Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Mycobiome changes in the vitreous of post fever retinitis patients

  • Kotakonda Arunasri,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Validation, Writing – original draft

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Malleswarapu Mahesh,

    Roles Data curation, Methodology

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Gumpili Sai Prashanthi,

    Roles Data curation, Software

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Rajagopalaboopathi Jayasudha,

    Roles Data curation, Software

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Sama Kalyana Chakravarthy,

    Roles Data curation, Software

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Mudit Tyagi,

    Roles Resources

    Affiliation Smt. Kanuri Santhamma Centre for Vitreo Retinal Diseases, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Rajeev R. Pappuru,

    Roles Conceptualization, Resources

    Affiliation Smt. Kanuri Santhamma Centre for Vitreo Retinal Diseases, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India

  • Sisinthy Shivaji

    Roles Conceptualization, Funding acquisition, Supervision, Writing – review & editing

    Affiliation Jhaveri Microbiology Centre, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, L V Prasad Marg, Banjara Hills, Hyderabad, India


Fungi have been associated with various diseases of the eye like keratitis, uveitis and endophthalmitis. Despite this fact, fungal microbiome (mycobiome) studies compared to the bacterial microbiome studies have remained neglected. In the present study, using metagenomic sequencing, the mycobiomes of the vitreous of healthy control individuals (VC, n = 15) and individuals with post fever retinitis + non-PFR uveitis (PFR+, n = 9) were analysed and compared. The results indicated that Ascomycota was the most predominant phylum in both VC and PFR+ groups. Further, at the genera level it was observed that the abundance of 17 fungal genera were significantly different in post fever retinitis (PFR, n = 6) group compared to control group. Of these 17 genera, it was observed that 14 genera were relatively more abundant in PFR group and the remaining 3 genera in the VC group. Genus Saccharomyces, a commensal of the gut and skin, was predominantly present in the vitreous of both the cohorts, however it was significantly less abundant in PFR group. Further, significant increase in the genera that have a pathogenic interaction with the host were observed in PFR group. On the whole the mycobiome in both the groups differed significantly and formed two distinct clusters in the heatmap and Principal co-ordinate analysis. These results demonstrate significant changes in the mycobiome from the vitreous of post fever retinitis patients compared to healthy controls thus implying that dysbiotic changes in the fungal vitreous microbiome are associated with PFR.


Fungi are causative agents of several infections of the eye and range from infections of the ocular surface as in keratitis and could also cause intraocular infections as in endophthalmitis, uveitis and retinitis. In addition, fungi could infect eyelids, conjunctiva, and lacrimal system as in blepharitis, conjunctivitis and dacryoadenitis. Fungal infection of paranasal sinuses and other structures close to the orbit as in fungal sinusitis and acute invasive fungal rhinosinusitis was also reported [1, 2]. Among the regions of eye that the fungi infect, the ocular surface infections appear to be most prevalent. The fungi causing ocular surface infections include Lasiodiplodia theobromae, Candida spp., Fusarium, Alternaria, and Aspergillus [1, 3]. Several of these fungi also cause fungal endogenous endophthalmitis and include Cryptococcus neoformans, Sporotrichum schenckii, Histoplasma capsulatum, Blastomyces dermatitidis, and Coccidioides immitis in addition to Aspergillus and Fusarium [4]. Fugal retinitis is rare and the causative agents include Torulopsis glabrata [5], Pseudomyxoma peritonei [6] and Candida albicans [7].

Fungi, apart from their potential ocular pathogenicity are also associated with the healthy human eye [8, 9]. Culture-dependant methods showed that fungi could be detected on the ocular surface of healthy individuals in up to 28% of the individuals [8]. However, in a recent publication based on NGS technology fungi was identified on the ocular surface of 73.5% healthy individuals. This study also showed that genera Aspergillus, Setosphaeria, Malassezia, and Haematonectria were predominantly present in healthy eyes [9]. Interestingly all these genera are also known to be associated with ocular fungal infections [3] implying that these are opportunistic pathogens. In a subsequent study based on the abundance analysis of the fungal genera in the conjunctival swabs and corneal scrapings of fungal keratitis individuals indicated that 11 core genera namely Aspergillus, Setosphaeria, Malassezia, Haematonectria, Candida, Emericella, Penicillium, Fusarium, Cladosporium, Choiromyces, and Cochliobolus showed differential abundance in the keratitis patient eyes compared to healthy conjunctival swabs [10]. On a similar note in culture negative endophthalmitis patient samples, four genera Aspergillus, Fusarium, Exserohilum and Candida were observed to be predominantly present in 73.7% of the samples analysed [11]. In the present study the fungal diversity and abundance was assessed by metagenomic sequencing of the vitreous inpost fever retinitis individuals compared to the vitreous collected from individuals undergoing ocular surgery who had no ocular and systemic infection. Post fever retinitis (PFR) or retinitis post febrile illness is a form of retinitis that prevails mostly in the Indian subcontinent [12, 13]. Clinical manifestation of PFR happens after 2 to 4 weeks of post febrile illness and occurs mostly after bacterial [14], viral [15, 16] and protozoan [17] epidemics. A recent study also reported a case of fungal retinitis following a viral systemic illness in an immune competent individual [7]. Although the clinical impression of PFR and retinitis are well documented [13], changes in the intraocular microbial composition remained unknown. Thus the present study adopted next generation sequencing (NGS) approach to unravel the fungal microbiome (mycobiome) in retinitis patients compared to healthy individuals. The rationale of the present study was to ascertain whether mycobiome composition in the vitreous body was altered in PFR individuals. Our results indicated significant differences in the mycobiome in the vitreous of post PFR patients compared to healthy controls thus implying that dysbiotic changes in the fungal vitreous microbiome are associated with PFR.

