Fungal Pathogen in Digital Age: Review on Current State and Trend of Comparative Genomics Studies of Pathogenic Fungi

Pathogenic fungi continue to be a problem to humankind to this date. Fungal diseases create severe issues in public health and leave a catastrophic impact on agriculturally important crops. This review intends to highlight the importance of the continuous effort in studying fungal diseases amid more recently popular areas of study, such as virology and bacteriology while suggesting essential bioinformatics resources and techniques that can accelerate new knowledge discovery in this area of interest.

Several massive sequencing projects worldwide have produced many genomic resources for studying fungal pathogenicity. For instance, general genetic sequence resources such as GenBank (Benson et al. 2017), DDBJ (Fukuda et al. 2020), and EMBL (Hingamp et al. 1999) provide a plethora of genetic sequences for research. Specialized databases like the Fungal Genome Initiatives by Broad Institute (Fungal Genomics 2008), FungiDB (Basenko et al. 2018), and EnsemblFungi (Howe et al. 2020), to name a few, are fungi genome databases that serve as a repository for fungal genome and genetic sequences. NCBI, DDBJ, and EMBL are universal repositories for all sequence data types, such as raw sequencing data, whole genome assemblies, gene annotations, protein sequences, and variant calls. These data cover all organisms, including various species of fungi across the kingdom of fungi. Accordingly, bioinformatics analysis of pathogenic fungi is highly challenging because enormous effort is needed for data clean-up to obtain the specific datasets of interest, which creates a gap to be filled by specialized databases or repositories.

FungiDB contains 220 fungus genome sequences for fungi species associated with infectious diseases with mammalian hosts and invertebrate disease vectors (Basenko et al. 2018). FungiDB is an integrated platform for data mining and functional genomics analysis besides containing fungi sequence data. FungiDB provides online bioinformatics tools to allow homology studies using BLAST tools (Camacho et al. 2009), enabling downstream analysis in comparative genomics efforts in various studies such as those performed on Aspergillus fumigatus (Abad et al. 2021) and Cryptococcus isolates (Yu et al. 2021). FungiDB Enrichment Analysis in FungiDB allows GO annotations of the studies and contains many other tools that enable convenient downstream analysis of fungi genomics study. Publicly available fungi genomics data can help accelerate in silico research for the bioinformatics community, and various findings can be discovered more quickly without performing genome or DNA sequencing projects. Launched in November 2000, the Fungal Genome Initiative by Broad Institute was anchored by a group of fungal geneticists and biologists who believed that the speed of discovery in biomedical research was limited by minimal publicly available fungal genome data (Fungal Genomics 2008). Since then, the initiative has focused on species of fungi that are important to human health and commercial activities (i.e., agriculture) and are valuable for fungal diversity and comparative genomics.

Publicly available fungal genomics data are a valuable starting point for downstream analysis for comparative genomics studies. Using annotation data such as genes, proteins, exons, and transcript sequences, researchers can create secondary databases based on data in primary databases. The Pathogenic Host Interaction Database, PHI-base, is a specialized database that catalogues experimentally verified pathogenicity, virulence, and effector genes from fungal, oomycete, and bacterial pathogens (Urban et al. 2020). The database is beneficial as it provides validated experimental data on genes that participate directly and impact the pathogenicity of fungus within host-pathogen interactions. The database is used in various genomics studies of pathogenic fungi in comparative genomics studies, pathogenic genes annotation, and homology searches. For example, the database has been employed to annotate pathogenic genes in Ganoderma boninense (Ramzi et al. 2019) and successfully identify candidate genes that participate in the virulence of Ganoderma boninense in oil palm. It has also been used to predict virulence determinants in draft genomes of Apophyso myces variabilis, where the species are prevalent causative agents of mucormycosis in India (Prakash et al. 2017). The most recent PHI-base release 4.16 contains 9,666 gene sequences in 21,676 interactions. These entries are available for public download for local usage of the data. This provides an opportunity to build a fungal pathogenic genes annotation pipeline that can quickly predict the presence of candidate pathogenic genes in new genome sequence projects.

Fungal pathogenicity in plants involves specific mechanisms to challenge the rigid plant cell wall during the proliferation of the host organism. Fungus secret enzymes that can break down the rigid plant cell wall, and these enzymes are also commonly known as carbohydrate-active enzymes. The Carbohydrate-Active enZymes Database (Lombard et al. 2013), known more popularly by its acronym CAZy, is a database that contains protein sequences of structurally related catalytic and carbohydrate-binding modules that are known to have different modes of interactions with glycosidic bonds, which is a significant linkage and type of covalent bond that joins carbohydrate molecules to other groups. Glycosidic bonds are fundamental linkages in cellular walls (Joseleau and Pérez 2016). Thus, they are the prime target of carbohydrate-active enzymes, which are considered candidate fungal pathogenic genes because of their capability to degrade the plant cell wall. As described in Table III, these enzyme classes and associated modules are involved in various biological pathways of the host organism.

