For calling genetic variants we used the same short-read sequencing libraries as described for the abundance estimations, but we excluded all samples from the Walenbos population as the endosymbiont genomes were reconstructed from this population, and we excluded all samples with a coverage <30. Afterwards, we applied breseq (v0.36.1) [76 (link)] to call genetic variants present in the sequencing libraries in comparison to the reference genomes. We excluded all variants with a frequency <1 and variants affecting transposases or introns from the output and calculated the total number of mutations/kb and relative abundances of variant types per sample.
Free full text: Click here