Taxonomic and Functional Composition of Soil Mycobiome of Two Agricultural Sites in Khartoum State, Sudan

Fungi are one of the most diverse groups of organisms and considered as one of the least-explored biodiversity resources. Soil fungal community was investigated in two agricultural sites in Khartoum state, Sudan, during two seasons. A total of 42 soil samples were collected, their physicochemical properties were determined, then subjected to metabarcoding and metagenomic analyses. fungal community composition, diversity and microbial trophic modes were determined utilizing R software packages. From both sites, a total of 15 different phyla were detected, out of them, 11 were the most abundant and frequent. Ascomycota was the dominant phylum (86.54% total abundance), followed by the Basidiomycota (8.29%). The dominant class was Sordariomycetes (41.02%), followed by Dothideomycetes (19.80%). Aspergillus (6.2%), Curvularia (6.0%), Neurospora (5.8%) and Fusarium (4.9%) were the most abundant genera. Deniquelata for the rst time being recorded in Sudan. Apha diversity measures revealed sample richness ranging from 71 to 361 ASVs, and Shannon index ranging from 2.794 to 5.087. The two sites had signicantly different alpha diversity. Land-use types were also signicantly different in their diversity regardless of site. Season had no effect on alpha diversity of soil fungal communities. Beta diversity analysis indicated signicant differences between the two sites and the different land-use types. No signicant differences in the community structure recorded between the two seasons. The dominant trophic mode among the assigned ASVs in soil mycobiome was saprotroph mode (22.11%). Results of this study reveals that fungal community structure is affected by site and land-use type. It gives a comprehensive database for the mycobiome of the agricultural soil in Khartoum state.


Introduction
The soil represents one of the most dynamic and complex ecosystems in the world inhabited by a diverse group of microbial consortia in close association with roots of different plants and are responsible for occurrence of variety of biological and physiological processes (Choudhury and Jain 2012). It is highly needed to expand our knowledge regarding diversity and function of soil microbial community which is a necessary task to alleviate the harmful effects of soil degradation . Also, most of the world's soils are known to lack the nutrients that needed for plant growth. Farmers tend to the extensive use of chemical fertilizers to meet the de ciency of nutrients. Therefore, there is an urgent need to explore the potential of soil microbes for proper nutrient recycling and to recognize alternative, sustainable, environment-friendly options for reducing the use and impacts of synthetic fertilizers (Malav et al. 2015; Prasad et al. 2021).
Fungi are one of the most diverse groups of organisms on earth and are integral ecosystem agents that play pivotal ecological roles as mycorrhizal partners of plants, saprotrophs, and agents of disease. The estimated number of fungal species on earth is about 1.5 million in 2001 and re-estimated to 3.8 million in 2017, whereas the number of described fungal species is 120,000 species only. They are widely distributed in all terrestrial ecosystems; however, distribution of fungal taxa or groups has been poorly documented (Tedersoo et al. 2014; Tedersoo and Nilsson 2016). Fungi are considered as one of the least-explored biodiversity resources of our planet because most of them are uncultivable in the laboratory (Webster and Weber 2007). The number of cultured fungi is believed to be a small fraction of the total number of extant species (Rajendhran and Gunasekaran 2008). Therefore, rapid, and accurate rate of taxonomic and functional identi cation of fungi from complex environments such as soil, water, and tissues of plants and animals is of utmost importance (Tedersoo and Nilsson 2016). Hawksworth and Lücking(2017) in their article that estimating the number of fungi concluded that there were major sources for unrecognized fungal diversity, and the geographic areas and ecological habitats that are largely understudied, particularly in tropical regions and biodiversity hot spots are of the main sources.
Sudan, as many other African countries, also lacking information about fungal diversity particularly soil fungi. Although the importance of agriculture to Sudan, little is known about soil microbial community and microorganisms, particularly fungi. As far as we know, very few investigations on soil myco ora have been conducted previously, with no study using the modern molecular culture-independent approaches.
The rst attempt was done very early by Nour (1955). In which investigated the microscopic soil fungi in ten alkaline soil samples from Khartoum state, Khartoum north with clay content ranging from 15-35 % using conventional culture-based methods. Other studies were conducted in different areas of Sudan viz Gezira, Sennar and White Nile state (Amin and Abdalla 1980;Abdel-Rahim et al. 1983; El- Amin and Saadabi 2007). Also, few studies were conducted to investigate arbuscular myorrhizal fungi (AMF) association with some crops. For instance, a survey was performed in 2009 in the White Nile state, Central Sudan, to assess AM root colonization and AMF spore densities and species richness in nine elds planted with 13 different important crop species (Abdelhalim et al. 2014).
In the last decade, advancement in high-throughput sequencing brought unprecedented growth in understanding the world of fungi through sequencing of targeted metabarcoding marker genes directly obtained from environmental samples. Metabarcoding uses universal PCR primers to mass-amplify a taxonomically informative gene (barcodes) from mass collections of organisms or from environmental DNA (Ji et al. 2013). The importance of metabarcoding in ecology is increasing dramatically, particularly in the ecology of microorganisms that are often di cult to be identi ed by other tools rather than molecular biology (Balint et al. 2014). In addition, metabarcoding is a reliable method for recovering diversity information from large-scale, eld-collected data sets (Ji et al. 2013).
In fungal community ecology, metabarcoding is becoming a very necessary tool (Schmidt et al. 2013 At each site and for each land-use, soil samples were collected from the top 20 cm of the surface soil of three locations (points); ve replicates for each location were taken. Samples of the ve replicates were mixed and pooled to make a composite sample for each collection point. GPS coordinates for each sampling point were recorded. A representative sample of each collection (about 1 Kg) was put in a plastic bag to be used for determination of soil properties. Also, 20 grams of the sample was taken in zip-lock plastic bags and kept cooled till being transferred to the laboratory for DNA extraction.
The sampling was conducted in two seasons, winter, and summer, following the same sampling technique.