Materials and methods

Study group and sample collection

Collection of vitreous from healthy controls would not be ethical. Therefore in this cohort we included individuals who were to undergo ocular procedures like Macular hole surgery or Rhegmatogenous retinal detachment from whom vitreous could be collected. These controls were not symptomatic for PFR and had no other ocular infections. All the participants with a past history of Typhoid, Malaria and Chikungunya and febrile illness were excluded from the study. All participants who had history of inflammatory disorders of the eye, uncontrolled glaucoma, hypertension and diabetes were excluded from the study. A total of 15 individuals were recruited as controls and 300 μl of vitreous was collected from each individual.

PFR group included individuals with PFR which normally manifested 2 to 4 weeks after systemic infections like Typhoid [14], Malaria and Chikungunya [17] caused due to a wide array of infectious agents such as bacteria, parasites and viruses [18, 19]. All individuals who presented with features of retinitis (please see introduction) and who had a past history of fever were considered to be having post fever retinitis. All patients with HIV, Tuberculosis and Syphilis were excluded from the study. Individuals with history of fever and retinitis due to Behcet's disease and systemic lupus erythematosus [20] were also eliminated. PFR individuals who had inflammatory disorders of the eye, uncontrolled glaucoma, hypertension and diabetes were also excluded from the study. In both control and PFR groups any evidenceof endophthalmitis or anyother ocular infection in either of the eyes or current treatment for serious systemic infection at the time of examination were excluded. Vitreous (300 μl) was collected from the PFR individuals (n = 6) including 3 individuals with non-PFR uveitis who were scheduled for pars plana vitrectomy/vitreous biopsy which is the normal procedure followed during the treatment of PFR (S1 Table). The treatment of the patients was symptomatic and was managed by administering oral antivirals, antibiotics, and oral steroids. All the patients were subjected to a vitreous biopsy which was performed in the operating rooms under full aseptic conditions by trained vitreoretinal surgeons. There was no difference in preparation for PPV and Vitreous Biopsy. The vitreous samples collected from the two study groups was stored at -80°C until used. This study was approved by the Institutional review board (LV Prasad Eye Institute, Hyderabad) for 2 years (from 1st September 2017 to 31st August 2019) with the ethics reference number LEC 09-17-079. All the samples were collected during the study period and written consent was obtained from the study participants. The study was designed and conducted according to the tenets of the Declaration of Helsinki.

DNA extraction and metagenome sequencing

DNA was extracted from about 200 μl of vitreous sample by using PureLink DNA extraction kit (ThermoFisher Scientific, Mumbai India) according to the manufacturer’s protocol. The extracted DNA was quantified using Qubit fluorometer (Thermo Fisher Scientific, Carlsbad, CA, USA). The extracted nucleic acids were amplified with random hexamers using amplification kit (SeqPlex, Sigma Aldrich Chemicals Private Limited, Bengaluru, India). For Library preparation and sequencing, NEBNext Ultra DNA Library Prep Kit for Illumina Nextseq 500 PE sequencing protocol was followed using paired-end sequencing with 2x150bp chemistry on the Illumina Nextseq 500 platform. Care was taken to avoid microbial contamination from the environment by carrying out all the steps such as sample preparation, DNA extraction, PCR and whole genome amplification procedures in a dedicated BSL2 laminar flow hood. The laboratory where the experiments were conducted was UV sterilised prior to use. The equipment used was surface sterilised and the water used as control was autoclaved twice. Sterile water was used as a negative control instead of template DNA in PCR and consistently amplification was negative implying lack of contaminating DNA. No sequences could be generated from the negative controls.

Fungal metagenome analysis

FASTQ files of the raw reads were generated for all the 24 samples that were sequenced. These raw sequence reads were analysed for quality parameters such as read length, phred quality score (<25), GC content and presence of ambiguous bases. The sequencing adapters from the raw sequences were trimmed by using trim-galore (version 0.4.0; Felix Krueger) and Cutadapt version 1.2 [21]. Subsequently, all reads were subjected to FastQC (version 0.11.3) tool which helped in identifying reads with quality score greater than Q25. Human genome sequences were removed by using decontam ( and Bowtie 2 tools using NCBI-GRCh38 as reference genome. Reads were assigned to fungal taxa using a taxonomic classifier ‘What’s in my Pot’ (WIMP) [22]. For this the trimmed, filtered, FASTQ reads were subjected to the workflow of EPI2ME in the Oxford Nanopore Technologies [23]. Post taxonomy assignment, Meta Genome Analyser (MEGAN) v 5.11.3 was used for comparative analysis of the samples and generation of the fungal microbiome biome file for the 24 samples. The biome data file comprising abundance of all samples at genera level is provided in S2 Table. The metagenomic sequencing reads were deposited with National Center for Biotechnology Information (NCBI) and are accessible in a BioProject with accession number PRJNA646315.

