hin et al PD Whole Tree Observed Species Non-lens wearers Conjunctiva +- 3o8:中 Non-lens wearers Skin under he eye Lens wearers g Contact lens 10 20 30 100 Phylogenetic diversity Number of observed specie rarefication to 2,090 reads per sample. The nonparametric P values were calculated using 999 Monte Carlo permutations. + outlier samples excluded from the G FIG 1 Bacterial diversity in the conjunctiva, skin, and contact lens between non-lens wearers and lens wearers. Box plots of alpha diversity were generated with nity composition(linear discriminant analysis [LDA] score= mental material). However, there was no clustering by subject or >3.0; see Fig SIB)and structure(permutational analysis of vari- sampling time points( Fig 2A and B). Moreover, the conjunctival ance PERMANOVA] P value = <0.001; see Fig SIC), as mea- microbiota in lens wearers was more similar to human skin mi sured by unweighted/weighted UniFrac distances. In addition, the crobiota from a previous study (15)than to that in non-lens wear use of contact lenses caused only minor changes in the conjunct- ers(Fig. 3). Weighted Uni Frac distances depicted on principal Fig S2), supporting the idea of an effect of anesthetic. Due to the Fig S4). Lens wearers also had higher interindividual variability in parent perturbation caused by the eye drops, we performed their ocular microbiota than non-lens wearers(unweighted Uni- further analyses solely on the 20 subjects sampled in the laboratory Frac, nonparametric P value <0.001; Fig. 2C and D) without anesthesia The conjunctival microbiota of non-lens wearers was influ A total of 250 samples obtained in the laboratory (116 conjunc- enced by gender(PERManova P <0.001; see Fig S5A and B in tiva, 114 skin under the eye, and 20 contact lenses)were rarefied at the supplemental material). Numbers of Acinetobacter organisms ,090 sequences per sample(see Table S2 in the supplemental and of members of family Enterobacteriaceae were increased and material). Notably, the conjunctival samples showed higher alpha those of Anaerococcus were depleted in the ocular microbiota of diversity than the skin under the eye or contact lenses(P value= female subjects compared to male subjects(LDa score =>3.0; <0.05 [nonparametric Student's t test]; Fig. 1). In the same se- see Fig. S5C and D). Regardless of gender, lens wearers had quencing run in which we sequenced the eye project samples, we skin-like conjunctival microbiota compared to non-lens wearers sequenced samples of the vaginal microbiota from different sub-(see Fig S5E and F). Compared to levels seen with non-lens wear- jects. The alpha diversity of the vaginal microbiota was similar to ers, the ocular microbiota of lens wearers was enriched in Pseu that reported by the HMP Consortium(15), and the bacterial domonas, Acinetobacter, Methylobacterium, and lactobacillus diversity in the skin under the eye in our study was similar to the (LDA score=>3.0; Fig. 4). In non-lens wearers, these were de- bacterial diversity in the skin of the face reported by Bouslimani et tected at a higher relative abundance in skin samples than in the al.(16)(see Fig S3A). We consider these good positive controls conjunctiva(except for Lactobacillus)(Fig. 4), suggesting that and are confident in the diversity found in the conjunctiva and these bacteria could be classified as skin bacteria Levels of Haemoph kin under the eye lus, Streptococcus, Staphylococcus, and Corynebacterium were depleted There were no significant differences in bacterial alpha diver- in the ocular microbiota of lens wearers compared to non-lens wear- 4. y between the conjunctiva of lens wearers and that of non-lens ers(LDa score=>3.0; Fig 4). Contact lenses demonstrated ahigher wearers(nonparametric Student's t test; Fig. 1). The microbial relative abundance of Acinetobacter and Methylobacterium than the structures of the conjunctiva and skin under the eye were more conjunctiva(lda score=>3.0; Fig 4 dissimilar in non-lens wearers than in lens wearers(unweighted We compared the conjunctival microbiota with that UniFrac, nonparametric t test P value = <0.001; Fig 2A and B). under the eye. In relation to the skin, the conjunctiva Consistently, the conjunctival microbiota of lens wearers was eye(non-lens wearers) had higher abundances of Haer more similar to the microbiota of the skin under the eye than was Neisseria, Streptococcus, Staphylococcus, Rothia, and Corynebacte- the case with the non-lens wearers(unweighted Uni Frac distance, rium and lower abundances of Pseudomonas, Acinetobacter, Sph P value <0.001)(Fig 2A and B; see also Fig S4 in the supple- ingobium, and Methylobacterium (LDa score =>3.0; Fig. 4B) 2 mBio mbio. asm. org March/April 2016 Volume 7 Issue 2 800198-16nity composition (linear discriminant