Concerning the Paper “Ocular Surface Microbiome in Health and Disease”

To the Editor: Firstly, we would like to express our gratitude toward the authors, Gomes et al,1 for referring to our paper "Current Knowledge on the Human Eye Microbiome: A Systematic Review of Available Amplicon and Metagenomic Sequencing Data" published in Acta Ophthalmologica in 2020.2 We would like to, however, point out some misleading rephrasing of our systematic review's results, which have led to misinterpretation by other authors meanwhile. For our review, we obtained access to raw data from 11 studies. This allowed us to reanalyze them and define the commonly found microbiota or "core" ocular surface microbiota. In the conclusion section (page 510) of the paper published in your journal, this core bacteria have been described as "40% of the total ocular surface". Due to the sentence structure, this leads to confusion with area measures. It would be accurate to rephrase to "40% of the total ocular surface microbiome" when referring to our results as we did not analyze surface dimensions. Furthermore, several imprecisions lead to incorrect reports of our publication, we listed the main below. Page 507: "A recent systematic review of 11 published controlled cohorts used 16S rRNA sequencing to define, at a phylum level, the most commonly present bacteria or the core genera." We reviewed more than 11 cohorts, the 11 cohorts (page 507) mentioned in our review are the once we had access to raw data, making it possible to reanalyze the data. Furthermore, the last part of the sentence implies that we only defined the core microbiome at the phylum level, which is incorrect. Page 507: "A publication using 16S rRNA sequencing described the phyla levels of the microbiome core at a pediatric level (younger than 18 years old). The 3 main phyla were Actinobacteria, which represented 53% of bacteria, Proteobacteria, which represented 39%, and Firmicutes, which represented 8%. No core genera belong to the phylum Bacteroidetes." In terms of the pediatric core ocular surface microbiome, we only had raw data of 2 publications3,4 at pediatric level (<18 years old) and our criteria for "core" were met only by the genus Corynebacterium as it had a relative abundance >1% and was found in both publications. Page 510: "The core bacteria, including Actinobacteria (Corynebacterium and Propionibacterium), represent precisely 53% of the ocular surface microbiome, followed by Proteobacteria (Pseudomonas and Acinetobacter) accounting for 39%, and Firmicutes (Staphylococcus and Streptococcus) for 8%." Bacteria belonging to the phylum Actinobacteria represent 53% of the core ocular surface microbiome but not the ocular surface microbiome in general. The relative abundances of the general ocular surface are for Proteobacteria 41% ± 0.16 for Actinobacteria 27% ± 0.19 and for Firmicutes and Bacteroides 17% ± 0.08 and 7% ± 0.06 respectively. In our paper, we defined the healthy core microbiome as genera present in a minimum of 5 out of the 11 published control cohorts with available raw data (retrieved from publications or provided by the corresponding author) with a relative abundance of at least 1%. This definition is based on earlier work in the gut but is nevertheless arbitrary. The proportion of bacteria in the "core" should therefore not be confused with actual abundances.