Genomic Characterization and Functional Description of Beauveria bassiana Isolates from Latin America

Beauveria bassiana is an entomopathogenic fungus used in agriculture as a biological controller worldwide. Despite being a well-studied organism, there are no genomic studies of B. bassiana isolates from Central American and Caribbean countries. This work characterized the functional potential of eight Neotropical isolates and provided an overview of their genomic characteristics, targeting genes associated with pathogenicity, the production of secondary metabolites, and the identification of CAZYmes as tools for future biotechnological applications. In addition, a comparison between these isolates and reference genomes was performed. Differences were observed according to geographical location and the lineages of the B. bassiana complex to which each isolate belonged.


Introduction
Within Beauveria, about 25 species have been described, most of which originated in Asia-the species' center of origin [1]. Beauveria bassiana (Bals.) Vuill (Ascomycota, Hypocreales) is a versatile microorganism capable of killing insects, surviving as a saprophyte in soil, and creating symbiotic associations with plants as an endophyte [2][3][4]. Beauveria bassiana sensu lato has a worldwide distribution and can infect around 700 insect species [5]. Since the discovery of its effect on silkworms [6], Beauveria has been used in the biological control of numerous agricultural pests [7,8].
According to phylogenetic studies based on the nuclear ribosomal internal transcribed spacer (ITS) and the elongation factor 1-alpha (TEF1-a) gene, B. bassiana has a monophyletic origin [9]. Within the species, many lineages have been identified and linked to a specific geographic distribution [10]. However, among strains of the same lineage, genetic variation is not directly associated with abiotic or biotic factors [10]. The same phenomenon has been reported for Metarhizium anisopliae (Metschn.) Sorokin [7].
Currently, there are 17 fully sequenced genomes of B. bassiana from Europe, Asia, Oceania (Australia), and North (USA) and South America (Colombia) available in the GenBank database of the National Center for Biotechnology Information (NCBI, https: //www.ncbi.nlm.nih.gov/genome/browse/#!/eukaryotes/910/ (accessed on 16 May 2023)). More recently, eight additional genomes from Central America and the Caribbean, specifically Costa Rica, Honduras and Puerto Rico, have been added to the database [11], for a total of 25 sequenced genomes. Different levels of virulence and entomopathogenic responses have been linked to genetic variations, such as non-synonymous changes (NSCs) and copy number variations (CNVs) in important genes [12].

De Novo Genome Assembly Process
Prior to the de novo genome assembly, raw Illumina reads were quality-checked using FastQC [17] and processed using Trimmomatic v.0.38 [18] to remove the sequencing adapters and quality reads below a cutoff of 30. Quality passed reads were used for the de novo assembly using SPADES v.3.13.1 [19,20] at the default settings, except for the kmer size values, which were set as 89, 95, 97, 101, 107, 117 and 127.
Genome assemblies were compared and evaluated using Quast v.5.0.2 [21] against the Beauveria bassiana ARSEF 2860 (gb|ADAH00000000.1) reference. To identify the kmer's distribution through the genomes and to plot its distribution, Jellyfish v.2.2.10 [22] and GenomeScope [23] were used, respectively. This allowed us to rapidly determine the overall characteristics of the genomes, regarding genome size, heterozygosity rate and repeat content.

Gene-Calling and Identification of Putative Secreted Proteins
As a first proteome draft, proteins were predicted by implementing Augustus v2.6.1 [26] trained on Aspergillus oryzae, due to its well-documented necrotrophic lifestyle. In order to identify putative secreted proteins within each fungal isolate, SignalP 4.0 [27] was run locally; predicted proteins were considered to be secreted if SignalP identified a secretion signal peptide and no transmembrane domains.

