Fig. 1

Experimental pipeline of SUPeR-seq, and its sensitivity at the whole-transcriptome scale. a The schematic of SUPeR-seq analysis. A single cell is lysed to release RNAs. RNAs are then reverse transcribed into first-strand cDNAs using random primers with a fixed anchor sequence (AnchorX-T₁₅N₆). Unreacted primers are then digested using ExoSAP-IT, followed by adding poly(A) tails to the 3′ ends of the first-strand cDNAs using dATP doped with 1 % ddATP to restrict the length of poly(A) tails. Second-strand cDNAs are synthesized using poly(T) primers with a different anchor sequence (AnchorY-T₂₄). Then the double-stranded cDNAs are evenly amplified by PCR using AnchorX-T₁₅ and AnchorY-T₂₄ primers. Finally the purified single cell cDNAs are used to prepare sequencing libraries following Illumina’s TruSeq DNA sample preparation protocols. b Detection sensitivity of SUPeR-seq on poly(A)- genes in individual cells. We identified 696 poly(A)- genes by bulk RNA sequencing, of which around 30 % could be recovered in a single cell by SUPeR-seq (for details, see Additional file 3). When merged the SUPeR-seq data of the seven single cell samples together, over 50 % of these 696 genes could be successfully recovered by SUPeR-seq in at least one cell. c The number of genes detected from individual and bulk HEK293T cells using different protocols. SUPeR-seq detected 10,911 genes on average from an individual cell with FPKM ≥1(Fragments Per Kilobase of exon model per Million mapped reads), 19.3 % more than the Tang2009 protocol did (9148 genes on average). For comparison, 14,931 genes were detected with FPKM ≥1 in the four rRNA-depleted total RNA samples from bulk HEK293T cells, and 15,535 genes were detected in the four oligo(dT)-enriched total RNA samples from bulk HEK293T cells