory marker TCF7 21 (Fig. 1c).…”

Section: Cish Expression and Inducibility In Effector T Cellssupporting

confidence: 90%

Internal checkpoint regulates T cell neoantigen reactivity and susceptibility to PD1 blockade

Palmer

Webber

Patel

et al. 2020

Preprint

View full text Add to dashboard Cite

While neoantigen-specific tumor infiltrating lymphocytes (TIL) can be derived from in antigen-expressing tumors, their adoptive transfer fails to consistently elicit durable tumor regression. There has been much focus on the role of activation/exhaustion markers such as PD1, CD39 and TOX in TIL senescence. We found these markers were inversely expressed to Cytokine-Induced SH2 protein (CISH), a negative regulator of TCR signaling and tumor immunity in mice. To evaluate the physiological role of CISH in human TIL we developed a high-efficiency CRIPSR-based method to knock out CISH in fully mature TIL. CISH KO resulted in increased T cell receptor (TCR) avidity, tumor cytolysis and neoantigen recognition. CISH expression in the tumor resections correlated with TIL inactivity against p53 hotspot mutations and CISH KO in TIL unmasked reactivity against these universal neoantigens. While CISH KO resulted in T cell hyperactivation and expansion it did not alter maturation, perhaps by preferential PLCγ-1 and not AKT inhibition. Lastly, CISH KO in T cells increased PD1 expression and the adoptive transfer of Cish KO T cells synergistically combines with PD1 antibody blockade resulting in durable tumor regression and survival in a preclinical animal model. These data offer new insights into the regulation of neoantigen recognition, expression of activation/exhaustion markers, and functional/maturation signals in tumor-specific T cells.

show abstract

Section: Cish Expression and Inducibility In Effector T Cellssupporting

confidence: 90%

Internal checkpoint regulates T cell neoantigen reactivity and susceptibility to PD1 blockade

Palmer

Webber

Patel

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…To describe the hepatic expression of TJ genes, we used our recently published single-cell RNA sequencing (scRNA-seq) data set of parenchymal and nonparenchymal cells from a C57BL/6 liver. 26 Unsupervised clustering identified 14 unique cell clusters ( Figure 1 A ). A defined set of marker genes and clustering for cell classification identified the populations of hepatocytes, cholangiocytes, endothelial cells, immune cells, and stellate cells ( Figure 1 B ).…”

Section: Resultsmentioning

confidence: 99%

“…The unique molecular identifiers (UMI) matrix of our recently published scRNA-seq was downloaded (GEO accession number: GSE134134 ). 26 We removed cells with more than 15% UMIs coming from mitochondrial genes and cells with more than 25% UMIs coming from globin genes. In addition, a cell containing an abnormally high number of UMIs (110,270) was excluded.…”

Section: Methodsmentioning

confidence: 99%

“…After data preprocessing, the UMI matrix was processed as previously described. 26 Shortly, we transformed the UMI matrix into a Seurat object (Paul Hoffman, Satija Lab, New York Genome Center, New York) with Seurat 2 (PMID: 31178118 ). The data of the Seurat object were log-normalized, the variable genes were identified, and the data were scaled.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Loss of Claudin-3 Impairs Hepatic Metabolism, Biliary Barrier Function, and Cell Proliferation in the Murine Liver

