Using the phred-scores to estimate base error rates by default isn't ideal. These typically underestimate the true error. Furthermore, have a separate error rate for each base position is unnecessarily complicated and slows down the MCMC.
It would be better to remove use of the phred-scores and default to a sensible constant error rate. E.g., 0.0024 as in Pfeiffer (2018).
Using the phred-scores to estimate base error rates by default isn't ideal. These typically underestimate the true error. Furthermore, have a separate error rate for each base position is unnecessarily complicated and slows down the MCMC.
It would be better to remove use of the phred-scores and default to a sensible constant error rate. E.g., 0.0024 as in Pfeiffer (2018).