Fig. 1: Developing Bact-Builder.

a Pipeline overview. Bact-Builder takes raw fast5 sequencing data, files, assembles, generates a consensus, and polishes bacterial genomes. b Heatmap comparison of genome sizes of four de novo long read assemblers from laboratory stocks of H37Rv sequenced in triplicate (H37Rv.1-3). The sequence coverage sampled for each analysis is shown in each row on the Y axis. Boxes marked by an X indicate that the assemblies did not pass the Trycycler stage because they could not be reconciled with the other assemblies. * Indicates that 3 out of 4 assemblers could not be reconciled, necessitating that Trycycler was run with only 1 assembler. c Three replicates of laboratory stocks of H37Rv (H37Rv.1-3), showed variability in size depending on assembler used, and consistent sizes when Trycycler was followed by polishing (Bact-Builder output). Dotted line indicates the size of the established H37Rv reference. Data are plotted as means ± SD. d Heatmap of hierarchical clustering of the distance using Euclidean average linkage clustering of differences between all assemblies for H37Rv.1, the Bact-Builder output and the published reference (H37Rv ref) determined by DNAdiff. e Anvi’o pangenome comparing gene clusters in the reference (H37Rv ref) and H37Rv.1 individual assemblies, Trycycler output and the Bact-Builder output.