Sub-Cluster Identification through Semi-Supervised Optimization of Rare-Cell Silhouettes (SCISSORS) in single-cell RNA-sequencing

Abstract

Motivation Single-cell RNA-sequencing (scRNA-seq) has enabled the molecular profiling of thousands to millions of cells simultaneously in biologically heterogenous samples. Currently, common practice in scRNA-seq is to determine cell type labels through unsupervised clustering and the examination of cluster-specific genes. However, even small differences in analysis and parameter choosing can greatly alter clustering results and thus impose great influence on which cell types are identified. Existing methods largely focus on determining the optimal number of robust clusters, which can be problematic for identifying cells of extremely low abundance due to their subtle contributions towards overall patterns of gene expression. Results Here we present a carefully designed framework, SCISSORS, which accurately profiles subclusters within broad cluster(s) for the identification of rare cell types in scRNA-seq data. SCISSORS employs silhouette scoring for the estimation of heterogeneity of clusters and reveals rare cells in heterogenous clusters by a multi-step semi-supervised reclustering process. Additionally, SCISSORS provides a method for the identification of marker genes of high specificity to the cell type. SCISSORS is wrapped around the popular Seurat R package and can be easily integrated into existing Seurat pipelines. Availability SCISSORS, including source code and vignettes for example datasets, is freely available at https://github.com/jr-leary7/SCISSORS. Supplementary information Supplementary data are available at Bioinformatics online.

Myofibroblastic cancer‐associated fibroblast subtype heterogeneity in pancreatic cancer

“…To develop effective CAF-based therapies to cooperatively target the stroma, accurate preclinical models are needed. While CAF subtypes have been a promising development, 5,15 our results show that CAFs are very heterogeneous even within a specific subtype such as myCAFs.…”

“…Libraries were prepared using the TruSeq Briefly, BCL image files were converted to fastq files using "cellranger mkfastq," followed by unique molecular identifier quantification using "cellranger count" based on the human GRCh38 genome. 14 SCIS-SORS (Sub-Cluster Identification through Semi-Supervised Optimization of Rare-cell Silhouettes), 15 which is a wrapper of the Seurat package (v4), was employed for downstream analysis on the count matrix (filtered_feature_bc_matrix). The quality control steps include (1) the inclusion of genes expressed in more than two cells, (2) the inclusion of cells that have the number of genes captured within two times absolute deviation (+−2 SD) from the median (1500 < nFeatures < 4700), and (3) the inclusion of cells that have less than three times SD (+3 SD) from the median of the percent of mitochondrial reads (percent_MT < 10).…”

“…The filtered data were processed by SCISSORS to derive robust clusters as described before. 15 Resultant clusters were annotated using SingleR. 16…”

“…SCISSORS was used to derive myCAF-and iCAF-like exemplar genes from the Elyada data set 5 as previously described. 15 panCAF exemplar genes were defined by combining myCAF-, iCAF-, and antigen-presenting CAF-like clusters and compared to perivascular fibroblast and endothelial cell clusters to derive a candidate gene list (p < 0.05, Wilcoxon's rank-sum test). This gene list was filtered by excluding the genes that were highly expressed (top 10%) in other cell clusters.…”

“…To determine gene expression using more discriminatory genes, we next compared CAF subtype exemplar genes that we identified through SCISSORS. 15 We previously derived discriminatory exemplar genes for CAFs, based on the previously described iCAF and myCAF subtypes (SCISSORS iCAF-like and myCAF-like). Both CAF lines were consistent with the myCAF-like signature (Figure 4B)…”