Seq-Well: portable, low-cost RNA sequencing of single cells at high throughput

Abstract

Single-cell RNA-seq can precisely resolve cellular states, but applying this method to low-input samples is challenging. Here, we present Seq-Well, a portable, low-cost platform for massively parallel single-cell RNA-seq. Barcoded mRNA capture beads and single cells are sealed in an array of subnanoliter wells using a semipermeable membrane, enabling efficient cell lysis and transcript capture. We use Seq-Well to profile thousands of primary human macrophages exposed to Mycobacterium tuberculosis.

Figure 1. Seq-Well: A Portable, Low-Cost Platform for High-Throughput Single-Cell RNA-Seq of Low-Input Samples

(a) Photograph of equipment and array used to capture and lyse cells, respectively. (b) Transcripts captured from a mix of human (HEK293) and mouse (NIH/3T3) cells reveal distinct transcript mapping and single-cell resolution. Human (mouse) cells (> 2,000 human (mouse) transcripts and < 1,000 mouse (human) transcripts) are shown in blue (red). Among the 254 cells identified, 1.6% (shown in purple) had a mixed phenotype. (c,d) Violin plots of the number of transcripts (c) and genes (d) detected in human or mouse single-cell libraries generated by Seq-Well or Drop-Seq (Ref. ; Center-line: Median; Limits: 1^st and 3^rd Quartile; Whiskers: +/− 1.5 IQR; Points: Values > 1.5 IQR). Using Seq-Well (Drop-Seq), an average of 37,878 (48,543) transcripts or 6,927 (7,175) genes were detected among human HEK cells (n = 159 for Seq-Well; n = 48 for Drop-Seq) and an average of 33,586 (26,700) transcripts or 6,113 (5,753) genes were detected among mouse 3T3 cells (n = 172 for Seq-Well; n = 27 for Drop-Seq) at an average read depth of 164,238 (797,915) reads per human HEK cell and an average read depth of 152,488 (345,117) read per mouse 3T3 cell.

Figure 2. Combined Image Cytometry and scRNA-Seq of Human PBMCs

(a) The hierarchical gating scheme (with the frequencies of major cell subpopulations) used to analyze PBMCs that had been labeled with a panel of fluorescent antibodies, loaded onto three replicate arrays and imaged prior to bead loading and transcript capture (Methods). Myeloid cells (green) were identified as the population of hCD3(-) HLA-DR(+) CD19(-) cells; B cells (orange) as the subset of hCD3(-) HLA-DR(+) CD19 (+) cells; CD4 T cells (blue) as the subset of CD3(+) CD4(+) cells; CD8 T cells (yellow) as the CD3(+) CD8(+) subset of cells; and, NK cells (red) as the subset of CD3(-) HLA-DR (-) CD56 (+) CD16(+) cells. (b) t-SNE visualization of single-cell clusters identified among 3,694 human Seq-Well PBMCs single-cell transcriptomes recovered from the imaged array and the two additional ones (Methods; Supplementary Fig. 10–12). Clusters (subpopulations) are labeled based on annotated marker gene (Supplementary Fig. 10). (c) The distribution of transcriptomes captured on each of the 3 biological replicate arrays, run on separate fractions of the same set of PBMCs. All shifts are insignificant save for a slightly elevated fraction of CD8 T cells in array 1 (*, p=1.0×10⁻¹¹; Chi-square Test, Bonferroni-corrected). (d) A heatmap showing the relative expression level of a set of inflammatory and antiviral genes among cells identified as monocytes.

Figure 3. Sequencing of TB-Exposed Macrophages in a BSL3 Facility Using Seq-Well

(a) t-SNE visualization of single-cell clusters identified among 2,560 macrophages (1,686 exposed, solid circles; 874 unexposed, open circles) generated using 5 principal components across 377 variable genes (Methods). (b) Marker genes for the 3 phenotypic clusters of macrophages highlighted in (a). (c) Volcano plots of differential expression between exposed and unexposed macrophages within each cluster showing genes enriched in cells exposed to M. tuberculosis. In each plot, a p-value threshold of 5.0 ×10⁻¹⁶ based on a likelihood ratio test was used to establish statistical significance, while a log2-fold change threshold of 0.4 was used to determine differential expression. Genes with p-values less than 5.0×10⁻⁶ are shown in cyan and absolute log2-fold changes greater than 0.4; In magenta are genes with p-values less than 5.0×10⁻⁶ but absolute log2-fold changes less than 0.4; and, in black, are genes with p-values greater than 5.0×10⁻⁶ and absolute log2-fold changes less than 0.4.