I’m collecting some answers to FAQs here.
Why do you sequence short fragments?
Adapted from this longer guide: https://artic.network/quick-guide-to-tiling-amplicon-sequencing-bioinformatics.html
We have standardised on short amplicons (~400bp) for this protocol. We do this because short amplicons are more likely to amplify with degraded RNA. RNA can be degraded due to effects of sample storage (nuclease activity and freeze-thawing), sample preparation steps (e.g. heat treatment) and during the RNA extraction process. This is particularly notable with low genome copy number / high Ct value samples.
Amplicon dropout is expected as the Ct value goes up. Our thinking is that if you choose long amplicons then if those dropouts will have a much larger effect on reduction in genome coverage than with shorter amplicons.
Of course, in an ideal world, amplicons would be as long as possible: this would help with linking variants into haplotypes. For some samples with high viral copy numbers it should definitely be possible to amplify regions much larger than 400 bp (we have tested up to 2 kb with Ebola in the past and even managed 6 kb on one occasion).
The other advantage of a 400 bp scheme is that it is compatible with ligation protocols on multiple sequencing instruments (e.g. 2x250 on Illumina).