Fungal Phylogeny and Comparative Genomics
A phylogeny tree was constructed using single-copy BUSCO genes [25] for all B. bassiana isolates [11] and reference B. bassiana strains ARSEF 2860 (GenBank accession no. ADAH00000000.1); B. bassiana HN6 (GenBank accession no. GCA_014607475.1); and C. militaris CM01 (GenBank ID no. GCA_000225605.1). Single-copy BUSCO genes were identified within each organism and their corresponding proteins were aligned with Multiple Alignment using Fast Fourier Transform (MAFFT) v7.397 [28] with the options "--maxiterate 1000-auto", and each alignment was trimmed using TrimAl v1.4.rev22 [29] with the option "--gappyout". For each of the 293 single-copy BUSCO genes from each species, we implemented IQTree v.2.2.0-beta software [30]. An automatic detection to identify the best-fitting model with the option "-m MFP" was used, which led to the best JTT + F + R model. To obtain the best-scoring ML tree, five independent tree searches were conducted with the option "--runs 5"; the topological robustness of each gene tree was evaluated using the option "-bb 10,000" with 10,000 ultrafast bootstrap replicates. C. militaris CM01 was used [31] as the outgroup of the tree. Orthofinder version 2.5.2 [32] aided the identification of unique genes and orthologous relationships between the proteomes from fungal isolates and the reference B. bassiana ARSEF 2860 through standard mode parameters.

Metabolic Pathways Description and Fungal Functionality
For the discovery of the potential biological meaning of the fungal isolates, the BlastKOALA tool was implemented as an automatic annotation server for the genome and metagenome sequences [33]. KEGG orthology assignments were performed with BlastKOALA to characterize individual gene functions and for the reconstruction of the KEGG pathways and modules in the fungal isolates.
B. bassiana genome annotation for Carbohydrate-Active Enzymes (CAZymes) was performed automatically using the Augustus called-proteins through the dbCAN2 meta server [34] with the databases dbCAN, HMMM and CAZy for pattern recognition. Hits were considered when an annotation was present in all three databases. Additionally, antiSMASH [35] software version 6.1.1 was used locally to identify important gene clusters associated with secondary metabolites, such as NRPS, TPKS and Terpenes.
To identify the presence of potential virulence-associated genes, blast analysis was conducted against the host-pathogen interactions database (HPIDB 3.0) [36]. The Blosum 62 matrix was implemented and only genes that showed a percentage of identity greater than 70% and a coverage greater than 50% were selected. We searched for PHI genes involved in virulence, genes responsible for degrading insect cuticles, mating-type genes, and core genes involved in the biosynthesis of secondary metabolites. The pathogenicity or virulence reported for a particular fungus was presumed to be similar or identical in B. bassiana, as there are no entries for entomopathogenic fungi in the PHI-base [37]. Matingtype loci identification in B. bassiana isolates and fungal references were identified using homology to characterized MAT genes specific to Clavicipitaceae and Cordyceps sp. using BLASTp. Selected hits fulfill both an e-value of 0.05 or lower and a query coverage filter of at least 55 (blast command line options-qcov_hsp_perc 55-evalue 1 × 10 −5 ).

Genome Assembly and Completeness Statistics
Unprocessed short raw reads were analyzed using Jellyfish and GenomeScope to plot the kmer distribution to check for genome size, heterozygosity rate and repeat content. Globally we observed an optimal genome size, haploidy and low repeat content for all isolates (Supplementary Figure S1).

Fungal Phylogeny and Comparative Genomics
To construct the phylogenomic analysis, as inputs, we used single-copy full-length BUSCO genes from all the B. bassiana sequenced isolates from this study; reference strains Bbas ARSEF 2860 (gb|GCA_000280675.1); BbasHN6 (gb|GCA_014607475.1); and C. militaris CM01 (gb|GCA_000225605.1) were included as outputs. This analysis produced a total of 293 core-genes. For each gene, its corresponding BUSCO protein was selected, concatenated and aligned using MAFFT, and an ML tree was inferred using IQ-Tree ( Figure 1).

Fungal Phylogeny and Comparative Genomics
To construct the phylogenomic analysis, as inputs, we used single-copy full-leng BUSCO genes from all the B. bassiana sequenced isolates from this study; reference stra Bbas ARSEF 2860 (gb|GCA_000280675.1); BbasHN6 (gb|GCA_014607475.1); and C. m taris CM01 (gb|GCA_000225605.1) were included as outputs. This analysis produced total of 293 core-genes. For each gene, its corresponding BUSCO protein was select concatenated and aligned using MAFFT, and an ML tree was inferred using IQ-Tree (F ure 1).