Massive sequence data and literature published on fungal pathogenicity also allow the opportunity to create a database based on these experimental data published in the literature. The Database of Virulence Factors in Fungal Pathogens (DFVF) (Lu et al. 2012) was a project aimed at filling the missing gaps in understanding fungal pathogenicity by aggregating all known virulence factors. This project also develops an algorithm that allows the prediction of potential candidate genes that will contribute to the development of fungal pathogenicity. The database was built by leveraging the textmining technique used by PubMed and the Internet by looking for fungal disease virulence keywords. In-house tools were also developed to allow relevant supporting literature to be searched. With this methodology, the database currently contains 2058 protein sequences.

Other fungal pathogen-related databases provide information about antifungal genetics properties, and one such example is the AFRbase. AFRbase is a database that keeps information about protein mutations responsible for antifungal resistance (Jain et al. 2023), which will help the scientific community understand antifungal resistance as clinicians. Biologists are looking for better treatments and cures for fungal diseases. The database was created through text mining of publicly available research papers and further enhanced with information from publicly available databases such as NCBI. Other similar databases include FunResDB (Weber et al. 2018) and MARDy (Nash et al. 2018). FunResDB focuses on susceptibility testing of Aspergillus fumigatus, which has high public health importance. At the same time, MARDy includes data on existing antifungal resistance in humans, animals, and plants with its associated antifungal agents.

Developing bioinformatics databases and tools that focus on different study paradigms is vital to increasing the spectrum of understanding and helping expand the perspective of biological research on pathogenic fungi. These bioinformatics databases and tools are developed to deal with data in various stages of readiness, ranging from tools like FastQC (de Sena Brandine and Smith 2021) that enable quality control of DNA/RNA sequences generated by sequencing machines to downstream through that deal with the more complex interpretation of data such as Cytoscape (Shannon 2003), VisANT (Hu 2014), Pathway Studio (Nikitin et al. 2003) and Patika (Demir et al. 2002) that enable scientists to explore biological networks as a mean to understand better integrative biology, system biology, and integrative bioinformatics.

Standalone tools include BLAST+ (Camacho et al. 2009), a universally common tool for comparing two or more DNA/RNA/protein sequences to understand the degree of similarity and identity between sequences, which implies the degree of conservation of sequences among subjects of studies. It is often utilized to understand the relationship between species of organisms. ClustalW (Thompson et al. 2002), MAFFT (Katoh 2002), and MUSCLE (Edgar 2004) are other examples of such standalone tools that incorporate statistical analyses of subject sequences, building multiple sequence alignments. Phylogenetics tools such as PHYLIP (Retief 2000) and MEGA (Hall 2013) generate relationship trees of input sequences that allow not only understanding but also visualize relative relationships between multiple sequences in the study. Recent trends in bioinformatics tool development indicate an increasing need within the scientific community to have integrated tools that serve as a “one-stop centre” for biological data analysis. An integrated bioinformatics platform will allow more biological scientists without high computing skill sets to perform bioinformatics analysis, such as executing sequence analysis via multiple bioinformatics tools and visualizing results. This is extremely important as it requires time to understand different bioinformatics tools, and as such, this requirement is a higher barrier to entry for most scientists. Given such unique demands, significantly more integrated bioinformatics analysis platforms are being developed for scientists to conduct integrated sequence data analysis.

Unipro UGENE (Okonechnikov et al. 2012) is an example of a bioinformatics tool that provides a platform for developing an integrated pipeline. UGENE delivers a user-friendly interface for scientists to build their desired bioinformatics pipeline and workflows for sequence data analysis. With many popular standalone bioinformatics tools available within UGENE, it also provides a user-friendly interface for scientists to easily build desired workflows with a drag-and-drop feature that requires minimum computer programming knowledge.

Comparative genomic analysis involves the comparative analysis of sequence data from multiple sources, some within species and some across numerous species. These analyses usually involve multi-stage data analysis and, therefore, require a combination of bioinformatics tools and applications to draw meaningful discussions and deductions while answering experimental hypotheses. Most comparative genomics platforms allow comparative analysis of DNA sequences and streamline the process from data analysis to visualization of results. EDGAR (Dieckmann et al. 2021) is an example of an integrated comparative genomics platform. It is one of the most popular platforms for gene-based comparative genomics and differential gene content analysis. Venn diagrams or synteny plots can be generated to provide a user-friendly and visually appealing interpretation of results. A list of all databases and tools and a link to the website can be found in Table IV.

