increasing false negatives by disregarding good data. This again reinforces the need to understand individual sequencing runs, even if they originate from the same technology (in our case, the sequencing runs used to create error models for the two Illumina simulations were performed by the same technician on the same machine). Without simulation, it would be difficult to choose an optimal M.A.Q. value and almost impossible to interpret any findings. Following this simulation, for a B. aphidicola sequencing experiment resembling our Illumina v5 simulation a M.A.Q. of 39 will give confident and accurate results. Furthermore, we expect to identify 100% of SNPs with frequency > = 3% and 71% of SNPs with frequency of 1% (see Figure 2). We also expect that across the length of the genome, 101 false positives with frequency of 1% (+/-1%) will also occur.
