Fig 1.
Key steps in the Unicycler pipeline.
Table 1.
Reference genomes for simulated read sets.
Fig 2.
Simulated short-read assemblies: Errors.
Misassembly and small-error (mismatches and indels) rates for assemblies of simulated short-read sets, summarising results across all reference genomes and replicate tests (total 360 per assembler).
Fig 3.
Simulated short-read assemblies: NGA50.
NGA50 for assemblies of simulated short-read sets, summarising results across all reference genomes and replicate tests (total 360 per assembler).
Fig 4.
Simulated hybrid assemblies: Errors.
Error rates for hybrid assemblies of simulated short-read and long-read sets, summarising results across all reference genomes and replicate tests (total 2520 per assembler).
Fig 5.
Simulated hybrid assemblies: NGA50 against long-read depth.
Mean NGA50 values for hybrid assemblies of simulated read sets. Mean values were calculated across all read lengths, read accuracies and replicate tests for each reference genome (210 hybrid-read sets each); the top panel shows mean values for all 12 reference genomes (2520 hybrid-read sets). Horizontal dashed lines indicate the N50 size of the reference genome. For the bacterial genomes, this is the size of their only chromosome; for Saccharomyces, it is the size of chromosome XIII, an intermediate-sized replicon in the genome.
Fig 6.
Simulated hybrid assemblies: Read length and accuracy.
NGA50 values segregated by read length and read accuracy. These plots summarise results across all reference genomes and replicate tests, but only include the tests of 8x long-read depth. For read lengths, the p-value is from a two-tailed t-test. For read accuracies, the p-value is from a one-way ANOVA test.
Table 2.
Real E. coli K-12 substr. MG1655 read sets.
Fig 7.
E. coli K-12 assemblies: NGA50 against long-read depth.
Mean NGA50 values for hybrid assemblies of real E. coli read sets, summarised across 20 replicate tests at each depth. Top panel shows mean values for all six long-read sets.
Fig 8.
Klebsiella pneumoniae INF274 assembler comparison.
Final assemblies of Klebsiella pneumoniae INF274 produced by Unicycler, SPAdes, HGAP and Canu. The contigs/graph of the assembly are shown on the left, coloured by replicon. The read depth plot of plasmid 1’s contig is shown on the right. Low read depth at the ends of the contig is indicative of start-end overlap.
Fig 9.
Klebsiella pneumoniae INF125 ONT assemblies over sequencing time.
Assembly metrics of K. pneumoniae INF125 produced by Unicycler, SPAdes, npScarf and miniasm over a four-hour period of sequencing. Miniasm assemblies contain error rates comparable to that of the raw reads and are therefore excluded from the error rate plots.