Big Data for a Small World: A Review on Databases and Resources for Studying Microbiomes

Abstract

Microorganisms are ubiquitous in nature and form complex community networks to survive in various environments. This community structure depends on numerous factors like nutrient availability, abiotic factors like temperature and pH as well as microbial composition. Categorising accessible biomes according to their habitats would help in understanding the complexity of the environment-specific communities. Owing to the recent improvements in sequencing facilities, researchers have started to explore diverse microbiomes rapidly and attempts have been made to study microbial crosstalk. However, different metagenomics sampling, preprocessing, and annotation methods make it difficult to compare multiple studies and hinder the recycling of data. Huge datasets originating from these experiments demand systematic computational methods to extract biological information beyond microbial compositions. Further exploration of microbial co-occurring patterns across the biomes could help us in designing cross-biome experiments. In this review, we catalogue databases with system-specific microbiomes, discussing publicly available common databases as well as specialised databases for a range of microbiomes. If the new datasets generated in the future could maintain at least biome-specific annotation, then researchers could use those contemporary tools for relevant and bias-free analysis of complex metagenomics data.

Introduction

Microorganisms carry out numerous functions in essential ecosystem processes and collectively shape biodiversity. Along with their crucial role in human health and diseases¹¹⁴, microorganisms across other diverse environments help us with a range of applications from the agricultural industry⁸⁸ to food processing⁹⁴ and generating a sustainable utility of wide environmental resources¹⁵. However, microorganisms do not prefer solitary existence and demonstrate utilisation of resources through a complex trade-off between cooperation and competition⁷². Understanding these co-existing microbial communities is crucial in the post-genomic era for re-engineering communities to serve designated functional deliverables.

With recent advances in high-throughput sequencing techniques in terms of cost reduction, an enormous amount of complex metagenomics data has been generated. Over the last decade, it has helped researchers to investigate a variety of microbial communities, from human gut⁹⁹, livestock¹⁰⁹, plant rhizosphere⁶⁶ to deserts⁵, polar ends¹¹⁶, deep-sea hydrothermal vents^12,31, earth surface⁹⁸ and beyond^96,103 in a culture-independent way. However, these large quantities of data have not been matched with the comparable computational resources available. The lack of standardised protocols, proper data annotation, data archiving with metadata, common analysis pipelines, and inadequate resource organisation has restricted further discoveries through underutilizing the available microbial sequences and constrained meta-learning and meta-mining. This insufficient data interoperability led to the bulk of microbiome data being “single-use” without being repurposed beyond the original scope of the study⁵⁵.

In this perspective, we provide an overview of available microbiome resources and a brief overview of the current state of computational microbiome research and its importance. In addition, we highlight some of the advances made in analysing complex microbial sequencing data and understanding their interactions in the communities. We believe it will further stimulate microbiome researchers to initiate large-scale meta-analyses of the available sequencing data and develop robust computational tools to generate more valuable insights from the available individual studies.

Microbiome Databases

To study complex microbial networks, community structures, and host–microbiome interactions, microbiome databases derived from metaomics experiments are essential. Common data repositories like Integrated Microbial Next Generation Sequencing (IMNGS)⁵⁶, Metagenomic RAST (Rapid Annotation using Subsystems Technology)⁵⁰, MGnify by EMBL-EBI⁷⁸, NCBI (https://www.ncbi.nlm.nih.gov), SILVA⁸⁶ host bulk of the microbial metagenomics and other metaomics data sets. However, it is important to understand the nature of the host environment which directly impacts the composition and conformation of its microbial community. Several biome-specific databases have been developed over the past decades which could be utilised to understand microbiomes across different habitats. Below we classified the available microbiome databases into five broad categories namely environmental, human, animal, plant, and industrial microbiome. Selected databases with prominent and structured datasets belonging to these categories are reported in Table 1 with brief descriptions and their biome content.

Table 1:

Some examples of microbiome databases: The table provides a brief description of some key example microbiome databases, along with web URLs and references

