| Geminiviruses are plant infecting viruses with circular single stranded DNA genomes. Prior studies have shown that geminiviruses have a relatively high mutation rate approaching that of RNA viruses, suggesting they may utilize a quasispecies-like replication strategy. These prior studies estimated geminivirus mutation rates by dideoxy sequencing of virus progeny cloned out of infected plants. While dideoxy sequencing is sufficient for estimating overall mutation rates, the limited depth of this approach is insufficient for studies aimed at identifying mutation hot spots and conserved regions. Theoretically, next generation sequencing (NGS) platforms have enough power to sequence the progeny of small viruses thousands of times and estimate the mutation rates of each position within the genome. While the throughput of these platforms is sufficient for these studies it is not clear whether the accuracy of these platforms is high enough for identification of mutations within populations of viral progeny. Here we report on the use of ultra-deep sequencing with Ion Torrent to characterize the mutation rate of Beet Severe Curly Top Virus (BSCTV). The reported error rate of Ion Torrent is over a hundred times higher than the known geminivirus mutation rates, thus it was necessary to use a combination of read trimming and software assisted error correction. After extensive error cleaning and consequent sequence loss, the average mutation frequency for virus replicated in Nicotiana benthamiana was 8.53 x 10 -6substitutions per site. This is ten to a hundred times lower than previous estimates and it is the lowest mutation rate reported for geminiviruses so far, possibly because the necessary data cleanup removed many of the real mutations present in the data along with the errors. The mutations that remained in the datasets were mapped to non-vital regions on the genome, however many occurred in regions conserved in other curtoviruses. Long deletions associated with Dl DNAs were also found in cloned virus progeny and Ion Torrent reads. But although ultra-deep sequencing is a powerful technique to produce data, the complications of error cleanup compromise its reliability for quantitative analysis at precision level required for mutation rate estimations. |