analysis [LDA] score 3.0; see Fig. S1B) and structure (permutational analysis of vari￾ance [PERMANOVA] P value 0.001; see Fig. S1C), as mea￾sured by unweighted/weighted UniFrac distances. In addition, the use of contact lenses caused only minor changes in the conjuncti￾val microbiota of subjects at the ophthalmology practice (see Fig. S2), supporting the idea of an effect of anesthetic. Due to the apparent perturbation caused by the eye drops, we performed further analyses solely on the 20 subjects sampled in the laboratory without anesthesia. A total of 250 samples obtained in the laboratory (116 conjunc￾tiva, 114 skin under the eye, and 20 contact lenses) were rarefied at 2,090 sequences per sample (see Table S2 in the supplemental material). Notably, the conjunctival samples showed higher alpha diversity than the skin under the eye or contact lenses (P value 0.05 [nonparametric Student’s t test]; Fig. 1). In the same se￾quencing run in which we sequenced the eye project samples, we sequenced samples of the vaginal microbiota from different sub￾jects. The alpha diversity of the vaginal microbiota was similar to that reported by the HMP Consortium (15), and the bacterial diversity in the skin under the eye in our study was similar to the bacterial diversity in the skin of the face reported by Bouslimani et al. (16) (see Fig. S3A). We consider these good positive controls and are confident in the diversity found in the conjunctiva and skin under the eye. There were no significant differences in bacterial alpha diver￾sity between the conjunctiva of lens wearers and that of non-lens wearers (nonparametric Student’s t test; Fig. 1). The microbial structures of the conjunctiva and skin under the eye were more dissimilar in non-lens wearers than in lens wearers (unweighted UniFrac, nonparametric t test P value 0.001; Fig. 2A and B). Consistently, the conjunctival microbiota of lens wearers was more similar to the microbiota of the skin under the eye than was the case with the non-lens wearers (unweighted UniFrac distance, P value 0.001) (Fig. 2A and B; see also Fig. S4 in the supple￾mental material). However, there was no clustering by subject or sampling time points (Fig. 2A and B). Moreover, the conjunctival microbiota in lens wearers was more similar to human skin mi￾crobiota from a previous study (15) than to that in non-lens wear￾ers (Fig. 3). Weighted UniFrac distances depicted on principal coordinate analysis (PCoA) plots also supported these results (see Fig. S4). Lens wearers also had higher interindividual variability in their ocular microbiota than non-lens wearers (unweighted Uni￾Frac, nonparametric P value 0.001; Fig. 2C and D). The conjunctival microbiota of non-lens wearers was influ￾enced by gender (PERMANOVA P 0.001; see Fig. S5A and B in the supplemental material). Numbers of Acinetobacter organisms and of members of family Enterobacteriaceae were increased and those of Anaerococcus were depleted in the ocular microbiota of female subjects compared to male subjects (LDA score 3.0; see Fig. S5C and D). Regardless of gender, lens wearers had a skin-like conjunctival microbiota compared to non-lens wearers (see Fig. S5E and F). Compared to levels seen with non-lens wear￾ers, the ocular microbiota of lens wearers was enriched in Pseu￾domonas, Acinetobacter, Methylobacterium, and Lactobacillus (LDA score 3.0; Fig. 4). In non-lens wearers, these were de￾tected at a higher relative abundance in skin samples than in the conjunctiva (except for Lactobacillus) (Fig. 4), suggesting that these bacteria could be classified as skin bacteria. Levels ofHaemophi￾lus, Streptococcus, Staphylococcus, andCorynebacteriumwere depleted in the ocular microbiota of lens wearers compared to non-lens wear￾ers (LDA score3.0; Fig. 4). Contact lenses demonstrated a higher relative abundance of Acinetobacter and Methylobacterium than the conjunctiva (LDA score 3.0; Fig. 4). We compared the conjunctival microbiota with that in the skin under the eye. In relation to the skin, the conjunctiva of normal eye (non-lens wearers) had higher abundances of Haemophilus, Neisseria, Streptococcus, Staphylococcus, Rothia, and Corynebacte￾rium and lower abundances of Pseudomonas, Acinetobacter, Sph￾ingobium, and Methylobacterium (LDA score 3.0; Fig. 4B). FIG 1 Bacterial diversity in the conjunctiva, skin, and contact lens between non-lens wearers and lens wearers. Box plots of alpha diversity were generated with rarefication to 2,090 reads per sample. The nonparametric P values were calculated using 999 Monte Carlo permutations. , outlier samples excluded from the analyses; **, P value 0.05. Shin et al. 2 ® mbio.asm.org March/April 2016 Volume 7 Issue 2 e00198-16 mbio.asm.org on June 29, 2016 - Published by mbio.asm.org Downloaded from