FIGURE LEGENDS
FIGURE 1 . Conjunctival microbiome compositions of healthy and AC participants. A, Microbiome of healthy subjects (n = 48) and AC patients with SAC/PAC (n = 21) or VKC (n = 18) classified at the kingdom level. B, Microbiome of healthy subjects and AC patients classified at the genus level.
FIGURE 2 . Taxonomic differences between individuals with AC (n = 39) and healthy individuals (n = 48). A, Alpha diversity measured with the Shannon index was computed for AC and healthy samples. B, Beta diversity measured with Bray-Curtis dissimilarity within AC and healthy samples. P values of alpha and beta diversity were computed using a two-sided Wilcoxon test. C, PCoA of samples from all 87 participants based on species-level Bray-Curtis distance. D, LDA effect sizes of species differentially abundant (LDA effect size > 4) in the participants with and without AC. E, Relative abundance distributions of species enriched and reduced in AC. Species with fold change > 5 were displayed. Relative abundances in the heatmap were centered and scaled across all the samples.
FIGURE 3 . Functional differences between individuals with AC (n = 39) and healthy individuals (n = 48). A, KEGG orthology, EC gene families, EggNOG clusters of orthologous groups, and Pfam protein families that are enriched in AC patients versus healthy subjects. Only the protein family signatures identified with LEfSe (Kruskal-Wallis rank-sum test P < 0.05 and LDA effect size > 2) are displayed in the scatter plots. The gene families (enriched in AC) with LDA effect size > 3 are highlighted in red. RP, ribosomal protein. B, Prevalence of CARD antibiotic resistance genes in the conjunctival metagenome from patients with AC and healthy individuals.
FIGURE 4 . Microbiome differences between individuals with SAC/PAC and VKC. A, Alpha diversity measured with the Shannon index was computed for healthy subjects (n = 48) and patients with SAC/PAC (n = 21) and VKC (n = 18). B, Beta diversity measured with Bray-Curtis dissimilarity within healthy, SAC/PAC, and VKC samples. C, PCoA of samples from 39 AC participants with SAC/PAC or VKC based on species-level Bray-Curtis distance. D, LDA effect sizes of species differentially abundant (LDA effect size > 3) in SAC/PAC or VKC.
FIGURE 5 . Relative abundance distributions of species differentially abundant between individuals with SAC/PAC and VKC. A, Overview of the distributions of species associated with the two types of AC. Relative abundances in the heatmap were centered and scaled across the samples from individuals with AC. B, Relative abundances of the species enriched in SAC/PAC. C, Relative abundances of the species enriched in VKC. Only the species that were present in > 50% samples with mean relative abundances > 0.1% in the overabundant group were displayed in the B and C.