Studying fungal pathogenicity continuously centers around treatment and diagnosis, intending to identify methods for early diagnosis or eradication of diseases. Studies in this field are ongoing, and sequencing technologies serve as an enabling platform for various downstream research and development projects. They set the foundation for bioinformatics research and development. The discovery of different polymorphic markers, such as single nucleotide polymorphism, insertions and deletions, copy number variations, and the presence of genes, is essential. Each of these polymorphisms plays an important role in causing pathogenicity in fungi, and this could confer pathogenicity to pathogenic isolates, as shown in human research.

The emergence of sequencing technologies has increased the resolution of research into underlying molecular causative factors in molecular plant pathology. Through genome sequencing of plant pathogens like Magnaporthe oryzae (Dean et al. 2005), Botrytis cinerea (Amselem et al. 2011), Ustilago maydis (Kamper et al. 2006), and Puccinia graminis (Duplessis et al. 2011) coupled with improvements in bioinformatics methodology such as genome assembly, genome annotation, comparative genomics, pathologists can identify genomics features in fungal pathogens that play important roles in fungal pathogenicity. On top of that, further understanding of these genomics’ features will allow scientists to pursue and develop faster and more accurate diagnostic tools for fungal-related diseases.

Whole genome sequencing of plant fungal pathogens allows high-quality genome assembly to identify and reveal the underlying sequences of the fungus. The genome annotation of the assembled genome then predicts gene models based on ab initio prediction and homology searches (Yandell and Ence 2012) to known nucleotide or protein sequences. The availability of an annotated genome allows downstream bioinformatics analysis, such as polymorphic markers identification through genome mapping (Davey et al. 2011) and comparative genomics (Wei et al. 2002). A study on Verticillium dahliae proposed the possibility of horizontal gene transfer (HGT) from bacterial origins, which directly contributed to the pathogenicity of the fungus, which is known to be a plant pathogen that affects hundreds of plant species and causes substantial economic losses annually (Shi-Kunne et al. 2019).

The same effort was applied to the comparative genomics of human pathogenic fungi. Candida and Aspergillus are the most prevalent fungal genera that cause significant health implications in human (Moran et al. 2010). Hence, elucidating the sequences at a genomic level is extremely important to allow the development of effective antifungal therapy and understand the emergence of drug resistance. A study was done to understand the drug resistance of Candida auris using genomic data such as epidemiology and evolutionary information (Chybowska et al. 2020). A comparative genomics study was also done on Aspergillus to improve understanding of genome heterogeneity between Aspergillus fumigatus, Aspergillus lentulus, and Aspergillus fumigatiaffinis (dos Santos et al. 2020). These three species are highly similar in morphology, making it challenging to distinguish one species from another by phenotype observation (Izquierdo et al. 2014). Molecular markers can be developed into rapid serology-based test methods that can yield fast results, and one example of such assay is the (1,3)-β-D-glucan (BDG) based assay (Fang et al. 2023). This assay detects a polysaccharide fungal cell wall component from Candida spp. Pneumocystis jivoveci, Aspergillus spp., Acremonium spp., Fusarium spp (Tissot et al. 2013) in patients to determine disease causes. This demonstrated the importance of genomics studies as sequencing and downstream bioinformatics analysis can uncover unique genomics features of each species, allowing accurate diagnostic results.

Comparative genomics techniques were applied in studying not only genetic diversity but also in the discovery of critical genomic markers such as short sequence repeats (SSR), short tandem repeats (STR), long tandem repeats (LTR), and single nucleotide polymorphisms (SNP). A recent study on Fusarium oxysporum is an example of such application of comparative genomics in uncovering genomics markers for quicker detection of pathogenic isolates of the species (van Dam et al. 2018). The study includes candidate effector genes from 88 Fusarium oxysporum genomic assemblies for comparative genomics. It aimed to distinguish the isolates based on the traits where it could differentiate between cucurbit-affecting formae speciales from each other and differentiate between pathogenic and non-pathogenic isolates.

General identification of pathogenic and non-pathogenic fungi often investigates genetic features such as the presence of pathogenicity-related genes and proteins. The presence or absence of pathogenicity-related genes is important in understanding fungal pathogenicity and its viability. This was demonstrated in a study comparing Fusarium graminearum and Fusarium venenatum, non-pathogenic and pathogenic fungus species, respectively (King et al. 2018). The study presented helpful insights to support such a hypothesis. Through comparative genomics involving a comparison of the proteomes of each species, the scientists discovered 15 putative secondary metabolite gene clusters, 109 secreted proteins, and 38 candidate effectors that are not identified in the non-pathogenic subject. One recent study identified gene presence-absence variation (PAV) in Magnaporthe oryzae, which can help understand fungal pangenome evolution (Pierre and Ksenia 2024). This approach is helpful to understand and identify core genes of a specific that potentially supports fungal pathogenicity (Chen et al. 2023). As new methods and tools appear, new ways of studying fungal pathogenicity will continue to evolve and improve.