Figure 1.
Maximum likelihood tree constructed from single-copy BUSCO genes for B. bassiana semblies, with Bbas ARSEF 2860, BbasHN6, and C. militaris CM01 strains as references. This t was constructed using MAFFT alignments of 2930 full single-copy protein coding sequen through IQ-Tree, implementing the JTT + F + R5 model and 10,000 replicates for ultrafast bootstr support, depicted in percentages. The tree depicts the Costa Rican isolates (B0, B01, B26 B27 a B31), Honduran isolate (B13), and the Puerto Rican isolates (B43 and B44). Our results show h values for bootstrap and, interestingly, C. militaris CM01, considered as an outgroup clustered to bassiana references (HN6 and Bbas2860) and isolate B31.
The Orthofinder ortholog comparison between the fungal isolates revealed t 96,233 genes (98.1% of the total) were assigned into 10,356 orthogroups. Fifty percent all genes were assigned in orthogroups with 10 or more genes (G50 was 10) and w contained in the largest 4546 orthogroups (O50 was 4546). There were 7047 orthogrou within all species and 6284 of these consisted entirely of single-copy genes. In additi there were 1899 unassigned genes accounting for 1.9% of the total number of genes (Figu S1). All isolates shared a large number of orthologous genes, between 8526 and 9353, cluding the reference genome (ARSEF2860; Figure 2). Interestingly, isolate B31 shared highest number of proteins with C. militaris CM01. Additionally, when in combinat with the shared proteins among the genomes, B26 shared the lowest number of ortho gous genes with other isolates (Figure 2).

Figure 1.
Maximum likelihood tree constructed from single-copy BUSCO genes for B. bassiana assemblies, with Bbas ARSEF 2860, BbasHN6, and C. militaris CM01 strains as references. This tree was constructed using MAFFT alignments of 2930 full single-copy protein coding sequences through IQ-Tree, implementing the JTT + F + R5 model and 10,000 replicates for ultrafast bootstraps support, depicted in percentages. The tree depicts the Costa Rican isolates (B0, B01, B26 B27 and B31), Honduran isolate (B13), and the Puerto Rican isolates (B43 and B44). Our results show high values for bootstrap and, interestingly, C. militaris CM01, considered as an outgroup clustered to B. bassiana references (HN6 and Bbas2860) and isolate B31.
The Orthofinder ortholog comparison between the fungal isolates revealed that 96,233 genes (98.1% of the total) were assigned into 10,356 orthogroups. Fifty percent of all genes were assigned in orthogroups with 10 or more genes (G50 was 10) and were contained in the largest 4546 orthogroups (O50 was 4546). There were 7047 orthogroups within all species and 6284 of these consisted entirely of single-copy genes. In addition, there were 1899 unassigned genes accounting for 1.9% of the total number of genes ( Figure S1). All isolates shared a large number of orthologous genes, between 8526 and 9353, including the reference genome (ARSEF2860; Figure 2). Interestingly, isolate B31 shared the highest number of proteins with C. militaris CM01. Additionally, when in combination with the shared proteins among the genomes, B26 shared the lowest number of orthologous genes with other isolates (Figure 2).

Metabolic Pathways Description and Fungal Functionality
To introduce biological information using KEGG categories and the number of gene counts, we focused on specific functional categories (Table 1). B43 and B44 showed the closest amount of protein counts compared with the reference genome but B44 had a greater protein count than ARSEF 2860. On the other hand, B26 showed the lowest protein count with respect to the reference and the other isolates. All the KEGG terms, in general, were uniformly distributed among the isolates.

