Fig 1.
Patient genotype influences on different aspects of the chronic wound microbiome.
a, Boxplot of Hill1 diversity as a function of genotype at rs8031916 (TLN2) and rs7236481 (ZNF521), each explaining significant components of variation in the two-cohort mbGWAS. b, Distance-based redundancy analysis based on Bray-Curtis dissimilarities with relative abundance species effects indicated by vectors. rs8031916 significantly explained beta diversity (p < 0.01). The distribution of genotypes at rs8031916 across MDS1 and MDS2 are illustrated by boxplots. c, Boxplot illustrating how residual relative abundances of P. aeruginosa and S. epidermidis were significantly explained by rs8031916 (p < 0.05 for each). d, Communities in which P. aeruginosa (Pa) was present were significantly less diverse than communities in which S. epidermidis occurred (p < 0.05).
Fig 2.
Interspecific interactions inferred from relative abundance correlations.
a, Interaction network with positive associations denoted by black edges and negative associations denoted by orange edges. Species included had a study-wide average relative abundance greater than 5% and had at least one correlation greater than or equal to r = 0.10. b, Barplot of summed positive and negative correlations for each species. Species key b: 1) Anaerococcus hydrogenalis, 2) Anaerococcus lactolyticus, 3) Anaerococcus prevotii, 4) Anaerococcus vaginalis, 5) Corynebacterium tuberculostearicum, 6) Finegoldia magna, 7) Fusobacterium nucleatum, 8) Peptoniphilus harei, 9) Porphyromonas levii, 10) Pseudomonas aeruginosa, 11) Staphylococcus aureus, 12) Staphylococcus epidermidis, 13) Staphylococcus lugdunensis, 14) Streptococcus agalactiae.
Fig 3.
Chronic wound healing time is a function of the wound microbiome and demographics.
Individual regressions of predictor variables on healing duration. Predictor variables included are those that were retained through backward selection for models of healing. For Pseudomonas aeruginosa and Anaerococcus vaginalis, only samples in which these species were observed are plotted. Variables EV1, EV4, and EV5 are population Eigen vectors encompassing genetic ancestry.
Fig 4.
TLN2, functioning in cell adhesion and structure, exhibits genotype-specific alternative transcription.
a, Select ontological terms that were enriched in this study. See S2 Table for a full list of terms. Terms involving TLN2 are colored yellow and number of included genes (out of 15 total included in testing) per term is given in parentheses next to each term. b, Distribution of TLN2 protein coding transcripts on human chromosome 15 with functional domains illustrated per Gough and Goult 2018. Red shading, from left to right denotes the transcript regions amplified from either the canonical isoform, three isoform, or four isoform assay used to compare alternative transcription by genotype (see Methods). c, Standard curve of RT-qPCR assays, with calculated efficiencies shown per assay. d, TLN2 log isoform ratios (calculated as the log of either the three or four isoform Ct divided by the canonical Ct) were significantly associated by rs8031916 genotype. Here, two levels of pairwise comparison are shown: above the dashed line are overall genotype comparisons, while within-genotype assay comparisons are indicated below the dashed line. Significance annotation is as follows: ns = (p>0.05), * = (p<0.05), and ** = (p<0.01).
Fig 5.
Microscopic localization of Tln2 in infected mouse wound bed.
Tln2 was observed in adipocytes (a, b, c), and striated muscle (d, e, f). H&E staining (a, d), IHC negative control (b, e), IHC detection of Tln2 with Alexa Fluor 488 secondary (c, f). Nuclei were stained with DAPI. Scale bars are given for each image.