Statistical analysis

The vegan package in R ( was used to generate rarefaction curves and for quantifying diversity indices. Genera with a mean abundance >0.002% were used for the analysis. Alpha diversity indices viz., Shannon diversity, Simpson index and Observed number of genera were calculated and the degree of variation in the fungal diversity of VC and PFR group was ascertained. Unpaired t test was conducted by using GraphPad Prism to determine statistical significance of the alpha diversity indices ( Significantly different taxa in both VC and PFR groups at phylum and genus level were identified by a non-parametric, Wilcoxon signed rank test (with P<0.05 as significant). For this analysis only 6 patient samples of PFR group (PFR01 to PFR06) (S1 Table) that were clinically identified as post fever retinitis were included. These samples also tested PCR negative for the viral etiology (HSV, CMV and VZV). For visualising the relative abundances based clustering of the genera, a rank-sum normalised heatmap was generated for the fungal genera of both the cohorts. Principal co-ordinate analysis (PCoA) plot was generated for the 24 fungal microbiomes using ade4 package in R (v3.2.5) by employing Jensen–Shannon divergence distance metric. K-means clustering (k = 2) was employed to identify VC and PFR clusters on the PCoA plot.

Correlation Network analysis of fungal genera

CoNet [24] is a Cytoscape [25] plugin that was used to detect interactive networks of the fungal microbiome in VC and PFR+ groups independently. Spearman correlation coefficient (r) was used to analyse the interactions among the different fungi (mutual exclusions/negative and co-presence/positive interactions) at the genus level.


Total reads and taxonomic assignment of mycobiomes

A total of 105.5 million reads were generated for all the 24 vitreous samples of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups. Reads assigned to fungi comprised a total of 6.3127 million reads with 359506 and 102233 reads assigned to fungi in the control and PFR+ groups respectively (Table 1). Rarefaction curves plotted for all the 24 mycobiomes showed a tendency towards saturation indicating that majority of the fungal diversity was identified at the genera level (S1 Fig). Alphadiversity analysis showed significant difference in Simpson and Shannon diversity indices (Fig 1). While the Observed number of Genera were not statistically different between VC and PFR+ groups.

Fig 1. Alpha diversity indices of mycobiomes in the vitreous fluid of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups.

Median values (horizontal line) and interquartile ranges of the diversity indices are depicted in the box plots.

Table 1. Number of fungal reads generated from vitreous fluid of control (VC, n = 15) and post fever retinitis+ non-PFR uveitis (PFR+, n = 9) groups.

Relative abundance at the phyla and genera level

The relative abundance of the fungal taxa in control and post fever retinitis groups are summarised at both the phyla and genera level. At phyla level, Ascomycota, Basidiomycota, Chytridiomycota and Microsporidia were detected in both control (VC) and in post fever retinitis + non-PFR uveitis (PFR+) groups (Table 2). Reads assigned to phylum Ascomycota were predominantly present in both control and post fever retinitis groups with a mean abundance of 96.2% (range 76.9 to 99.3%) and 93.06% (range 60.25 to 99.6%) in the VC and PFR cohorts. These changes at phyla level in both VC and PFR group were not significantly different (Fig 2A and 2B). Further, the ratio of Basidiomycota to Ascomycota in VC and PFR+ is 3.4 and 6.0 respectively. At the genera level, the number of genera detected in VC and PFR+ were similar with 134 and 133 genera respectively. In addition it was observed that the most abundant top 10 genera Saccharomyces, Malassezia, Colletotrichum, Aspergillus, Lobosporangium, Paracoccidiodes, Exserohilum, Metacordyceps and Talaromyces were shared between all the 24 vitreous samples analysed (S3 Table; Fig 2C). Fungi incertae sedis that comprise fungi that could not be classified, undefined or unknown accounted for a mean abundance of 0.47% in the control and 1.27% in the PFR+ groups.

Fig 2.

Relative abundance of different fungal phyla (A and B) and fungal genera (C) in the vitreous fluid of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups. The yellow line in ‘C’ separates VC and PFR groups. Median values (horizontal line) and interquartile ranges have been depicted in the boxplots of ‘B’.

Table 2. Relative abundance of fungal phyla in the vitreous of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups.

Differentially abundant fungal genera

Wilcoxon signed rank test was performed between PFR group (n = 6) and control (VC) group (n = 15) to measure the significant differences in abundance at genera level (P<0.05). This resulted in identifying seventeen fungal genera that were significantly differentially abundant in post fever retinitis group compared to control group (Table 3). Out of the 17 differentially abundant genera, 14 genera were increased in abundance in PFR group and 3 genera were increased in VC group (Fig 3). Significant differences in the relative abundance of the 17 fungal genera in VC and PFR groups was also clearly visualised using boxplots (S2 Fig). Discriminative genera like Paracoccidioides, Saccharomyces, Trichoderma and Kluyveromyces were present in all the samples of VC and PFR groups. Further genera Nectria and Nematocida were present only in VC group. At the same time when all the 9 samples of PFR group were included for the analysis, 18 discriminative genera were identified (S4 Table). 13 out of 18 discriminative genera viz., Setosphaeria, Arthroderma, Isaria, Paracoccidiodes, Sordaria, Nectria, Saccharomyces, Exserohilum, Nematocida, Microsporum, Trichoderma, Pseudogymnoascus and Kluveromyces were common to the discriminative genera identified using only the 6 post fever retinitis patient samples.

Fig 3. Discriminating fungal genera exhibiting significant (P<0.05) differential abundance in the vitreous fluid of control (VC, n = 15) and post fever retinitis (PFR, n = 6) groups.

The horizontal bars represent the statistically significant genera as Log2 fold change. Purple and blue colour bars indicate increase in the relative abundance of the fungal genera in VC group and PFR group respectively. Genera Nectria and Nematocida were present only in VC group.