Metabolic Pathways Description and Fungal Functionality
To introduce biological information using KEGG categories and the number of gene counts, we focused on specific functional categories (Table 1). B43 and B44 showed the closest amount of protein counts compared with the reference genome but B44 had a greater protein count than ARSEF 2860. On the other hand, B26 showed the lowest protein count with respect to the reference and the other isolates. All the KEGG terms, in general, were uniformly distributed among the isolates. From our analysis, a total of 2349 protein models were assigned to CAZymes distributed into the Auxiliary Activities (AA), Carbohydrate Binding Modules (CBM), Carbohydrate Esterase (CE), Glycoside Hydrolase (GH), Glycosyl Transferase (GT), and Polysaccharide Lyases (PL) families. Most proteins (54%) were associated with the GH family, followed by the GT family (30%) and various other gene families (Table 2). We focused on identifying key members from the GH family, which are involved in the enzymatic deconstruction of chitin [38] and harbor chitinases that are targets for the biological control and biotech industry. For each assembled and annotated B. bassiana isolate, we identified 18 putative glycoside hydrolase family 18 (GH18) CAZymes; whereas for the C. militaris CM01 and B. bassiana ARSEF2860 references, we identified 21 genes for each reference. Other important chitinases in the families GH3 and GH16 were absent in our isolates, while for the families GH75 (Chitosan) and GH76, two and nine enzymes were identified, respectively. Furthermore, family GH10 enzymes, present in our isolates, have only been reported in Beauveria bassiana [39] and are absent in other entomopathogenic fungi (Cordyceps militaris (L.), Metarhizium anisopliae and Metarhizium acridum (Driver & Milner) J.F. Bisch.), while the GH84 and GH95 families, which were reported previously as exclusive for entomopathogenic fungi [39], were observed in all the isolates included in this study. Families GH5_5 and GH45, which target cellulose, were also present in our data as in previous studies [39].
On the other hand, the CBM66 member from the Carbohydrate-Binding Module family is reported to be involved in binding fructans [40]. Interestingly, only C. militaris reported one combination of GH18 + CBM66 (Table S2). From the annotated GH18 subfamily, according to the CAZyme database, some enzymes report a characterized deconstruction of chitin. From these, the chitinase enzyme (EC. 3 In terms of secondary metabolite potential, the majority of the biosynthetic gene clusters were associated with Non-ribosomal Peptide Synthetases (NRPS), whereas those for Polyketide synthases (PKS) and Terpenes were comparable ( Figure 3). All genomes had the same number of terpenes compared with ARSEF 2860 and C. militaris. However, in the B31 isolate, no NRPS and PKS were observed. The Puerto Rican isolates B43 and B44 consistently showed similarity throughout the analysis.

Identification of Functional Pathogenic and Important Virulent Elements and Mating-Type Genes
To identify the presence of potential virulence-associated genes, a blast analysis was conducted against the Pathogen-Host Interaction database v4.5 (PHIbase). Hits above 70% were the most informative and reliable. We identified PHI genes for important gene families, such as virulence-associated genes, degrading insect cuticles genes, mating-type genes, and core genes for the biosynthesis of secondary metabolites (Table 3). From this, we identified orthologues of the most important pathogenic genes [41]. Each gene was grouped into a single-copy gene orthogroup, except for the CYP52X1 and HSP90, as there are two copies for B26 and the Bbas reference, respectively. MrpacC is a transporter involved in ion transport and/or toxin secretion and was only present in the B0 and B1 isolates. Similarly, a hydrophobin (Hyd2), a protein involved in hydrophobic interactions on the insect cuticle, was identified in B0, B1 and B31 (Table 3).
Concerning mating-type genes, we compared the proteome for each B. bassiana isolate from this study against a database downloaded from the NCBI database specific to the Clavicipitaceae and Cordyceps species. This protein database contained 372 unique entries. We identified seven potential orthologues for each genome respective to the database. Interestingly, each gene was in a single-copy gene orthogroup, except for one orthogroup that lacked a protein for B31, and the genome references B. bassiana ARSEF 2860 and C. militaris CM01; this corresponded to a mating-type protein a-1 (orthogroup OG0009027). On the other hand, the six mating-type candidates had a corresponding single-copy protein in C. militaris, and all homologues showed a high level of conservation ( Figure S3). enzymes, the chitinases EC. 3.2.1.96 and EC 3.2.1.17 were not identified.
In terms of secondary metabolite potential, the majority of the biosynthetic gene clusters were associated with Non-ribosomal Peptide Synthetases (NRPS), whereas those for Polyketide synthases (PKS) and Terpenes were comparable (Figure 3). All genomes had the same number of terpenes compared with ARSEF 2860 and C. militaris. However, in the B31 isolate, no NRPS and PKS were observed. The Puerto Rican isolates B43 and B44 consistently showed similarity throughout the analysis.   Two orthologues coded for an HMG-box protein; members of this family include the fungal mating-type gene products MC, MATA1 and Ste11 [42]. The other orthologues coded for a DNA repair protein, rad10, DNA lyase, timeless protein and the fungal pheromone mating factor STE2 GPCR. This orthologue was missing in B31, and both references coded for the mating-type protein a-1. All the isolates showed similarity for the mating-type protein MAT1-2-1, except B31, which showed a MAT1-1-1 gene as the C. militaris CM01 and B. bassiana ARSEF 2860 reference genomes.

