Extended Data Fig. 2: Barcode pairs dictionary generation with paired-end short read sequencing.
From: Pool-packaged AAV libraries exhibit extensive length-dependent and homology-dependent chimerism
(a) Schematic of procedure to generate the BC1-BC2 pairs dictionary. Parental plasmid dock p146 was used as template for PCR (primers o945 + o946v2) to append Illumina P5 and P7 handles, and sequenced on NS2000 with paired-end sequencing to retrieve barcode pair representation. (b) Top: Table of number of barcodes and reads for the different categories shown in panel c. Bottom: Table of number of barcode pairs, reads, and proportion displaying retention at every filtering step (i.e., both barcodes present in the well-represented set and uniquely paired [>99% reads] with a single other barcode). (c) Read count distribution by summing only on respective barcodes (not pairs) for BC1 (left) and BC2 (right). Dashed lines indicate cut-offs used for the mid- and high-counts classes of BCs. (d) Two-dimensional distributions showing the proportion of reads to the BC arising from the pair on y-axis (BC1 left panel, BC2 right panel) vs. the total read count to the given pair (x-axis). Retained pairs are within the green boundary (>5 reads and >99% unique proportion). (e) Similar to panel d, but now showing unique pairing proportions for BC1 (y-axis) vs. BC2 (x-axis) as a two-dimensional histogram. Retained pairs are in the top right corner within the green boundaries (i.e., BC1-BC2 pairs showing >99% uniqueness of read mapping to both barcodes). Marginalized histograms for BC1 and BC2 are shown on right and top respectively.
