Comparison of pretreatment techniques for single-cell RNA sequencing of airway aspirate specimens / 中华微生物学和免疫学杂志

ABSTRACT

Objective:To investigate the appropriate pretreatment methods for single cell RNA sequencing of airway aspirate cells.Methods:Four fresh airway aspirate specimens were collected from four patients with acute respiratory tract infections. These specimens were digested with airway aspirate digester and prepared into single cell suspension. The cells were used for library construction directly (DE), or fixed with 10×Genomics Chromium Next GEM Single Cell Fixed RNA Sample Preparation Kit and then mixed to construct the library (DF), or cryopreserved, thawed, fixed (FF) before mixed to construct the library. All three methods were treated with oil emulsion using 10 4 cells and subjected to single-cell sequencing using the 10×Genomics platform. The number of obtained cells, data quality, annotated cell types and expression of marker genes were analyzed. Differences in the expression of highly variable genes (HVGs) of the same cell subsets obtained by the three pretreatment methods were compared using Pearson correlation. Expression of the differentially expressed genes in the same cell subpopulation obtained by different pretreatment methods was also compared. The correlation of the expression of differentially expressed genes between the same cell subsets obtained by the three pretreatment methods was analyzed by Pearson correlation. Results:The median numbers of single cells obtained using DE, FF and DF methods were 2 733, 1 140 and 5 897 ( P>0.05). The unique molecular identifiers were higher than 500. The median numbers of genes obtained using the three methods were 801, 887 and 1 259 ( P>0.05). The cells with novelty score over 0.8 accounted for 99%, 87% and 93%, respectively. There were nine cell subsets obtained by the three methods, including squamous cells, secretory cells, ciliated cells, T cells, B cells, macrophages, plasma cells and neutrophils. DF and FF methods could obtain more basal cells with specific high expression of keratin 5 than DE method. The differentially expressed and highly variable genes in the same cell subsets obtained by the three pretreatment methods showed high consistency in their expression with a significant correlation ( P<0.001). Conclusions:Under the same sequencing data volume, the quality of data obtained from fixed airway aspirate single-cell suspensions using the method of probe hybridization and transcriptome sequencing was comparable to that obtained directly from fresh cells. This method was more suitable for the pretreatment of clinical samples used for single-cell RNA sequencing.