The effect of variant interference on de novo assembly for viral deep sequencing

Christina J. Castro; Rachel L. Marine; Edward Ramos; Terry Fei Fan Ng

The simulated experiments suggested that genome length and read length influence VI; A longer 225 genome length will produce proportionally more contigs during VI, whereas a longer read length decreases 226 the PID range where VI occurs [ Figure 4 ]. While longer read length improves assembly, unfortunately, 227 platforms that produce long reads such as Oxford Nanopore and PacBio have higher error rates. 23 Until long 228 reads can be produced at high fidelity, researchers must continue to rely on combining long-and short-read The large number of contigs generated due to VI may be overwhelming for most researchers, and for 232 viral ecology studies, could lead to over-estimation of species richness for methods that use contig spectra to 233 infer richness, such as PHACCS or CatchAll. 24, 25, 26 This phenomenon may also impact studies differently 234 depending on the overall goal for generating viral sequence data. For example, some researchers may only be 235 concerned with generating a single major consensus genome, even when variants are detected in the data. The effects of VI could potentially be mitigated by running multiple assembly programs. A previous 243 study testing bioinformatics strategies for assembling viral NGS data found that employing sequential use of 244 All rights reserved. No reuse allowed without permission.