Fig 1.
A schematic of the computational pipeline used to determine the reproductive manipulator infection status and symbiont titer of a sequencing run.
The pipeline (A) takes in a sample’s unique identification number, then downloads two million reads (includes symbiont and host genomic reads). Then, (B) reads are aligned to Wolbachia, Arsenophonus, Spiroplasma, Cardinium, and Rickettsia reference genomes. Also, reads are aligned to a set of 1066 single copy ancestral orthologs obtained from ORTHODB v9 to estimate host coverage without requiring a reference genome. (C) Summary statistics for sample reads aligned to each reference are computed. If a sample had between 0.1 and 0.9 breadth of coverage, the full dataset was downloaded and the workflow repeated to prevent false negative calls. We apply coverage breadth and depth cutoffs to classify infection status as positive, or negative. To estimate symbiont titer, we compare the depth of coverage of host reads mapped to a set of single copy orthologs, to the coverage of symbiont reads mapped to a symbiont reference genome. Please see the methods section for more details on the approach for classification and titer computation.
Table 1.
Raw infection counts of Wolbachia, Spiroplasma, Rickettsia, Cardinium and Arsenophonus infecting arthropods and nematodes.
The frequency of positive samples or species is listed in parentheses.
Table 2.
Estimated infection frequencies and confidence intervals from the data for Wolbachia, Spiroplasma, Rickettsia and Arsenophonus infecting arthropods and nematodes.
All species in the dataset were downsampled to a maximum 100 individuals for frequency inference. We used a minimum infection frequency of 0.001 to classify a species as positively infected (See Methods).
Fig 2.
Phylogeny of Arthropoda orders tested and number of reproductive manipulator positive species within each order.
The frequency of reproductive manipulator-positive species listed in parentheses. No frequency is listed if there was no infection within an arthropod order. We used the Tree of Life taxonomic and phylogenetic package and rotl [74], to group host species by their orders. We labeled arthropod clades containing two or more taxa with subphylum or class.
Table 3.
Wolbachia global infection frequencies and confidence intervals generated for arthropod orders.
All species in the dataset were downsampled to a maximum 100 individuals. Confidence intervals were generated using 1000 bootstrap replicates fitting a beta-binomial model to species infection frequency data among orders. A minimum infection frequency of 0.001 was used to classify a species as positively infected (See Methods).
Fig 3.
Titer variation across species infected with Wolbachia and between diverse reproductive manipulator clades.
To increase readability of both plots, categories were randomly downsampled to show 100 samples. The y-axis is log10 scaled. (A) Titer for Wolbachia positive arthropod species with at least three samples were plotted from low to high titer and color coded by taxonomic order. (B) Titer for Wolbachia, Rickettsia, Arsenophonus, Spiroplasma infected samples. We plotted up to three samples for every species infected with Wolbachia to show the range of Wolbachia across tested arthropod species. Titer variation within host species is significant, and this variation is not due to pooled sequencing samples. Our results suggest a symbiont and host genetic contribution to shaping within-host infection densities. Asterisks indicate statistically significant relationships between titer and arthropod species where 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’.