Abstract
Purpose:
Changes in the homeostatic ocular surface (OS) microbiota may indicate an ongoing development of the OS disease, which may impair the visual potential or cause blindness. We applied modern bioinformatic tool and multivariate statistics in an attempt to analyze the microbial compound of the healthy OS and track the infection-related changes.
Methods:
We applied 454 Roche sequencing of the 16S rRNA libraries of the OS bacteria in order to analyze the OS microbiome. The sequences were processed via MOTHUR and classified with RDP II Classifier. A deeper taxonomic assignment achieved with a novel BLAST-based tool TUIT. Between-group comparison was performed with MEGAN4, multivariate statistics and HMPTree R-package.
Results:
We compared microbial compositions at the human cornea and conjunctiva with that on the facial skin of the same subjects; samples from both genders were examined. We performed comparative analysis using HMPTree-based statistics analysis and revealed a significant difference between the composition at OS and corresponding facial skin samples. Multivariate statistical analyses suggested that certain within-group taxonomic subclusters may exist within groups. Visual bacterial group comparison, performed with MEGAN4, allowed us to compare the most abundant taxa at phylum and genus levels in healthy OS and during the development of ulcerative bacterial keratitis. Application of the HMPTree-based statistics analysis helped us to analyze the infection-dependent changes during the development of ulcerative keratitis.
Conclusions:
Successful application of a set of effective bioinformatics and statistics provided a new insight into the ocular microbiome composition and infection-related changes. Our results showed that the ocular surface is a solitary microbiome niche, distinct from most of other human sites.
Keywords: 473 computational modeling •
482 cornea: epithelium •
594 microbial pathogenesis: experimental studies