Database name	Type of microbiome	Description	Link	References
A Database for Domestic Animal Gut Microbiome Atlas (ADDAGMA)	Animal	Domestic animal gut microbiome data associated with specific phenotypes such as health and production traits	http://addagma.omicsbio.info/	111
Amadis	Human	Database for association between diseases and gut microbes	http://gift2disease.net/GIFTED/	64
Animal Microbiome Database (AMDB)	Animal	16S rRNA gene profiles for animal gut microbiome across multiple projects	http://leb.snu.ac.kr/amdb	113
AnimalMetagenome DB	Animal	Metagenomic data from domestic and wild animals	http://animalmetagenome.com/	47
Atacama Database	Environmental	Information on microbes present in Atacama desert	https://www.atacamadb.cl/	25
BASE: the Biomes of Australian Soil Environments soil microbial diversity database	Environmental	Database of amplicon data, complete with rich contextual information on edaphic, aboveground diversity and climate	http://www.bioplatforms.com/soil-biodiversity/	17
Beebiome Data Portal	Animal	Catalogue of bee-associated microbes and viruses	https://beebiome.org/	35
BeeExact	Animal	Facilitate standardized classification of bee-associated microbial communities, improve cross-study reproducibility, and help to highlight novel candidate taxa in need of characterization	https://github.com/bdaisley/BEExact	27
biogasmicrobiome	Industrial	Microbiome data from anaerobic digesters and processes such as lignocellulose degradation	https://biogasmicrobiome.env.dtu.dk/	21
BugSigDB	Human	Manually curated database that standardises results from previous studies on microbial signatures	https://bugsigdb.org/Main_Page	41
Coral Microbiome Database (Coral MD)	Animal	Compilation of SSU rRNA sequences from corals	https://www.bco-dmo.org/dataset/724355	48
Disbiome	Human	Cataloguing microbiome composition changes in disease conditions—queries can be made based on data such as disease or microorganism name	https://disbiome.ugent.be/home	49
Earth Microbiome Project (EMP)	Environmental	An initiative to catalogue the microbiomes across the globe	https://earthmicrobiome.org/	43
Extreme Microbiome Project (XMP)	Environmental	Description and characterisation of microbes surviving under extreme environmental conditions	http://extrememicrobiome.org/	100
Food Microbe Tracker	Industrial	Ribotype, phenotype, gene sequences, PFGE (pulsed-field gel electrophoresis) data for food-associated microbes	https://www.foodmicrobetracker.net/login/login.aspx	105
FoodMicroBionet	Industrial	Data from and correlation network analyses for 17 studies on food-associated microbes	http://www.foodmicrobionet.org/	83
Genomic Features Of Bacterial Adaptation to Plants (GFOBAP)	Plant	Data on plant-associated microbes and genes that contribute to this association	http://labs.bio.unc.edu/Dangl/Resources/gfobap_website/	62
GMRepo	Human	Curated metagenomic data from the human gut	https://gmrepo.humangut.info/	26
GreenGenes	Common	16S rRNA database that checks for chimeras	https://greengenes.secondgenome.com/	29
Human Microbiome Project	Human	Microbiome data from human skin and mucosal samples	https://hmpdacc.org/hmp/	68
Human Oral Microbiome Database	Human	Taxonomic and sequence data from bacteria associated with the mouth and aerodigestive tract	https://www.homd.org/	37
Integrated Microbial Next Generation Sequencing (IMNGS)	Common	Platform to screen for and process all 16S rRNA sequences available in SRA	https://www.imngs.org/	56
Life Science Data Archive (LSDA)	Environmental	Life Science Data Archive (LSDA) is a searchable NASA-federated data archive containing information on flight analog data along with the respective human, plant and animal data	https://nlsp.nasa.gov/explore/lsdahome	108
MG-RAST	Common	Metagenomics data repository which includes microbiome data	https://www.mg-rast.org/	50
Mgnify	Common	Submission and analysis of microbiome data from various biomes (including human, aquatic, plant and wastewater microbiomes) and including multiple analysis types such as amplicon, assembly and metagenome data	https://www.ebi.ac.uk/metagenomics/	78
MicrobiomeDB	Human	Database and tool for querying microbiome studies (including human and animal) based on metadata queries, including sample details and taxon abundance	https://microbiomedb.org/mbio/app	82
MiDAS	Industrial	16S rRNA and taxonomic data for wastewater plant-associated microbes	https://www.midasfieldguide.org/	33
NASA Genelab	Environmental	NASA Genelab is a collection of omics datasets for the biological samples exposed to radiation and ions with high atomic number and energy. Along with the datasets, Genelab goes a step further to provide analytic and visualization tools for host-associated molecular signals	https://genelab.nasa.gov/	16
PAMDB	Plant	MLST/MLSA data for plant-associated microbes	http://genome.ppws.vt.edu/cgi-bin/MLST/home.pl	4
SILVA	Common	rRNA database for organisms from 3 domains of life—Bacteria, Archaea and Eukaryota	https://www.arb-silva.de/	86
SKIOME	Human	Curated skin microbiome datasets with metadata	https://github.com/giuliaago/SKIOMEMetadataRetrieval	1
SpaceLID	Environmental	Space Life Investigation Database (SpaceLID) records investigations published in peer reviewed academic journals. In SpaceLID, each publication constitutes an entry, which commonly includes the procedure and outcomes presented in figures and tables of multiple experiments	https://bidd.group/spacelid/	108
The Integrative HMP (iHMP) Research Network Consortium	Human	Expanded version of HMP—includes IBD, Type-2 diabetes and preterm birth	https://hmpdacc.org/ihmp/	67
UNITE	Common	Database for eukaryotic ITS sequences—provides representative sequences for a Species Hypothesis (SH) which serves to act as a bridge between OTU and species classification for fungi	https://unite.ut.ee/	81
Vaginal Microbiome Consortium	Human	Data from the Vaginal microbiome project, including 16S rRNA, whole metagenome shotgun sequencing data, and phenotype and clinical metadata	https://www.vmc.vcu.edu/index.html	39

