Fig. 1

Single-cell transcriptomic profiles in plants. a The technical noise profile between two single cells of the same cell type, showing high dispersion for transcripts expressed at a low level. The axes are read-counts representing gene expression levels on a log2 scale. As most genes are expected to be expressed at similar levels, the two axes evaluate replication and show that, at these scales, genes expressed at higher levels show the potential to distinguish biological from technical noise. b (upper) The expression distribution of a gene among pooled samples typically shows a peak frequency on a positive expression value. (lower) Gene expression among single-cell samples typically shows a peak frequency at zero, with a subset of cells showing a second peak of positive read counts in a subset of samples. Density represents the frequency of cells showing a given expression level (read count). c Several gold-standard markers in single-cell profiles of cells with known tissue origins. These functional markers are expressed at higher levels (e.g., more replicable expression in a and non-zero expression in b (lower). In these real samples collected from plant cells, markers for the quiescent center (QC), stele, and epidermis all show detectable expression in target cells and are largely absent in non-target cells, with some false-positive and false-negative expression