Table 3. Discriminative fungal genera in the vitreous fluid of control (VC) and post fever retinitis (PFR) groups (P<0.05).

Rank normalised hierarchal clustering based heatmap analysis was done for all the 24 samples using the 17 discriminative genera (Table 3; Fig 4A). Heatmap showed two distinct clusters differentiating the control (VC) and post fever retinitis + non-PFR uveitis (PFR+) groups. However one sample from control group (VC16), clustered with PFR+ group while a sample from PFR+ group (PFR06) clustered with control group (Fig 4A). Using Jensen-Shannon Divergence (JSD) approach principal coordinate analysis was performed for all the 24 samples of control (VC) and PFR+ groups. This showed two distinct clusters for both control (VC) and post fever retinitis + non-PFR uveitis (PFR) groups (Fig 4B). All the nine samples from post fever retinitis group + non-PFR uveitis formed a separate cluster from the 15 control samples suggesting the divergence in their mycobiomes. It is worthwhile to mention that the PFR+ group also included 3 individuals viz., PFR07, PFR08 and PFR09 who were identified as having non-PFR uveitis but nevertheless in the Principal co-ordinate analysis they formed a single group implying that the mycobiomes of PFR and non-PFR uveitis are not different.

Fig 4. Heat map of discriminative fungal genera in the mycobiomes of vitreous fluid of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups (P<0.05).

A. Rank normalised abundances of the genera were scaled from 0 to 1 that are depicted by red and green colour respectively in the heat map. B. Principal co-ordinate analysis of fungal genera between control group (VC) and post fever retinitis + non-PFR uveitis (PFR+) group.

Network analysis

Fungal-fungal interactions within the control group and PFR+ group were inferred through Cytoscape generated CoNet interactions. A total of 69 of the 134 genera showed fungal-fungal interactions in the control group. Eleven genera viz., Saccharomyces, Talaromyces, Valvraia, Marssonina, Puccinia, Kluyveromyces, Xylona, Histoplasma, Leptosphaeria, Metarhizium and Thielavia (S4 Table) showed >10 interactions. Two major network hubs were formed by Saccharomyces and Talaromyces which showed 49 and 43 negative interactions respectively (S5A Table; Fig 5A). Positive interactions were absent for these genera. On the contrary from PFR+ group, 124 genera out of 133 genera showed more than 1 interaction with other genera (S5B Table and Fig 5B). However, only 4 genera viz., Torulaspora, Cordyceps, Isaria and Mixia showed more than 10 interactions with other genera. Genus Torulaspora forms the largest cluster in PFR+ group which had 31 negative interactions with other fungal genera.

Fig 5.

CoNet network analysis of fungal genera that co-occur in the mycobiomes of vitreous fluid of control (VC, n = 15) (A) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups (B) (P<0.05). Nodes indicate the fungal genera. Green colour edges indicate positive interaction and red colour edges indicate negative interaction.


Alterations in the bacterial microbiome have been implicated in human health and disease. However, similar studies implicating fungal microbiomes (mycobiomes) in health and disease are limited. Taking cue from theses bacterial microbiome studies there has been an increased interest in studying mycobiomes in different niches of the human body such as skin [26], oral cavity [27], gut [2831] and ocular surface [9] in healthy individuals and in individuals with diseases like urinary tract diseases [32], Inflammatory bowel disease [33], respiratory diseases [34], ocular diseases like keratitis and uveitis [29, 35] etc. These studies showed dysbiotic changes in the mycobiomes under diseased conditions and also revealed that certain fungal species exist at low levels without causing disease but may become pathogenic under immune-compromised conditions [36]. Compared to the above human niches that have been sampled for fungi, studies on the eye are rare [37, 38] and even more rare are studies analysing the mycobiome of the vitreous fluid of the eye [1, 11, 39]. Further, detection of fungi in intraocular fluids secondary to systemic infection is extremely rare and the causative agents are often difficult to cultivate. Therefore, in the present study whole metagenomic sequencing approach was employed to demonstrate the mycobiome in 24 vitreous samples that included samples from healthy controls (VC, n = 15) and individuals with post fever retinitis + non-PFR uveitis (PFR+, n = 9).