Environmental Microbiome

The complex interplay among microorganisms in a community and their role in utilising environmental resources drive species richness of the biosphere. Understanding this microbial diversity, complex patterns of distribution, interactions, and community assembly across diverse ecosystems is essential to find ecological selections, the effect of geological parameters, and surveilling the acquisition of antimicrobial resistance (AMR). From 2010 onwards, the Earth Microbiome Project (EMP) has assembled 27,751 samples from 97 different studies to reveal the microbial landscape of the earth’s environments ranging from plant surface, rhizosphere to fresh water, salt water, soil, sediment, animal gut, and many more^43,98. RefSoil+³⁴ and BASE (abbreviated from the Biomes of Australian Soil Environments)¹⁷ are two extensive databases on soil microbiomes. The former contains plasmid information from soil microbiomes with the potential to transfer antibiotic resistance genes (ARGs) from the environment to clinical strains, while the latter directs us to soil metagenomics data from Australia and Antarctica, initiating an improvement of agricultural outcomes and restoring biological functionalities complementing physicochemical properties of the soil.

Apart from ‘hospitable’ environments, microorganisms thriving in extremely harsh environments, ranging from the deep oceans to dry deserts, exhibit growth under limiting conditions⁷⁷. In 2014, a scientific consortium initiated by the Association of Biomolecular Resource Facilities (ABRF) Metagenomics Research Group (MGRG) called The Extreme Microbiome Project (XMP), attempted to discover and characterise the extremophiles and novel organisms from Lake Hillier in Australia, Darvaza gas crater in Turkmenistan, Permafrost tunnels in Alaska, Ethiopian toxic hot springs, and many more extreme places¹⁰⁰. Along with the novel biomolecular techniques for sample processing at extreme locations, XMP projects have benchmarked several computational tools for de novo taxa classification and identification of novel biosynthetic gene clusters (BGCs) from minimally assembled genomes³². Microorganisms that evolved and adapted in these extreme environments possess heterogeneous biological resources with potential biotechnology applications. Atacama Database is a freely available, curated database with 2302 cultured and uncultured microorganisms from one of the aridest and rich in diverse microorganism deserts on the planet, Atacama, in northern Chile²⁵.

Moreover, microorganisms present in human inhabitance strongly influence human health. Confirming biological cleanliness in built environments is essential for protecting human health and avoiding contamination in industry-level assemblies⁷³. Along with inhabiting humans, the microbiome of a built environment depends on many factors, including geographical location, ventilation, usage, and constructional design²³. An interesting study led by Jack Gilbert in 2014, Home Microbiome Project, aimed to understand the dynamic microbial interactions between home and their habitats. A total of 1609 samples from a combination of seven houses, 18 people, three dogs, and one cat were collected. This longitudinal study showed that unique microbial patterns in the homes could be predicted by their occupants, and microorganisms are able to colonise rapidly in the new environments as the habitats move house⁵⁹. Later, the International Metagenomics and Metadesign of Subways and Urban Biomes (MetaSUB) Consortium, established in 2015, gathered over 4500 samples from 60 cities over a period of 3 years and generated a global catalogue of the urban microbiome²⁸. MetaSUB initiative noticed a core set of non-human microorganisms consistently present across the cities. They identified novel bacterial and viral contents in the samples and the global widespread of ARGs. However, other built-in cleanroom environments like operation theatres (OTs) and intensive care units (ICUs) need to maintain a standard operating environment with minimal microbial content⁸⁷. Recent studies from hospitals investigated a detailed understanding of cleanroom microbiomes and their requirement for the maintenance and regulation of the microbial community structure of the built environment^87,90,110.