Discussion
B. bassiana isolates from different locations in the neotropics were previously sequenced and assembled [11]. For each isolate, Jellyfish and GenomeScope were implemented to plot the kmer distribution and to check for the genome size, heterozygosity rate and repeat content from unprocessed short reads. Although there are no B. bassiana reference genomes at BUSCO, Castro-Vásquez et al. [11] reported values ranging from 96% to 97% using the BUSCO eukaryota_odb9 database fixed for Fusarium graminearum. In the study herein, we additionally used the BUSCO fungi_odb9 fixed set for Aspergillus oryzae, which gave a variation percentage ranging from 88% to 98%, indicating lower universal single-copy orthologous genes found in the assembled genomes with the chosen training data set. According to Waterhouse et al. [43], BUSCO genes have been widely used as a measure of genome completeness, and additionally as markers for fungal phylogenomic relationship inferences [44,45]. Both datasets have different gene categories that made this analysis more robust due to the lack of specific B. bassiana databases.
Phylogenetic analyses performed in previous studies determined that all the isolates (B0, B1, B13, B26, B43, and B44), except for B27 and B31, belong to an African-Neotropical lineage of B. bassiana [15] defined by Rehner et al. [10]. In the analysis herein, of 293 core-genes, almost each isolate represented a particular cluster (B0, B1, B13, B26, and B27). Only the Puerto Rican isolates (B43 and B44) were together in the same cluster. On the other hand, B31 grouped with the reference B. bassiana and C. militaris genomes, corroborating that this isolate is genetically distinct from all the other isolates analyzed. Throughout several analyses, B31 had particular similarities with C. militaris (number of proteins, lack of the orthogroup OG0009027, same mating-type gene) and, at the same time, showed a complete absence of NRPS and PKS compared with the other isolates and reference genomes. In previous analyses, B31 also showed unique alleles in several SSR markers and its lineage could not be determined, while the rest of the isolates were highly diverse but, hence, distinct from Cordyceps [15].
The Orthofinder ortholog comparison between the fungal isolates and B. bassiana (ARSEF 2860, HN6) and C. militaris (CM01) reference genomes confirmed the singularity of the B31 isolate, whose proteome had a greater similarity to C. militaris than the B. bassiana references and isolates. The direct ancestors of B. bassiana were Asian Cordyceps species, according to Xiao et al. [1]. Additional evidence has also confirmed the link between Beauveria anamorphs and Cordyceps teleomorphs [46][47][48]. Among the principal differences between B. bassiana and the Cordyceps species is the capability of the former to infect a wide host range, while Cordyceps species usually manifest host specificity [1]. Metarhizium species have been shown to evolve from specific-host species to generalist-host species [49]. This raises an interesting question as to whether B31 represents an intermediate species between Beauveria and Cordyceps or a new lineage of the B. bassiana sensu lato complex. New species of Beauveria have been reported in South America [50][51][52] and more could still be found in the rest of the continent. Similar phenomena have been observed in Asia, where a notable abundance of newly identified species demonstrates high biodiversity. However, among these species, the prevalence of one or two Beauveria species, typically B. bassiana, tends to overshadow the presence of the newly determined species [53,54].
KEEG analysis of important metabolic pathways, such as the environmental information processing, biosynthesis of secondary metabolites, and lipid and amino acid metabolism, were completed in our genomes. Those pathways are related to the capability of the isolates to develop a pathogenic response and toxin production during the infection process [1]. Previous reports have established that the expression of those genes can be strain differentiated, and as a consequence, can have different environmental nutrition acquisition capabilities [55]. We also analyzed CAZyme genes due to their capabilities in degrading lignocellulose development and the stress response [39]. CAZymes have been reported for B. bassiana genomes and other entomopathogenic fungi [39], and herein, we provide further support for those findings as well as a list of potential genes that emphasize carbohydrate-active enzymes in tropical isolates of B. bassiana. In the case of secondary metabolite precursors, the B. bassiana genome contains 45 SM core genes, including 13 NRPS, 12 PKS, 7 NRPS-like, 1 PKS-like, 3 hybrid NRPS-PKS, and 12 genes related to FAS/terpene/steroid biosynthesis [1]. A reduction in the NRPS and PKS genes in entomopathogenic species that cause systemic infection of host tissues, such as Metarhizium acridum, has been reported by Gao et al. [56]. The reason for the total lack of NRPS and PKS genes in the B31 isolate is unknown, but the differences between this isolate and the other isolates is consistent throughout several of our analyses. C. militaris contains enough genes for the infection of specific hosts, while B. bassiana and M. anisopliae are mostly generalist species and have the same NRPSs as the ARSEF2860 reference and fewer PKS components.
