Fig 1.
Overview of the workflow from gathering fleas (panels A and B) to DNA extraction (panel C), sequencing (panel D), and analysis (panel E) in a study evaluating short- and long-read 16S rRNA gene sequencing of pooled fleas from different sources (cats, dogs and traps) in flea-infested households in Florida.
Fig 2.
Circle packing diagram of the flea pools sequenced by household and source.
Outer gray circles represent a household and are labeled by house number. Inner gray circles represent a source (cat, dog, or trap) within the household, and the inner colored circles represent the number of pools sequenced from that source. Ex: Household 15 includes 1 sampled trap with 4 sequenced pools, and 2 sampled cats, one cat with 1 sequenced pool and a second cat with 2 sequenced pools.
Fig 3.
Proportion of (a) bacterial ASVs and (b) bacterial reads classified to different terminal taxonomic levels from short-read and long-read 16S rRNA gene sequences.
Short- and long-read 16S sequencing data was generated from the same set of pooled flea samples. Automated taxonomic assignment using the Silva reference database was performed in the DADA2 R package following the developer recommendations.
Fig 4.
The prevalence and abundance of the top 5 most abundant and prevalent genera following short- and long-read 16S rRNA gene sequencing.
Points are colored based on the categorization of genera as known microbiome members (pink), possible microbiome members (green), or suspect contaminants (black). Known microbiome members include the well-documented taxa Wolbachia, Rickettsia, and Bartonella. Possible microbiome members and candidate taxa were designated as such following evaluation of decontam scores and scrutiny of literature describing taxa identified in prior C. felis sequencing studies, as well as literature describing common contaminants in microbiome studies.
Table 1.
Decontam scores, prevalences (presence or absence in a tested pool), and abundances for core and candidate taxa found in short- and long-read 16S sequencing. Scores were defined by the decontam-frequency method at the ASV level. Scores range from 0 to 1 with high scores suggesting true sample origin and low scores suggesting contaminant origin. For genera, the range and median are across all ASVs assigned to that genus.
Fig 5.
Relative abundance and decontam frequency score for all bacterial ASVs present in two-or-more samples following short- (left) and long-read (right) 16S rRNA gene sequencing.
Decontam scores range from 0 to 1 with high scores suggesting true sample origin and low scores suggesting contaminant origin. Colors indicate known core taxa (Wolbachia in pink, Rickettsia in green, Bartonella in yellow) and a proposed fourth core taxa (Spiroplasma in blue) based on its high prevalence, abundance, and decontam frequency score in the long-read dataset.
Table 2.
Summary of species-level taxonomic assignments between two different methods for Rickettsia and Bartonella ASVs generated from long-read/full-length 16S sequencing of the C. felis microbiome.
Fig 6.
PCoA plot depicting lack of major effect from any one source (dog, cat, or trap) on community composition following long-read 16S rRNA gene sequencing of flea pools from different sources in flea-infested homes in Florida.
Table 3.
Top 5 results of two-sided Wilcoxon rank sum tests evaluating the differential abundance of long-read ASVs between sources (cats, dogs, and traps) and between households with and without dogs.