Integration of Multiple, Diverse Methods to Identify Biologically Significant Marker Genes

Identification of genes that reliably mark distinct cell types is key to leveraging single-cell RNA sequencing to better understand organismal biology. Such genes are usually chosen by measurement of differential expression between groups of cells and selecting those with the greatest magnitude or most statistically significant change. Many methods have been developed for performing such analyses, but no single, best method has emerged. Validating the results of these analyses is costly in terms of time, effort and resources. We demonstrate that applying an ensemble of such methods robustly identifies genes that mark cells that cluster together and that show restricted expression assessed by antisense mRNA in situ and immunofluorescence. This technique is easily extensible to any number of differential expres-sion methods and the inclusion of additional methods is expected to result in further improvement in performance. (c) 2022 Elsevier Ltd. All rights reserved.

Automated summary

Identifying genes that reliably mark distinct cell types is crucial for using single-cell RNA sequencing to enhance our understanding of organismal biology. This study demonstrates that applying an ensemble of differential expression methods robustly identifies genes that mark clustered cells with restricted expression, as validated by antisense mRNA in situ and immunofluorescence techniques. This technique can easily be extended to incorporate additional methods, leading to further improvements in performance.