The study of extraterrestrial environments and spacecrafts equips us with insights into the effect of space on microbiota, their survival, metabolite production, biofilm formation, development of drug resistance, as well as the impact of space travel on astronauts’ health. Through their extensive specialised experiments, Spaceflight Life Investigations (SLIs) generate new perspectives on basic microbiology studies and provide unique opportunities for pharmaceutical industries¹⁰⁸. Life Science Data Archive (LSDA) is a searchable National Aeronautics and Space Administration (NASA)-federated data archive containing information on flight analogue data along with the respective human, plant, and animal data⁹³. Further, NASA Genelab is a collection of omics datasets for the biological samples exposed to radiation and ions with high atomic numbers and energy. Along with the datasets, Genelab goes a step further to provide analytic and visualisation tools for host-associated molecular signals¹⁶. Another initiative, Space Life Investigation Database (SpaceLID), collected and reported procedures and outcomes of 448 SLIs across 90 species from multiple studies¹⁰⁸. A recent study by Kumar et al. revealed the central role of microbial metabolism and metabolite dependencies in the development of the International Space Station (ISS) microbiome through combined graph-theoretical and constraint-based modelling approaches⁵³. Computational approaches such as these can provide significant insights towards unravelling the complexity of various microbiomes.

Human Microbiome

Microorganisms inhabit collectively on multiple human body sites, from the digestive tract and skin to respiratory and reproductive tracts, and possess 100 times the gene content of a human⁴². The complex microbial ecosystem in different parts of the body is unique in its composition and provides more genetic diversity and flexibility to the host⁴². Blooming research on the human microbiome provides evidence for their implications in critical host functionalities like innate and adaptive immune system development¹¹⁷, energy homeostasis^22,24, and metabolism¹¹⁵. For instance, more abundant gut microbiota can partly explain the immune system response modulation⁹¹. Despite the significance of the human microbiome, the microbiota composition can be easily influenced by several intrinsic (host genetics and age)^30,70 and extrinsic (lifestyle, geographical location, diet, and antibiotic uptake)^54,79,80 factors and leads to higher interpersonal variation. The alteration in microbiota composition at spatial and temporal scales within a person (intrapersonal variation) along with interpersonal variation makes it difficult to untangle the microbiota association with human wellness or with the disease onset and progression and understand the influence of external factors on the microbiota composition.

To unravel the underlying host–microbiota and microbe–microbe interactions, mechanistic studies are done on different populations present across the world. One such interdisciplinary study, Human Microbiome Project (HMP)⁶⁸, attempted to find new ways of determining disease predisposition and improving health through the monitoring and manipulating human microbiota by analysing the skin and mucosal (mouth and gut) microbiota samples collected from the United States. The project is expanded with the longitudinal microbiome samples of people suffering from Inflammatory Bowel Disease (IBD), Type-2 diabetes, and preterm birth to understand the impact on the dynamics of human health and disease (The Integrative HMP (iHMP) Research Network Consortium)⁶⁷. Since the human gut microbiota is highly dense and involved in critical host functions, several studies on the European population (MetaHIT^63,85), African population², and Asian population (Asian Microbiome Project^52,97, Westlake Gut Project⁴⁵) have been done to investigate the interaction of intestinal microbiota with the human host and the trajectories of microbiota development.

The compositional analysis of the human microbiota offers only a restricted view of the microbial assembly, and employing ecosystem-specific gene and genome functional information with the composition could help in probing the evolutionary trajectory of the microbiota structure. The Unified Human Gastrointestinal Genome (UHGG) catalogue³ showcased a community effort to build a reference sequence repository comprising 2,04,938 genomes from 4644 microorganisms and provided a glimpse of the structural and functional diversity of the human gut microbiota. Following the gut microbiome, multiple studies on the oral microbiome are collated into Human Oral Microbiome Database (HOMD) to improve the taxonomic assignment quality using standard reference sequence of the microorganism present in the nose and alimentary canal³⁷. Though these databases allow the researcher to access genomic information from multiple studies, raw sequences need to be analysed through the metagenomic pipeline (QIIME2¹⁹, MOTHUR⁹², MetaPhLAn4¹⁸ and HUMAnN¹⁴) and visualised using a bar graph to identify the variation in microbiota composition. But, MicrobiomeDB, a data-discovery and analysis tool, aids in collecting samples from 20 human microbiome studies, including HMP, ECAM, BONUS, and MAL-ED cohorts in a biom format, which contains taxonomic information of the microbiome rather than raw sequences⁸². In-built characterization of microbiome samples for diversity metrics, relative abundance, and differential abundance analysis and interpreting results through visualisation enables the user to understand the data and design the experiments accordingly.

