(A) Drops containing RNA and RT-PCR reagents are created with a microfluidic T-junction and carrier oil. Brightfield microscopy images of the drop formation are shown below, the middle image showing the generation of one population of drops from a single reaction mixture, and the lower the generation of two populations from two mixtures. The red arrows indicate the direction of emulsion flow in the illustrations. (B) After formation, the drops are picoinjected with reverse transcriptase using a picoinjection channel triggered by an electric field, applied by an electrode channel immediately opposite the picoinjector. Picoinjection fluid is pictured as dark gray in the schematic diagram. (C) The picoinjected drops are collected into a tube, thermocycled, and imaged with a fluorescent microscope.

(A) Control RT-PCR reactions containing PC3 cell total RNA were emulsified on a T-junction drop maker, thermocycled, and imaged. FAM (green) fluorescence indicates Taqman detection of an EpCAM transcript and Cy5 (red) indicates detection of CD44 transcripts. Brightfield images (BF) of the same drops are shown in the image panel on the far right. The red arrows indicate the direction of emulsion flow in the illustrations. (B) RT-PCR reactions lacking reverse transcriptase were emulsified on a T-junction drop maker and subsequently picoinjected with reverse transcriptase. Picoinjection fluid is pictured as dark gray in the schematic diagram on the left. Brightfield images demonstrate that the drops roughly doubled in size after picoinjection. (C) RT-PCR reactions subjected to picoinjection omitting the reverse transcriptase show no Taqman signal for EpCAM and CD44, demonstrating the specificity of the Taqman assay. Scale bars  = 100 µm.

(A) Scatter plots of FAM and Cy5 drop intensities for a control sample (left) and picoinjected sample (right). The gating thresholds used to label a drop as positive or negative for Taqman signal are demarcated by the lines, and divide the scatter plot into quadrants, double negative drops (–,–), FAM positive (–,+), Cy5 positive (+,–), positive for both FAM and Cy5 (+,+). Numbers of drops in each quadrant are indicated. (B) The bar graph shows the average Taqman positive drop count with picoinjection relative to the normalized count for CD44 and EpCAM Taqman assays for control populations. The control detection rate value is defined as 1 for each replicate. The data represent the average of four independent experimental replicates.

(A) Non-picoinjected drops. Control RT-PCR reactions containing mixed PC3 cell total RNA and Raji cell total RNA were emulsified with a T-junction drop maker, thermocycled, and imaged. Merged FAM and HEX fluorescent images are shown with FAM (green) fluorescence indicating Taqman detection of an EpCAM transcript and HEX (red) indicating the presence of PTPRC transcripts. The yellow drops indicate the presence of multiplexed Taqman assays, where EpCAM and PTPRC transcripts were co-encapsulated in the same drop. The brightfield images (BF) are shown in the panel on the right. The red arrows indicate the direction of emulsion flow in the illustrations. (B) Picoinjected drops. A double T-junction drop maker simultaneously created two populations of drops that were immediately picoinjected. One drop maker created drops containing only Raji cell RNA, and the other drops containing only PC3 cell RNA. Both drop types initially lack reverse transcriptase, which is added via picoinjection just downstream of the drop makers. The overwhelming majority of drops display no multiplexing, demonstrating that transfer of material during picoinjection is very rare. Scale bars  = 100 µm.

(A) Scatter plots of FAM and HEX drop intensities for a co-encapsulated control sample (left) and dual population picoinjected sample (right). Using this analysis, large numbers of Taqman multiplexed drops were identified in the co-encapsulated controls that were virtually absent in the dual population picoinjected drops (upper right quadrants of gated scatter plots). (B) A bar graph of different bright drop populations relative to the total bright count for co-encapsulation control and for dual population picoinjection. The data represent the average of three experimental replicates.

(A) An emulsion was made consisting of two populations of drops, one containing RNA recovered from Raji cells, and the other from PC3 cells. The drops were collected into a syringe, incubated off chip, and then re-introduced into a microfluidic device to picoinject. The drops were then collected, thermocycled, and imaged. These drops are somewhat more polydisperse and displayed higher multiplexing rates (1%) than the drops picoinjected on the same device on which they were formed, which is most likely due to merger of some of the drops during incubation and reinjection. The ability to reinject emulsions following incubation to add reagents is critical for numerous droplet-based molecular biology assays. (B) Brightfield images of picoinjected emulsions. Scale bars  = 100 µm.