Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

Abstract

Stably Expressed Genes (SEGs) are a set of genes with invariant expression. Identification of SEGs, especially among both healthy and diseased tissues, is of clinical relevance to enable more accurate data integration, gene expression comparison and biomarker detection. However, it remains unclear how many global SEGs there are, whether there are development-, tissue- or cell-specific SEGs, and whether diseases can influence their expression. In this research, we systematically investigate human SEGs at single-cell level and observe their development-, tissue- and cell-specificity, and expression stability under various diseased states. A hierarchical strategy is proposed to identify a list of 408 spatial-temporal SEGs. Development-specific SEGs are also identified, with adult tissue-specific SEGs enriched with the function of immune processes and fetal tissue-specific SEGs enriched in RNA splicing activities. Cells of the same type within different tissues tend to show similar SEG composition profiles. Diseases or stresses do not show influence on the expression stableness of SEGs in various tissues. In addition to serving as markers and internal references for data normalization and integration, we examine another possible application of SEGs, i.e., being applied for cell decomposition. The deconvolution model could accurately predict the fractions of major immune cells in multiple independent testing datasets of peripheral blood samples. The study provides a reliable list of human SEGs at the single-cell level, facilitates the understanding on the property of SEGs, and extends their possible applications.

Chromatin enables precise and scalable gene regulation with factors of limited specificity

“…Because we evaluate expression levels at steady state, our results most directly apply to genes that must stably maintain a constant expression level over a long timespan relative to the rates of transcription and protein decay. The most salient biological analogs include "housekeeping" genes that carry out basal functions in all cells, as well as tissue-or cell type-specific genes that show little cell-to-cell or temporal variation in transcript levels (75)(76)(77). In addition to constant expression, three features make stably expressed genes (SEGs) particularly suitable for analysis with our paradigm.…”

Housekeeping protein-coding genes interrogated with tissue and individual variations

“…Indeed, this observation is particularly true for the numbers of diverse tissues investigated for a given study. This would be more sever with increasing single cell transcriptome analyses to cover even greater cell types 17,18 . It is anticipated that the amounts of housekeeping genes can vary from a few hundred to several thousand genes, depending on the numbers of tissue samples; specific criteria and experimental settings 14 .…”