Common microbiome databases such as MGnify and CNGB-MicrobiomeDB, and microbiota-specific databases such as GMRepo (gut)²⁶, Human Microbiota Bioactive Resources (gut), SKIOME (skin)¹, Vaginal Microbiome Consortium³⁹ presented a collection of datasets on various human microbiomes from different demographics. The collected data can be harnessed to predict the characteristic microbial signature or biomarker associated with the physiological condition of the host using graph-based⁶¹, constraint-based⁴⁶ or machine-learning-based approaches⁷⁴. Databases such as BugSigDB⁴¹, Disbiome⁴⁹ and Amadis⁶⁴ curated the microbial association with various disease conditions based on experimental evidence and differential abundance studies. These comprehensive databases could help the research community to understand the context-specific effects of a microorganism in microbiota. Apart from studying the microbiome-host phenotype association to engineer the microbiota for diagnostic/ therapeutic interventions, human microbiome research can also assist in exploring the evolutionary history of human microbiota and discovering novel bioactive compounds to address the challenges associated with the treatment of multi-drug resistant (MDR) human pathogens.

Animal Microbiome

In recent years, apart from the environmental and human microbiome, researchers have started to explore the microbiome of species such as livestock^8,109, bees³⁵, howler monkeys⁶, and corals¹⁰⁶ to understand their impact on pathogen resistance, chemical resistance, and climate change⁹. Among non-human microbiomes, the animal gastrointestinal microbiome comprises a wide variety of microorganisms that play a crucial role in the physiological processes of the host, including development, immunity, and metabolism^38,112 and also induce pathogenesis at times⁵¹. For investigating bidirectional host–microbe interactions across cross-host and cross-phenotypes, robust databases have become necessary lately. The Animal Microbiome Database (AMDB) by Yang et al.  is a manually curated, user-friendly database containing bacterial 16S ribosomal RNA (rRNA) gene profiles of 10,478 bacterial taxa across 467 animal species from 34 major projects¹¹³. The interactive interface of AMDB is convenient for searching by taxa of interest, taxonomic profiles of the samples, and summary of included projects. Another comprehensive database, ADDAGMA (abbreviated from A Database for Domestic Animal Gut Microbiome Atlas), is a publicly available database for domestic animal gut microbiome equipped with browsing and query-based searching for annotated microbial and metadata from major four domestic animals (cattle, chicken, horse, and pig) along with 48 relevant phenotypes¹¹¹. Databases of these volumes are quite useful for capturing the variation in the animal gut microbiome and their association with host phenotypes. Despite possessing a large quantity of microbial data, AMDB and ADDAGMA only contain bacterial diversity, though animal guts comprise archaea, fungi, protozoa, and viruses as well. A recent database CRAMdb has facilitated the users to study consistently annotated microbiomes (278 archaea, 9430 bacteria, 2216 fungi, and 458 viruses) across 516 animals and 43 phenotypes⁶⁰. This exhaustive database provided a microbiome landscape from the gastrointestinal tract and other body sites (e.g. skin) of the host. The animal metagenome database, popularly known as, AnimalMetagenome DB, is another public animal metagenomic database majorly from wild animals⁴⁷. Other macro databases like IMNGS and MGnify also contain metagenomics data from multiple sources, including animals as hosts.

A specialised gut microbial community in honey bees plays a critical role in their growth and development and, subsequently, in the production of honey from nectar²⁰. Dysbiosis of bee gut microbiome through environmental factors such as stress, diet, and antibiotic exposure increases the risk of pathogenic attack, leading to significant economic loss to apiarists⁸⁹. Beebiome Data Portal is a curated database with 14,418 biosamples from 292 bioprojects, along with metadata and reference sequences of the bee-associated microorganism³⁵. Moreover, to address under-annotated bee microbiome, exploration of uncultivated microbial dark matter, and inconsistent phylogenetic placement,²⁷ developed a curated, comprehensive and non-redundant reference database, BEExact, based on 16S rRNA sequencing²⁷. Further, away from the land, swallowed in the ocean, marine invertebrate corals provide an abode for a plethora of microbial communities with multiple inter-species interactions. Coral microbiomes, also known as Beneficial Microorganisms for Corals (BMC), predominantly consist of bacteria and archaea, and they play a quintessential role in the health and ecology of corals⁸⁴. In the era of anthropogenic perturbations, conserving marine ecosystems through coral protection and subsequent restoration, manipulation, and reversing microbial dysbiosis is crucial. Coral Microbiome Database (Coral MD) is a collection of 21,000 small subunit rRNA sequences from nine studies. It identifies a set of 39 bacteria and two archaea specific to coral holobiont⁴⁸.