Our results indicated that fungi affiliated to the phyla Ascomycota and Basidiomycota were the predominant phyla in the vitreous fluid of the healthy controls and PFR+ group. (Table 2 and Fig 2). That these two phyla were also reported as being predominant on the ocular surface [9, 10] and in the vitreous fluid of endophthalmitis patients [11]. We also observed that the ratio of the phyla Basidiomycota to Ascomycota increased in PFR+ group compared to VC group. This observation may be relevant considering that earlier studies had indicated that in Inflammatory bowel disease, increase in the ratio of Basidiomycota to Ascomycota is critical for the disease pathogenesis [33]. At the genera level, the top 10 genera viz., Saccharomyces, Malassezia, Colletotrichum, Aspergillus, Paracoccidiodes, Exserohilum, Lobosporangium, Metacordyceps and Talaromyces were assigned to 85% of the total fungal reads. It is difficult to ascertain whether this abundance status is in way related to the health status of the individual. But, some interpretation could be made based on changes in the abundance with the diseased state. The mean abundance of the genus Saccharomyces, a gut commensal [40] was significantly higher in control group compared to PFR group. This may be of an advantage to the healthy control group since the genus Saccharomyces is known to positively influence the immune system to better cope with secondary infection [41]. At the same time it was also observed that the abundance of the potential human pathogens viz., Colletotrichum [42], Aspergillus [1], Paracoccidiodes [43], Exserohilum [44], were high in PFR group compared to control (S3 Table). The analysis also indicated that VC and PFR groups could be discriminated based on 17 different genera and interestingly the functional attributes of these discriminative genera vis a vis the host could be deduced from literature (Table 4). Pathogens such as Setosphaeria, Arthroderma, Paracoccidiodes, Exserohilum, Microsporum, Trichoderma, Pseudogymnoascus and Kluyveromyces were all increased in PFR group compared to control group (Table 4; Fig 3). In addition a few non- pathogenic or commensal fungi such as Isaria, Sordaria, Komagataella and Fomitiporia were also increased in PFR group (Table 4; Fig 3). Among the above pathogens Setopsphaeria, Trichoderma and Kluyveromyces were increased in patients having fungal keratitis [10] (S6 Table). Earlier studies had also identified Setosphaeria, Exserohilum and Trichoderma in the vitreous fluid in endophthalmitis patients [11]. Besides the above pathogens, Fusarium and Candida the common ocular pathogens were also identified in the vitreous fluid of VC and PFR cohorts but their abundance in the two groups were not significantly different. This study demonstrates differences in the relative abundance of several fungi in the vitreous fluid of VC and PFR group. Based on these discriminative fungi VC and PFR mycobiomes could be separated into two distinct clusters both by heatmap and principal co-ordinate analysis implying that the vitreous fluid mycobiomes in PFR group is distinct from the control (VC) group. Although, the individuals in the PFR group had no clinical history of fungal infection, increase in the mean abundance of fungal genera in PFR group is difficult to explain.

Table 4. Attributes of the discriminative fungal genera in the vitreous of Control (VC, n = 15) and post fever retinitis (PFR, n = 6) mycobiomes.

CoNet network analysis of the vitreous fluid mycobiomes also indicated that the mycobiomes based on their interaction pattern are very distinct in the VC and post fever retinitis + non-PFR uveitis (PFR+) groups. For instance, in the VC group 42.2% of the total fungal genera showed fungal-fungal interactions compared to the 89.2% of the genera in PFR+ group implying that the PFR+ group is more interactive. Further, genera that exhibited interactions with more than 10 fungal genera were identified as hub genera and in VC and PFR+ groups the number of hubs identified include 11 and 4 respectively. Out of the 11 hub genera in VC group, two hubs were formed by discriminative genera such as Kluyveromyces and Saccharomyces. Kluyveromyces is an opportunistic pathogen [56], that showed 10 positive interactions and 2 negative interactions with other fungal genera (S5A Table). Further, Kluyveromyces exhibited 4 positive interactions with human pathogenic fungi and a negative interaction with genus Saccharomyces. Besides, Saccharomyces is a commensal organism [40] that formed the major hub in control group and interacted negatively with 49 other fungal genera out of which 24 are known human pathogens. Thus it would mean that in the healthy control group Saccharomyces may confer beneficial effects. In contrast in PFR+ group Saccharomyces negatively interacted only with 4 human pathogenic fungi, thus implying that the beneficial effects that were conferred in VC group may not exist in PFR+ group. In addition genus Isaria is the only discriminative genera that formed a hub in PFR+ group. Genus Isaria is a non pathogenic fungi that exhibited 5 positive and one negative interaction with human pathogenic fungi. On the whole it appears that the networks generated for PFR+ group lack the prominent hubs that could negate the pathogenic fungal interactions as observed in VC group.

Increasing evidences suggest the importance of microbiome studies as a potential modifiable factor in therapeutics [57, 58]. The results of the present study may not be directly impacting the clinical outcome. However the observations of the study reveal that there was an increase in the abundance of pathogenic fungi in patient samples compared to the control. Conversely, the abundance of few fungal genera that may have beneficiary effect were decreased in the patient cohort. Another important observation that could be made from the network analysis is that microbial interactions are crucial in maintaining the homeostasis and this was disrupted under the diseased condition. All these evidences put together imply that the changes in the mycobiome composition in retinitis individuals were significant but it would be difficult to assess its impact on the treatment outcomes from the available data.

The study involving generation of mycobiomes from the vitreous samples has a few limitations and biases. One of the limitations is the lack of vitreous sample from healthy individuals, as controls, for want of ethical clearance. The study has less number of individuals recruited in PFR group due to rareness of the conditions and due to ethical compliance for vitreous biopsy. Further, the collection of vitreous is an invasive process involving the use of a syringe to pierce the ocular surface, access the intraocular space and retrieve the vitreous. This procedure though accomplished in an operation theatre does not ensure possible contamination by the ocular surface mycobiome. Therefore analysis of the ocular surface mycobiome from all the participants may be required for comparison with vitreous. Since in the present study the ocular surface samples were not collected, the data of the present study was compared with the ocular surface mycobiome from published literature [911]. The mycobiome from blood in control and PFR group participants could have enhanced the understanding of the mycobiome penetration into the eye. Simultaneously mycobiome from the individuals with fungal retinitis would have helped in comparing the data to understand the changes in the vitreous due to fungal pathogens.


