Figure 2. Single-labeled-zone fluorescent barcodes.

(a) Superimposed TIRF microscopy images of five barcode species (top) and the statistics from manual counting (bottom). The barcode types are noted under the horizontal axis of the diagram. Each bar-graph is generated by manually counting the objects in a 50 × 50 μm² image (~40 barcodes, the exact sample size N is noted beside the corresponding bar-graph). (b) A representative image of the equimolar mixture of 27 barcode species. (c) Statistics obtained by analyzing twenty-seven 50 × 50 μm² images of the 27-barcode mixture (1,485 barcode molecules). Left: counts of the 27 species (55 ± 9; mean ± s.d.). A representative TIRF image (1.4 × 0.7 μm²) of each barcode type is placed underneath the corresponding bar. Right: object sorting result shown as a pie-chart. Color scheme used for the bar-graphs in (a) and the pie-chart in (c): blue, correct barcodes (qualified barcode with expected identity); green, incorrect barcodes (qualified barcode with unexpected identity); red, monomer nanorods (one spot or two touching spots); purple, barcodes with wrong geometry (bending angle <120°, see Methods in SI); orange, barcodes containing at least one spot with two colors. Note that in the 27-barcode pool, correct vs. incorrect barcodes were not distinguishable because all barcode types were expected. As a result, the bars and pie representing the qualified barcodes in (c) are shown in gray. Scale bars: 5 μm.