Baier

Sánchez-Taltavull

Yarahmadov

et al. 2021

Cellular and Molecular Gastroenterology and Hepatology

Self Cite

View full text Add to dashboard Cite

Delineating the cell type-specific expression of hepatic tight junction genes showed that claudin-3 is the predominant tight junction protein on hepatocytes and cholangiocytes. In vivo study of claudin-3 knockout mice showed that claudin-3 is necessary to maintain lipid metabolism, biliarybarrier function, and optimal liver regeneration.BACKGROUND & AIMS: Tight junctions in the liver are essential to maintain the blood-biliary barrier, however, the functional contribution of individual tight junction proteins to barrier and metabolic homeostasis remains largely unexplored. Here, we describe the cell type-specific expression of tight junction genes in the murine liver, and explore the regulation and functional importance of the transmembrane protein claudin-3 in liver metabolism, barrier function, and cell proliferation. METHODS:The cell type-specific expression of hepatic tight junction genes is described using our mouse liver single-cell sequencing data set. Differential gene expression in Cldn3 -/and Cldn3 þ/þ livers was assessed in young and aged mice by RNA sequencing (RNA-seq), and hepatic tissue was analyzed for lipid content and bile acid composition. A surgical model of partial hepatectomy was used to induce liver cell proliferation. RESULTS: Claudin-3 is a highly expressed tight junction protein found in the liver and is expressed predominantly in hepatocytes and cholangiocytes. The histology of Cldn3 -/livers showed no overt phenotype, and the canalicular tight junctions appeared intact. Nevertheless, by RNA-seq we detected a down-regulation of metabolic pathways in the livers of Cldn3 -/young and aged mice, as well as a decrease in lipid content and a weakened biliary barrier for primary bile acids, such as taurocholic acid, taurochenodeoxycholic acid, and taurine-conjugated muricholic acid. Coinciding with defects in the biliary barrier and lower lipid metabolism, there was a diminished hepatocyte proliferative response in Cldn3 -/mice after partial hepatectomy. CONCLUSIONS:Our data show that, in the liver, claudin-3 is necessary to maintain metabolic homeostasis, retention of bile acids, and optimal hepatocyte proliferation during liver regeneration. The RNA-seq data set can be accessed Q8 at: https:// www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc¼GSE159914 (token: wrmhoaccjrgrjyz).

show abstract

“…For example, Skinnider et al [35] recommends proportionality-based metrics, whilst Kim et al [36] recommends correlation-based metrics, specifically Pearson. However, Sanchez-Taltavull et al [37] recommend Bayesian correlation over Pearson. Several other studies have proposed novel and scRNA-seq specific proximity metrics after observing variable performance of traditional proximity metrics [38][39][40].…”

Section: Introductionmentioning

confidence: 99%

How does data structure impact cell-cell similarity? Evaluating the influence of structural properties on proximity metric performance in single cell RNA-seq data

Watson

Mora

Fard

et al. 2022

Preprint

View full text Add to dashboard Cite

Accurately identifying cell populations is paramount to the quality of downstream analyses and overall interpretations of single-cell RNA-seq (scRNA-seq) datasets but remains a challenge. The quality of single-cell clustering depends on the proximity metric used to generate cell-to-cell distances. Accordingly, proximity metrics have been benchmarked for scRNA-seq clustering, typically with results averaged across datasets to identify a highest performing metric. However, the ‘best-performing’ metric varies between studies, with the performance differing significantly between datasets. This suggests that the unique structural properties of a scRNA-seq dataset, specific to the biological system under study, has a substantial impact on proximity metric performance. Previous benchmarking studies have omitted to factor the structural properties into their evaluations. To address this gap, we developed a framework for the in-depth evaluation of the performance of 17 proximity metrics with respect to core structural properties of scRNA-seq data, including sparsity, dimensionality, cell population distribution and rarity. We find that clustering performance can be improved substantially by the selection of an appropriate proximity metric and neighbourhood size for the structural properties of a dataset, in addition to performing suitable pre-processing and dimensionality reduction. Furthermore, popular metrics such as Euclidean and Manhattan distance performed poorly in comparison to several lessor applied metrics, suggesting the default metric for many scRNA-seq methods should be re-evaluated. Our findings highlight the critical nature of tailoring scRNA-seq analyses pipelines to the system under study and provide practical guidance for researchers looking to optimise cell similarity search for the structural properties of their own data.

show abstract

Bayesian correlation is a robust gene similarity measure for single-cell RNA-seq data

Cited by 21 publications

References 79 publications

Internal checkpoint regulates T cell neoantigen reactivity and susceptibility to PD1 blockade

Internal checkpoint regulates T cell neoantigen reactivity and susceptibility to PD1 blockade

Loss of Claudin-3 Impairs Hepatic Metabolism, Biliary Barrier Function, and Cell Proliferation in the Murine Liver

How does data structure impact cell-cell similarity? Evaluating the influence of structural properties on proximity metric performance in single cell RNA-seq data

Contact Info

Product

Resources

About