Table 2 Benchmarking results for simulated PacBio CLR reads. HC haplotype coverage, ER error rate (mismatches + indels + N’s), MC misassembled contigs proportion. NGA50 is labeled with “-” if the uniquely aligned blocks cover less than half of the reference length. The total sequencing coverage in this table is 20,000 ×. Note that CliqueSNV is a reference-based method, whereas the others are de novo assemblers. For running CliqueSNV, we have tried various strategies (see Additional file 1: Table S2) but only the results of the best strategy are reported here. * If contigs are full-length, this number represents the estimated number of haplotypes or strains in the virus mixture. ^† Wtdbg2 consensus as reference, using reads corrected by Strainline error correction. ^‡ High quality reference (the highest abundant strain) using reads corrected by Strainline error correction

From: Strainline: full-length de novo viral haplotype reconstruction from noisy long reads