Unsurprisingly, there are similarities between pathogenic fungi that attack plants and animal hosts (Dickman and de Figueiredo 2011). Both groups of fungi are similar in the mechanisms of pathogenicity, which are all part of the fungi lifecycle from spore germination, invasion via physical openings, colonization and alteration of host, reproduction, and transmission. These similarities in the pathogenicity mechanism prompted the interest in studying fungi not as a separate group but as the same study group, which allows further understanding of pathogenic mechanisms in the kingdom of fungi.

By adopting a different perspective to compare pathogenic fungi that infect plant hosts and animal hosts, genomics identification provides a means of understanding the adaptation of these species of fungi based on host-specificity. Fungal species that infect plant hosts can have broad or narrow host ranges (Sexton and Howlett 2006). Specificity is defined by R genes known as resistance genes in the host and the virulence factors found in the pathogenic fungi (van der Does and Rep 2007). The range of hosts a fungus can infect is not limited to plants or animals. Some extreme examples, like within the genus Fusarium, cause disease across plant species and animals, including human (Sharon and Shlezinger 2013), which makes understanding the mechanism behind pathogenicity even more peculiar.

The same is true for bacterial pathogens; a study has found that Pseudomonas aeruginosa, which causes pneumonia and infections in the blood (CDC 2019) in humans, shows a high degree of conservation in the virulence mechanism used to infect both humans and plants. The pathogen also causes infection on Arabidopsis and sweet basil roots, forming a biofilm layer under specific physiological conditions (Walker et al. 2003). Evidence also showed that the bacterial pathogen used a common subset of virulence factors for pathogenesis in plants and animals (Walker et al. 2003), further demonstrating that pathogens that infect a range of hosts use a common pathogenesis mechanism. Understanding the common mechanism behind the range of potential hosts for infection can shed light and give rise to a better understanding of host specificity and the mechanism of pathogenicity in the kingdom of fungi. This is supported by studies from the past reporting commonalities of pathogenicity among mechanisms for plants and animals (Hamer and Holden 1997).

Comparative genomics studies will create a good foundation for using the identified pathogenicity-related genes and molecular markers for molecular diagnostic purposes. Fungal infections in human or animal hosts are easier to detect and identify compared to plant diseases caused by pathogenic fungi. For example, fungal nail infections known technically as “onychomycosis” can be diagnosed easily as the disease symptoms can be observed visually through the rotting of nails (Gupta et al. 2000). The same can be said for many fungal diseases, such as Aspergillosis, Candidiasis, Mucormycosis, Pneumocystis pneumonia and many more (CDC 2019). Furthermore, fungal infections in humans and animals jeopardize the health and livelihood of the subject; hence, early diagnosis is crucial. Early detection is easier with visible symptoms such as skin rashes or coughing. In addition, diagnosis can be done through direct microscopic examination of clinical samples, histopathology, culture, and serology of patient clinical samples (Kozel and Wickes 2014). Fungal diseases in plants, however, can be challenging to detect as the symptoms are not visible, and it could be too late when symptoms are finally observed. A classic example of this is the basal stem rot (BSR) and upper stem rot (USR) caused by Ganoderma boninense (Hushiarian et al. 2013). As the infection is not visually apparent during the early stages, it will be too late when its symptoms are visible. Once the symptoms are observable, palm trees die within 1 or 2 years to 3 or more years, depending on the age of the palm (Corley and Tinker 2003).

In the case of BSR or USR caused by Ganoderma boninense, traditional diagnostic methods are impractical as it will be too late when the disease symptom is observed. Molecular diagnostic methods using PCR amplifications are the way forward for early detection of fungal diseases that are not observable. This method requires the presence of a unique genome sequence of the target organism, which is usually a well-conserved region with polymorphic markers identifying different species. A specific primer (Hariharan and Prasannath 2021) will amplify the target region of interest. Example target regions of the fungal genome that have been identified for molecular diagnostics include the highly conserved internal transcribed spacer ITS-region in fungus, known for fungal diversity analysis and an important marker for fungal DNA barcoding (Bellemain et al. 2010), and alternative sequences such as cytochrome b gene which was used as a target region for Loop-Mediated Isothermal Amplification (LAMP) assay for detection of airborne Uromyces betae (Kaczmarek et al. 2019). Molecular diagnostic methods provide the possibility of early detection, and these methods can be applied to fungal diseases in plants, animals, and humans.

The negative impact of fungal disease on public health and the global economy requires continuous effort in research and development to expand our understanding of fungal pathogenicity with existing data repositories and bioinformatics tools. However, the rapid advancement of sequencing technology and computational biology calls for creating more fungal pathogen-focused bioinformatics databases and tools that identify causative factors of fungal pathogenicity and the underlying genetics. This will allow in-depth comparative genomics on fungal pathogens and help develop new fungal pathogen identification, diagnosis, and treatment methods.