The findings of the study point towards an imbalance in the vitreous mycobiome in PFR group compared to control group. Alpha diversity indices (Simpson and Shannon diversity indices) between VC and PFR+ group were significantly different. VC and PFR+ cohorts could also be discriminated based on the ratio of Basidiomycota to Ascomycota and the abundance of several genera. Significant decrease in the relative abundance of predominant commensal genus Saccharomyces and increase in the abundance of 9 fungal human pathogens viz., Setosphaeria, Arthroderma, Enterocytozoon, Exserohilum, Paracoccidiodes, Pseudogymnoascus, Trichoderma, Kluveromyces and Microsporum signifies the possible role of these genera in the pathogenesis of PFR. Thus the study provides evidence that the mycobiome of the vitreous fluid in post fever retinitis individuals is distinct compared to the healthy controls.

Supporting information

S1 Fig. Rarefaction curves of mycobiomes from vitreous fluid of controls ((VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups.

Curves were plotted for the genera having a mean abundance > 0.002%.


S2 Fig. Boxplot histograms of 17 discriminative fungal genera in the mycobiomes of vitreous fluid of control (VC, n = 15) and post fever retinitis + non-PFR uveitis (PFR+, n = 9) groups (P<0.05).

Median values (horizontal line) and interquartile ranges have been depicted in the plots.


S1 Table. Sample collection details of control samples (VC, n = 15) and post fever retinitis+ non-PFR uveitis samples (PFR+, n = 9).


S2 Table. Mycobiome biome file of control samples (VC, n = 15) and post fever retinitis samples + non-PFR uveitis (PFR+, n = 9).


S3 Table. Abundance of fungal genera in the vitreous of Control (VC) and post fever retinitis + non-PFR uveitis (PFR+) samples.

Genera having a mean abundance of >0.002% are listed in the table.


S4 Table. Discriminative fungal genera in the vitreous fluid of control (VC, n = 15) and post fever retinitis (PFR, n = 6) groups (P<0.05).


S5 Table.

A. Co-occurrence network analysis of control (VC) group determining the number of positive and negative interactions among the fungal genera. B. Co-occurrence network analysis of PFR+ group determining the number of positive and negative interactions among the fungal genera.


S6 Table. Presence or absence of discriminative genera of the present study in other ocular studies.



Our thanks to the Director, Prof Brien Holden Eye Research Centre, L V Prasad Eye Institute, Hyderabad for facilities.


