MinION-based long-read sequencing and assembly extends the Caenorhabditis elegans reference genome

“…In the case of the minimap2+miniasm assembly, the best assembly resulted from >10Kb or longest 41X filtering ONT reads, suggesting the inclusion of shorter reads may actually reduce the contiguity of the assembly. A similar observation was also reported in a study assembling the Caenorhabditis elegans genome with ONT reads [38]. However, due to Miniasm lacking a sequence consensus step, the base accuracy in the assembly can only be as good as the input reads, leading to a very poor BUSCO score (1.5%) (Table 1, Supplementary Table S2).…”

“…In our study, identifications of assembled rrn were based mostly on sequence identities higher than 99%, achieving a similar accuracy as in previous studies based on de novo genome assembly [20,22]. Furthermore, we have been able to detect the variability at strain level because the rrn sequences retrieved allowed the correct identification of bacterial species showing 89% and 91% sequence identity, against references, when using the 16S rRNA gene and the entire rrn region as query, respectively.…”

“…For example, the Nanopore RNA-seq consortium study estimated that a significant proportion of native RNA transcripts may be truncated by nanopore signal noise caused by electrical signals associated with RNA motor enzyme stalls, or by otherwise stray current spikes of unknown origin 18 . We also agree that nanopore native RNA read truncation is unlikely due to some fundamental limitation of nanopore-based sequencing, especially considering that ONT 1D genomic DNA sequence reads of several kilobases are consistently achieved without issue using the current pore type 16,24,25 used to sequence both DNA and RNA. Further, such problems could conceivably be addressed, to at least some extent, by training basecallers to reliably recognize relevant nanopore signal noise events which might cause single molecule sequence reads to be truncated or split.…”

“…In the case of the minimap2+miniasm assembly, the best assembly resulted from >10Kb or longest 41X filtering ONT reads, suggesting the inclusion of shorter reads may actually reduce the contiguity of the assembly. A similar observation was also reported in a study assembling the Caenorhabditis elegans genome with ONT reads [38]. However, due to Miniasm lacking a sequence consensus step, the base accuracy in the assembly can only be as good as the input reads, leading to a very poor BUSCO score (1.5%) (Table 1, Supplementary Table S2).…”

“…In our study, identifications of assembled rrn were based mostly on sequence identities higher than 99%, achieving a similar accuracy as in previous studies based on de novo genome assembly [20,22]. Furthermore, we have been able to detect the variability at strain level because the rrn sequences retrieved allowed the correct identification of bacterial species showing 89% and 91% sequence identity, against references, when using the 16S rRNA gene and the entire rrn region as query, respectively.…”

“…For example, the Nanopore RNA-seq consortium study estimated that a significant proportion of native RNA transcripts may be truncated by nanopore signal noise caused by electrical signals associated with RNA motor enzyme stalls, or by otherwise stray current spikes of unknown origin 18 . We also agree that nanopore native RNA read truncation is unlikely due to some fundamental limitation of nanopore-based sequencing, especially considering that ONT 1D genomic DNA sequence reads of several kilobases are consistently achieved without issue using the current pore type 16,24,25 used to sequence both DNA and RNA. Further, such problems could conceivably be addressed, to at least some extent, by training basecallers to reliably recognize relevant nanopore signal noise events which might cause single molecule sequence reads to be truncated or split.…”

“…In the case of the minimap2+miniasm assembly, the best assembly resulted from >10Kb or longest 41X filtering ONT reads, suggesting the inclusion of shorter reads may actually reduce the contiguity of the assembly. A similar observation was also reported in a study assembling the Caenorhabditis elegans genome with ONT reads [38]. However, due to Miniasm lacking a sequence consensus step, the base accuracy in the assembly can only be as good as the input reads, leading to a very poor BUSCO score (1.5%) (Table 1, Supplementary Table S2).…”

“…In our study, identifications of assembled rrn were based mostly on sequence identities higher than 99%, achieving a similar accuracy as in previous studies based on de novo genome assembly [20,22]. Furthermore, we have been able to detect the variability at strain level because the rrn sequences retrieved allowed the correct identification of bacterial species showing 89% and 91% sequence identity, against references, when using the 16S rRNA gene and the entire rrn region as query, respectively.…”

“…For example, the Nanopore RNA-seq consortium study estimated that a significant proportion of native RNA transcripts may be truncated by nanopore signal noise caused by electrical signals associated with RNA motor enzyme stalls, or by otherwise stray current spikes of unknown origin 18 . We also agree that nanopore native RNA read truncation is unlikely due to some fundamental limitation of nanopore-based sequencing, especially considering that ONT 1D genomic DNA sequence reads of several kilobases are consistently achieved without issue using the current pore type 16,24,25 used to sequence both DNA and RNA. Further, such problems could conceivably be addressed, to at least some extent, by training basecallers to reliably recognize relevant nanopore signal noise events which might cause single molecule sequence reads to be truncated or split.…”