However, owing to the large and decentralised nature of the data, it is still difficult for users to meta-analyse, mine, and compare the data sets. Qiita, a web-based, open-source microbiome study management platform, is popularly used for quick re-analysis of data sets using standardised platforms with graphical user interfaces⁴⁴. Also, MG-RAST is a repository of metagenomics data along with the tools for annotation and analysis of them⁵⁰. These tools are constantly evolving in terms of pipeline development and data standardisation to complement the pace of sequencing advances over the last decade.

Plant Microbiome

The plant microbiome plays an important role in plant health, nutrient uptake, disease resistance, and tolerance to stressful environmental conditions¹⁰². Typically, the plant-associated microbiome can be divided into the phyllosphere, a microbiome associated with the aerial or above-ground regions, and the rhizosphere, which is associated with the root or the regions of the plant that are below-ground. Additionally, endophytes colonise the region within the plant’s stem, roots, leaves, or other tissues⁷¹. Root-associated fungi are of two types, ectomycorrhizal and endomycorrhizal fungi, with ectomycorrhizal fungi playing an important role in root defence against pathogens⁷⁵. The plant microbiome is recruited from various “reservoirs”—the rhizosphere microbiome, for instance, is recruited from bulk soil via root exudates¹⁰⁷. One challenge that arises while identifying and cataloguing the microbial diversity associated with plants is the correct identification of different microbial species. In fact, due to the slow advances in the taxonomic assignment of sequences, around 43% of 16S rRNA records in GenBank are classified as “environmental” samples or “unclassified”²⁹. Databases such as SILVA⁸⁶, GreenGenes²⁹, and UNITE⁸¹ are used for the identification of microorganisms. The GreenGenes database checks for chimeras, which are common in 16S rRNA screening, and performs alignments and taxonomic classification of species. The SILVA database includes data on bacteria, eukaryotes, and archaeal microorganisms. It can further be used for 16S rRNA-based taxonomic assignment along with UNITE, a fungal Internal Transcribed Spacer (ITS) sequence database. A review by Lucaciu et al. discusses a number of such tools and databases that can be used in plant microbiome studies⁷¹.

While these are commonly used resources and can be used in different kinds of studies, some plant-specific databases can be used to deposit and access microbiome data. The PAMDB database allows Multilocus Sequence Typing (MLST) and Multilocus Sequence Analysis (MLSA) data to be deposited along with metadata such as the region and year of data collection⁴. The metadata also includes various substrates, such as soil and water, from which microorganisms have been isolated. Later, Trivedi et al. suggested a trait-based or functional approach for characterising plant microbiomes and their association with the plants¹⁰¹. A comprehensive and massive initiative by Levy et al. sequenced 484 genomes from root bacteriomes and compared more than 3500 genomes to identify and characterise candidate genes required at the bacteria-plant interface^{62 63}. Another challenge in plant microbiome studies involves contamination from the host DNA, which is difficult to resolve due to the close phylogenetic relationship between plant organelles such as chloroplasts and bacterial sequences. It is essential that DNA sequences from the host plant, including plastid sequences, are filtered out accurately⁴⁰. Databases such as the CpGDB database, which maintain a collection of chloroplast sequences from 3823 plant species, can be used to overcome this effect⁹⁵.

Effectively utilising the microorganism associated with crop species or the crop microbiome has been proposed as a strategy to increase crop productivity and food security. Crop improvement techniques that focus on the rhizosphere and endophytic microorganisms are more likely to succeed compared to techniques that utilise microorganisms from the bulk soil due to evolutionary processes that have shaped interactions between host plants and their microbial communities⁹⁵. The AgMicrobiomeBase project collected metagenomic data from the soil microbiome (bulk soil and the rhizosphere) for crops such as barley and wheat across nine different types of soil in the UK (https://agmicrobiomebase.org). However, there are some challenges associated with harnessing the crop microbiome, one of the major ones being standardising the use of specific microorganisms under different environmental or soil conditions. Moreover, algae-associated bacteria and fungi play an important role in the functioning of the holobiont. However, most studies on algal microbiomes or the phycosphere have been focused on wastewater treatment studies⁶⁵. Studies of the forest microbiome involve integrating data from a number of sources, including soil, roots, foliage, bark, rocks, and even dead wood. All these microbiomes together form a complex and dynamic ecosystem that is primarily influenced by the trees present¹⁰. However, we noticed a relatively small set of databases that explicitly targeted plant microbiomes. Macro databases such as MGnify, which are used to deposit microbiome data, reveal disparity as it has 99,926 samples for the human digestive system alone, outnumbering the number of plant samples (28,666) by more than three times (Accessed on 18th January 2023). We showed the distribution of the samples present in the MGnify database across the described biomes in Fig. 1 for providing an estimate of experiments conducted across these biomes. Approaches such as MLST/MLSA which are used in general for generating sequencing data could be exploited for getting more information on unexplored plant microbiomes.