  1. 1. Klotz SA, Penn CC, Negvesky GJ, Butrus SI. Fungal and Parasitic Infections of the Eye. Clinical Microbiology Reviews. 2000;13(4):662–85. pmid:11023963
  2. 2. Aribandi M, McCoy VA, Bazan C. Imaging Features of Invasive and Noninvasive Fungal Sinusitis: A Review. RadioGraphics. 2007;27(5):1283–96. pmid:17848691
  3. 3. Liesegang TJ, Forster RK. Spectrum of Microbial Keratitis in South Florida. American Journal of Ophthalmology. 1980;90(1):38–47. pmid:7395957
  4. 4. Patel AS, Hemady RK, Rodrigues M, Rajagopalan S, Elman MJ. Endogenous Fusarium Endophthalmitis in a Patient With Acute Lymphocytic Leukemia. American Journal of Ophthalmology. 1994;117(3):363–8. pmid:8129012
  5. 5. Fitzsimmons RB, Nicholls MD, Billson FA, Robertson TI and Hersey P. Fungal retinitis: A case of Torulopsis glabrata infection treated with miconazole. Br J Ophthalmol 1980;64:672–5. pmid:7191722
  6. 6. Mukherjee S, Weir C, Hammer H. Presumed fungal retinitis associated with Pseudomyxoma peritonei. Case Reports. 2009;2009(mar25 1):bcr0620080345-bcr. pmid:21686849
  7. 7. Menia N, Sharma S, Bansal R. Fungal retinitis following influenza virus type A (H1N1) infection. Indian journal of ophthalmology. 2019;67(9):1483. pmid:31436207
  8. 8. Wilson LA, Ahearn DG, Jones DB, Sexton RR. Fungi from the Normal Outer Eye. American Journal of Ophthalmology. 1969;67(1):52–6. pmid:5782862
  9. 9. Shivaji S, Jayasudha R, Sai Prashanthi G, Kalyana Chakravarthy S, Sharma S. The Human Ocular Surface Fungal Microbiome. Invest Ophthalmol Vis Sci. 2019;60(1):451–9. pmid:30703210
  10. 10. Prashanthi GS, Jayasudha R, Chakravarthy SK, Padakandla SR, SaiAbhilash CR, Sharma S, et al. Alterations in the Ocular Surface Fungal Microbiome in Fungal Keratitis Patients. Microorganisms. 2019;7(9):309.
  11. 11. Gandhi J, Jayasudha R, Naik P, Sharma S, Dave VP, Joseph J. Targeted High-Throughput Sequencing Identifies Predominantly Fungal Pathogens in Patients with Clinically Infectious, Culture-Negative Endophthalmitis in South India. Microorganisms. 2019;7(10):411. pmid:31581465
  12. 12. Achermann Y, Goldstein EJ, Coenye T, Shirtliff ME. Propionibacterium acnes: from commensal to opportunistic biofilm-associated implant pathogen. Clin Microbiol Rev. 2014;27(3):419–40. pmid:24982315
  13. 13. Kawali A, Mahendradas P, Mohan A, Mallavarapu M, Shetty B. Epidemic Retinitis. Ocul Immunol Inflamm. 2019;27(4):571–7. pmid:29370564
  14. 14. Prabhushanker M, Topiwalla TT, Ganesan G, Appandaraj S. Bilateral retinitis following typhoid fever. International Journal of Retina and Vitreous. 2017;3(1). pmid:28405476
  15. 15. Unnikrishnan SR, K V. Ocular manifestations in mosquito-borne viral fevers. J Clin Ophthalmol Res 2017;5(2):77–80.
  16. 16. Lalitha P, Rathinam S, Banushree K, Maheshkumar S, Vijayakumar R, Sathe P. Ocular involvement associated with an epidemic outbreak of chikungunya virus infection. Am J Ophthalmol. 2007;144(4):552–6. pmid:17692276
  17. 17. Vishwanath S, Badami K, Sriprakash KS, Sujatha BL, Shashidhar SD, Shilpa YD. Post-fever retinitis: a single center experience from south India. Int Ophthalmol. 2014;34(4):851–7. pmid:24362635
  18. 18. Ahuja Y, Couch S. M., Razonable R. R., and Bakri S. J. Infectious retinitis: A review. Retin Physician 2008. Nov-Dec:17–25.
  19. 19. Panchal B. Commentary: Post fever retinitis and vasculitis: A morphological conundrum. Indian journal of ophthalmology. 2018;66(12):1844. pmid:30451193
  20. 20. Palkar AH, Hossain MD, Majumder PD. Purtscher-like retinopathy as presenting manifestation of systemic lupus erythromatosus following high-grade fever with myalgia and arthralgia. Indian journal of ophthalmology. 2018;66(9):1317–9. pmid:30127152
  21. 21. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMB net journal. 2011;17:10–2.
  22. 22. Juul S, Izquierdo F, Hurst A, Dai X, Wright A, Kulesha E, et al. What’s in my pot? Real-time species identification on the MinION. bioRxiv. 2015.
  23. 23. Kim D, Song L, Breitwieser FP, Salzberg SL. Centrifuge: rapid and sensitive classification of metagenomic sequences. Genome Research. 2016;26(12):1721–9. pmid:27852649
  24. 24. Faust K, Raes J. CoNet app: inference of biological association networks using Cytoscape. F1000Res. 2016;5:1519. pmid:27853510
  25. 25. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. pmid:14597658
  26. 26. Grice EA, Segre JA. The skin microbiome. Nature Reviews Microbiology. 2011;9(4):244–53.
  27. 27. Dupuy AK, Andes DR, David MS, Li L, Heider TN, Peterson JD, et al. Redefining the Human Oral Mycobiome with Improved Practices in Amplicon-based Taxonomy: Discovery of Malassezia as a Prominent Commensal. PLoS ONE. 2014;9(3):e90899. pmid:24614173
  28. 28. Sam Q, Chang M, Chai L. The Fungal Mycobiome and Its Interaction with Gut Bacteria in the Host. International Journal of Molecular Sciences. 2017;18(2):330. pmid:28165395
  29. 29. Jayasudha R, Chakravarthy SK, Prashanthi GS, Sharma S, Garg P, Murthy SI, et al. Alterations in gut bacterial and fungal microbiomes are associated with bacterial Keratitis, an inflammatory disease of the human eye. J Biosci. 2018;43(5):835–56. pmid:30541945
  30. 30. Kalyana Chakravarthy S, Jayasudha R, Ranjith K, Dutta A, Pinna NK, Mande SS, et al. Alterations in the gut bacterial microbiome in fungal Keratitis patients. PLoS ONE 2018a;13:e0199640. pmid:29933394
  31. 31. Bacher P, Hohnstein T, Beerbaum E, Röcker M, Blango MG, Kaufmann S, et al. Human Anti-fungal Th17 Immunity and Pathology Rely on Cross-Reactivity against Candida albicans. Cell. 2019;176(6):1340–55.e15. pmid:30799037
