Gene content and arrangement of the maternally inherited markers
Mitochondrial minichromosomes . Neither the Spades nor the aTRAM
method was successful in reconstructing complete sequences of the eleven
mitochondrial minichromosomes. However, the aTRAM assemblies contained
whole coding regions and were used for both the phylogenetic
reconstruction and the comparison of gene content between the SE and SW
lineages. Although yielding considerable genetic differences (Table S2),
the mitochondrial minichromosomes show identical arrangement of the
genes (shared synteny) in both the SW and SE lineages (Table S3). This
arrangement is also very similar to that in the related louse species,P. spinulosa . Concatenation of the minichromosomes produced a
17,253 bp long matrix. When phylogenetically analyzed, it yielded a tree
with two very distant clusters corresponding to the SE and SW lineages
(Figure 4b, Figure S2). Within both clusters, the distances were
significantly lower than the distance between the clusters. However, the
overall range of distance was higher within the SW than within the SE
(Table S2, Figure 4), likely reflecting the broader geographic sampling
range of the SW lineage.
Legionella. Genomes of the symbiont L. polyplacis revealed
phylogenomic structure parallel to the mtDNA (Figures 4a, S3), with a
deep genetic split between the SW and SE lineages. The complete genomes
displayed a high degree of similarities with all pairwise comparisons
exceeding 99% identity. The contrast between the intra- and
inter-cluster comparisons is better illustrated by the counts of the
observed differences, which were 27-212 within the SW cluster and 1-57
within the SE cluster, compared to 3,683 - 3,699 between the clusters
(Table S2). When comparing the genome sequences, we did not find any
clear instance of missing genes. The majority of the gaps introduced by
genomic alignment span just one or two nucleotides and were placed in
intergenic regions (only one deletion span across 26 nucleotides, also
located between the gene coding sequences). The annotations provided by
RAST contained several differences between the two clusters, indicating
that a gene present in one lineage is shortened or missing in the other
cluster. In all of these cases, however, the differences were not caused
by a convincing absence of the gene sequence but rather by failure of
the algorithm to recognize the sequence as coding a gene, most likely
due to the aberrant nature of highly derived symbiont genomes.