Figure 1:

The pie chart in the centre depicts the segment of samples from MGnify, one of the major microbiome databases, across discussed biomes. The corresponding percentages represent biome-specific samples in MGnify. Other specific databases for each of the biomes are listed adjacently. This illustration was generated using Adobe Illustrator CS6.

Industrial Microbiome

Diverse microbiomes influence complex processes¹¹ like the synthesis of polymeric substances and maintaining geochemical cycles in nature, making them important contributors to industrial applications. Microbial cooperation is extremely important to perform these complex activities, which a single microbial species cannot perform individually. The food industry is one such industry that can make the most out of microbial communities¹⁰⁴. Microbiome studies in this field focus on the microbial composition of a healthy gut for designing probiotics⁷⁶. Assessing the metabolic capability of microbial communities for product production, sequencing microbial genomes in complex microbiomes and characterising gut microbiome is essential for healthcare⁷. A study by Beck et al. monitored changes in food microbiome in 31 high protein powder (HPP) samples in poultry meals¹³. A microbiome analysis pipeline was developed to detect the microbe composition in each sample and used it as an indicator for the presence of contaminants in the poultry meal. An interesting initiative, the Food Microbe Tracker database, contains ribotype, phenotype, gene sequences, PFGE (Pulsed-field Gel Electrophoresis), and reference data for 96,792 isolates in total¹⁰⁵. Moreover, the FoodMicroBionet database (while currently inaccessible) has collected data from 17 studies on microorganisms associated with foods such as dairy and meat, along with network-based correlation analyses⁸³.

Wastewater management is another concern in industries. Microorganisms can play a major part in treating wastewater owing to their ability to degrade contaminants and pollutants in industrial exudates. Understanding wastewater or sewer microbiomes has implications for human health and for wastewater treatment. These microbiomes vary with time and are influenced by the season and environmental conditions⁵⁷. The MiDAS database contains 16S rRNA and taxonomic data for wastewater treatment plant microbiomes collected by the MiDAS consortium³³. Moreover, anaerobic digester microbiomes are required for the proper degradation of the organic matter, and they are essential for the production of the final product, methane³². The biogas microbiome database (https://biogasmicrobiome.env.dtu.dk) contains data from various anaerobic digesters, including genomes from microorganisms associated with biogas reactors and processes such as lignocellulose degradation, cheese whey degradation, and saccharide degradation.

Conclusion

In the recent past, microbiome research has grown rapidly, with an emphasis on studying microbial interactions and their functions in communities. Being associated with human physiology to industry-level bioprocesses, microorganisms perform a range of functions through their diverse metabolic potential. Identification of novel microorganisms and understanding their co-occurrence patterns, community dynamics, and AMR are some of the important aspects of microbiome research. Despite multiple experimental and computational methods developed we are yet to explore the role of microorganisms in many other contexts due to cost and time constraints. Repurposing the existing data for designing novel strategies to study microbiota at temporal and spatial scale will be beneficial for deciding specific wet lab experiments and speeding up the research. To streamline the data comparison and re-analyses, we reviewed major microbiome databases and resources across environments and hosts, enabling the integration of data types.

In this paper, we argue the importance of establishing common microbiome data repositories for cataloguing the microbial composition of the community of interest and facilitating further novel analyses. We categorise the databases into five major biomes namely environmental, human, animal, plant, and industrial microbiomes, and discuss the benefits and limitations of the existing resources. Environmental microbiome being one of the most important microbiomes determine the sustainability, and safety of the environment we are living in. Temporal microbiome data (e.g. from MetaSUB) have the potential to track the dynamics of the microbiomes, characterise the AMR, and subsequently, find the BGCs. Moreover, microorganisms thriving in extreme environments exhibit auxotrophy and communicate cooperatively in the community. Identified small molecules from these environments show the extreme metabolic capacity of the extremophiles²⁸. Despite having many microbiome databases capturing the meta-omics of the environmental microbiome from different environments, they lack meta-information from the environments, which are very crucial factors in exploring the context-specific variation in altering the composition and functionality of the community⁶⁹.