There are no entries for entomopathogenic fungi in the PHI-base [37], therefore we assumed that the proof of pathogenicity/virulence reported in one fungus would also suggest a pathogenicity/virulence function in fungi within our study. Some of the gene families and orthogroups identified showed differences. Hyd1 was present in all isolates, but Hyd2 was absent in five (B13, B26, B27, B43, and B44) out of the eight isolates. Previous studies have shown that Hyd1 has a greater role on virulence than Hyd2; the main phenotype effect of Hyd2, when absent, is reduced surface adhesion [57]. On the other hand, five ABC transporters have been examined in B. bassiana including one B-type, one C-type and three G-type, and only the C-and G-type showed decreased virulence in topical bioassays [58] (Liu et al., 2011). The MrpacC transporter was absent in most of the isolates except for B0 and B1, and this must have been in consideration during the selection of these isolates as potential biocontrollers.
MAT-type genes identified in our isolates indicated that all possess MAT1-2-1 genes except for B31, which contained MAT1-1-1 genes. In both cases, these are the genes present in B. bassiana genomes [1] and a third kind of gene has been reported in B. bassiana isolates, MAT1-2-8 [42]. Regardless of these mating-type genes present in Beauveria genomes, sexual reproduction is infrequent in nature, according to Xiao et al. [1]. The lack of important genes, such as Spo11, which are crucial for meiotic recombination during sexual reproduction, was reported by Xiao et al. [1] as a possible reason as to why asexual reproduction is common in natural populations of B. bassiana. However, Valero-Jiménez et al. [42] described the presence of this gene in five isolates of B. bassiana and all of the genomes described herein also have this gene. Can the preference for asexual reproduction in B. bassiana be an adaptation for prolific geographical expansion, or is it used to enable the infection of a wide range of hosts? Or, as reported by Xiao et al. [1], are transposable elements the main force introducing genetic variation and genome evolution in Beauveria? Furthermore, is this mechanism effective enough to avoid sexual reproduction? These are questions that need to be answered in future studies.
Since allopatric speciation was previously described by Rehner et al. [10] as an important force to introduce genetic differentiation in B. bassiana complex lineages, we analyzed our isolates' unique genes, and found differences between the Puerto Rican isolates (B43 and B44) and other isolates, indicating a possible geographic differentiation. Similarly, continental isolates belonging to different lineages defined by Rehner et al. [10], such as B27 (North American lineage), showed 116 unique genes compared with B0 and B13 of the African and Neotropical lineages. We also observed a major differentiation in B31, whose lineage could not be determined; compared with the other isolates, a total of 533 unique genes were found in B31. High strain diversity has been reported in populations of entomopathogenic fungi around the globe, which indicates that complex interactions between abiotic and biotic factors are also important forces for gaining genetic variability at a local level [7]. According to Valero-Jiménez et al. [42], the Bb8028 isolate has 163 exclusive genes compared with four other isolates used for the control of malaria mosquitoes. Valero-Jiménez et al. attributed this genetic differentiation to the isolate's association with host adaptation due to its high virulence effect in Anopheles vectors. Perhaps, due to B31's association with Costa Rican sugarcane fields, a similar effect is occurring; however, further analysis is needed to investigate this pronounced genetic differentiation.
There are still countries, such as Belize, Guatemala, El Salvador, Honduras, Nicaragua, Panama, Dominican Republic, and Haiti, where there are no genomic studies or the molecular characterization of Beauveria or other entomopathogenic fungi. Studies from Mexico [59,60], Brazil [61] and Costa Rica [14,15,62], have shown high genetic diversity in Beauveria isolates. Untapped Beauveria diversity in the neotropics through genomics is crucial to develop a comprehensive international landscape of the B. bassiana complex and the different species of the genus. The above, including the finding of Cordyceps ancestor species, is needed to map their evolutionary relationships, host, niche and environmental adaptations.