  32. 32. Ackerman AL, Underhill DM. The mycobiome of the human urinary tract: potential roles for fungi in urology. Annals of Translational Medicine. 2017;5:31. pmid:28217696
  33. 33. Sokol H, Leducq V, Aschard H, Pham H-P, Jegou S, Landman C, et al. Fungal microbiota dysbiosis in IBD. Gut. 2017;66(6):1039–48. pmid:26843508
  34. 34. Weaver D, Gago S, Bromley M, Bowyer P. The Human Lung Mycobiome in Chronic Respiratory Disease: Limitations of Methods and Our Current Understanding. Current Fungal Infection Reports. 2019;13(3):109–19.
  35. 35. Kalyana Chakravarthy S, Jayasudha R, Sai Prashanthi G, Ali MH, Sharma S, Tyagi M, et al. Dysbiosis in the Gut Bacterial Microbiome of Patients with Uveitis, an Inflammatory Disease of the Eye. Indian J Microbiol. 2018b;58(4):457–69. pmid:30262956
  36. 36. Huffnagle GB, Noverr MC. The emerging world of the fungal microbiome. Trends in Microbiology. 2013;21(7):334–41. pmid:23685069
  37. 37. Borroni D, Romano V, Kaye SB, Somerville T, Napoli L, Fasolo A, et al. Metagenomics in ophthalmology: current findings and future prospectives. BMJ Open Ophthalmol. 2019;4(1):e000248. pmid:31276030
  38. 38. Hutchinson DS, Pflugfelder SC, Petrosino JF. Microbiome, Eye. Springer US. 2015:400–3.
  39. 39. Vilela RC, Vilela L, Vilela P, Vilela R, Motta R, Pôssa AP, et al. Etiological agents of fungal endophthalmitis: diagnosis and management. International Ophthalmology. 2014; 34(3):707–21. pmid:24081913
  40. 40. Limon JJ, Skalski JH, Underhill DM. Commensal Fungi in Health and Disease. Cell Host Microbe. 2017;9:156–65. pmid:28799901
  41. 41. Rizzetto L, De Filippo C, Cavalieri D. Richness and diversity of mammalian fungal communities shape innate and adaptive immunity in health and disease. European Journal of Immunology. 2014;44(11):3166–81. pmid:25257052
  42. 42. Lin L, Yang C, Wan J, Chang TC, Lee JY. Cutaneous Infection Caused by Plant Pathogen Colletotrichum gloeosporioides. JAMA Dermatol. 2015;151:1383–4. pmid:26287364
  43. 43. de Oliveira HC, da Silva JdFt, Scorzoni L, Marcos CM, Rossi SA, de Paula e Silva ACA, et al. Importance of adhesins in virulence of Paracoccidioides spp. Frontiers in Microbiology. 2015;6:303. pmid:25914695
  44. 44. Jayasudha R, Sharma S, Kalra P, Mishra D. Exserohilum keratitis: Clinical profile of nine patients and comparison of morphology versus ITS-Based DNA sequencing for species identification of the fungal isolates. Indian Journal of Medical Microbiology. 2018;36(4):564. pmid:30880707
  45. 45. Katragkou A, Pana ZD, Perlin DS, Kontoyiannis DP, Walsh TJ, Roilides E. Exserohilum infections: Review of 48 cases before the 2012 United States outbreak. Medical Mycology. 2014;52(4):376–86. pmid:24682112
  46. 46. Zhang H, Ran Y, Liu Y, Zhang R, Lin X, Yan W, et al. Arthroderma vanbreuseghemii infection in three family members with kerion and tinea corporis. Medical Mycology. 2009;47(5):539–44. pmid:19115135
  47. 47. Finamor LP, Muccioli C, Martins MC, Rizzo LV, Belfort R. Ocular and central nervous system paracoccidioidomycosis in a pregnant woman with acquired immunodeficiency syndrome. American Journal of Ophthalmology. 2002;134(3):456–9. pmid:12208267
  48. 48. F. Abdallah M, De Boevre M, Landschoot S, De Saeger S, Haesaert G, Audenaert K. Fungal Endophytes Control Fusarium graminearum and Reduce Trichothecenes and Zearalenone in Maize. Toxins. 2018;10(12):493. pmid:30477214
  49. 49. Coleman JJ, Rounsley SD, Rodriguez-Carres M, Kuo A, Wasmann CC, Grimwood J, et al. The Genome of Nectria haematococca: Contribution of Supernumerary Chromosomes to Gene Expansion. PLoS Genetics. 2009;5(8):e1000618. pmid:19714214
  50. 50. Munoz P, Bouza E, Cuenca-Estrella M, Eiros JM, Perez MJ, Sanchez-Somolinos M, et al. Saccharomyces cerevisiae Fungemia: An Emerging Infectious Disease. Clinical Infectious Diseases. 2005;40(11):1625–34. pmid:15889360
  51. 51. Deng L, Li W, Yu X, Gong C, Liu X, Zhong Z, et al. Correction: First Report of the Human-Pathogenic Enterocytozoon bieneusi from Red-Bellied Tree Squirrels (Callosciurus erythraeus) in Sichuan, China. PLoS One. 2016;11(12):e0168631. pmid:27992597
  52. 52. Sandoval-Denis M, Sutton DA, Martin-Vicente A, Cano-Lira JF, Wiederhold N, Guarro J, et al. Cladosporium Species Recovered from Clinical Samples in the United States. J Clin Microbiol. 2015;53(9):2990–3000. pmid:26179305
  53. 53. Sacheli R, Adjetey C, Darfouf R, Harag S, Huynen P, Meex C, et al. A one-year survey of Microsporum audouinii infections in Belgium: epidemiological and genotypic characterization. Clinical Microbiology and Infection. 2016;22(3):285.e9–.e17. pmid:26686810
  54. 54. Chouaki T, Lavarde V, Lachaud L, Raccurt CP, Hennequin C. Invasive Infections due to Trichoderma Species: Report of 2 Cases, Findings of In Vitro Susceptibility Testing, and Review of the Literature. Clinical Infectious Diseases. 2002;35(11):1360–7. pmid:12439799
  55. 55. Fuchs BB, Chaturvedi S, Rossoni RD, de Barros PP, Torres-Velez F, Mylonakis E, et al. Galleria mellonella experimental model for bat fungal pathogen Pseudogymnoascus destructans and human fungal pathogen Pseudogymnoascus pannorum. Virulence. 2018;9(1):1539–47. pmid:30289352
  56. 56. Taj-Aldeen SJ, AbdulWahab A, Kolecka A, Deshmukh A, Meis JF, Boekhout T. Uncommon opportunistic yeast bloodstream infections from Qatar. Medical Mycology. 2014;52(5):552–6. pmid:24934803
  57. 57. Behrouzi A, Nafari AH, Siadat SD. The significance of microbiome in personalized medicine. Clin Transl Med. 2019;8(1):16. pmid:31081530
  58. 58. Kashyap PC, Chia N, Nelson H, Segal E, Elinav E. Microbiome at the Frontier of Personalized Medicine. Mayo Clin Proc. 2017;92(12):1855–64. pmid:29202942