Human microbiome research has progressed tremendously compared to other ecosystems such as plant and animal microbiomes. The identification of probiotics/ biomarkers to engineer the microbiota using a computational approach has attained great attention due to their therapeutic/ diagnostic potential and better bioavailability. However, longitudinal studies on extensive sampling of multiple body sites (from each subject) with multi-omic analysis are needed to unveil the microbiota evolution, interactions between the various microbiome within humans (skin, oral, gut microbiome, etc.) and crosstalk between the host metabolism and community function. On the other part, the animal microbiome also becomes an essential domain of microbiomes owing to their contribution to our daily life. The animal gut is the most studied organ for microbiome research for its role in producing several by-products of interest. However, other microbiomes like skin microbiomes, are closely associated with the fitness of their animal host and are poorly studied⁵⁸. We further suggest that characterising other animal microbiomes will be useful for understanding microbial symbiosis, the immune system of the host, and tracking their exposure to environmental contaminants.

The plant microbiome is perhaps one of the underrated microbiomes. Microbial interactions in plant microbial communities control plant growth, health, nutrient uptake, and subsequently regulate beneficial traits. A recent study showed that data-driven modelling approaches in plant microbiomes can link the data to the phenotype through multiscale computational simulations¹⁰¹. Modelling engineered microbial communities will be effective for the safe and sustainable production of plants on a large scale. Moving forward, industrial microbiomes are very application-oriented as it involves studying microbes that are involved in industrial processes and wastewater management. Despite having several databases, in most cases, the data are generated for only a specific purpose and remain unusable beyond the scope.

Moreover, the recent outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in 2019 caused a global socio-economic disruption and snatched over 5 million lives to date. Though viruses are often minimally explored in microbiome studies, the COVID-19 pandemic has driven researchers to investigate host-virome interactions. VIROME is a web application with robust pipelines for analysing viral metagenomics sequences and secondary analyses³⁶. Rapid progress in computational tools during these necessary times accelerated the research and marked its capabilities. Nonetheless, with the development of new algorithms and quick analysis of data, storing the structured outcome in a database will increase the scope and the size of the domain. However, even with the many advances in computational methods, different ecosystems demand different data collection and preprocessing steps, restricting robust tool development and comparative studies. Development of tools to predict microbial interactions in different ecosystems could help in understanding the ecosystem-specific community structure and their role in the host–microbiome interactions. In summary, our review will encourage the community to develop a generic database of microbiomes and increase cross-community Microbiome Wide Association Studies (MWAS) and reveal the hidden potential of the tiny organisms that have a massive impact on our environment, health and well-being.

Biographies

Pratyay Sengupta

is a Prime Minister’s Research Fellow at the Department of Biotechnology, IIT Madras. He has a Bachelor of Biotechnology from the National Institute of Technology, Durgapur. His research focuses on developing computational tools for the analysis of genomics and metagenomics data, with a particular emphasis on microbial community analysis.

Shobhan Karthick Muthamilselvi Sivabalan

is a 4th year undergraduate from IIT Madras pursuing his dual degree in Biological Engineering, with a specialisation in Computational Biology. His love for biology and craze for computers have led him to the field of computational biology. His research interests include metabolic modelling and network-based analysis of microbial communities.

Amrita Mahesh

is currently pursuing her B.S.--M.S. in Biological Sciences at IIT Madras. Her research interests include understanding microbial interactions in microbiomes and cancer genomics.

Indumathi Palanikumar

is a Prime Minister’s Research Fellow at the Department of Biotechnology, IIT Madras. She completed her master’s degree in Biotechnology at IIT Guwahati, where she conducted research on characterising a recombinant ovine protein in Pichia pastoris. Currently, her doctoral research focuses on studying the transient impact of enteric pathogens on the gut microbiota of infants in low- and middle-income countries.

Dinesh Kumar Kuppa Baskaran

has a Master of Science (by research) from IIT Madras. His thesis involved developing methods to predict various features of microbiomes by analysing the interactions among the microbes within them. He utilises a combination of network-based and constraint-based approaches to model microbial communities.

Karthik Raman

has a PhD from IISc Bangalore (2009). After a postdoctoral stint at the University of Zurich, he joined IIT Madras in 2011 as a faculty at the Bhupat and Jyoti Mehta School of Biosciences, where he is now a professor. His research group works on developing algorithms and computational tools to understand, predict and manipulate complex biological networks, especially microbiomes. He has 50+ publications and has also authored a textbook on Computational Systems Biology.

Declarations

Conflict of interest

There is no conflict of interest between authors in writing this review.