Sample preparation:
In the laboratory, roots and rocks were removed from the samples before sieving. The 1 kgsized samples were left to air-dry at room temperature, whereas samples taken for DNA extraction were kept at -20 C˚ until processed.

Metabarcoding and metagenomic analysis:
Each soil sample (250 mg) was used to extract the DNA using Qiagen Dneasy® PowrSoil® DNA extraction kit (Qiagen) according to the manufacturer's instructions. The quality of the extracted DNA was checked using NanoDrop™ 2000c spectrophotometer (Thermo Fisher Scienti c Inc, USA).
To study the soil fungal community, the ITS1 region was ampli ed using forward primer, ITS1FKYO2 (5′-TAGAGGAAGTAAAAGTCGTAA-3′) and the reverse primer ITS2KYO2 (5′ -TTYRCTRCGTTCTTCATC- The PCR program consisted of initial denaturation step at 98 ºC for 30 seconds;35 cycles of: 98 ºC for 10 seconds, 53 ºC for 30 seconds, 72 ºC for 1 minute; then nal extension at 72 ºC for 2 minutes and incubation (in nite hold) at 12 ºC. The PCR products were veri ed using 1.2 % agarose gel in 1% TAE Buffer. An aliquot of 3 μl of each PCR product sample was loaded after mixing with 1 μl of SYBR green dye and 0.5 μl of 6x loading buffer. Then samples were allowed to separate for 20 minutes using 100 volts in MUPID-EXU horizontal electrophoresis system (Gel Company, Inc.). To purify the PCR amplicon from the excess primers, nucleotides, salts and enzymes, amplicons were subjected to high-throughput puri cation using AGENCOURT® AMPURE XP PCR Puri cation system (Beckman Coulter, Inc., Brea, CA). This system utilizes an optimized buffer to selectively bind PCR amplicons 100 bp and larger to paramagnetic beads.
The puri ed PCR products were quanti ed using Invitrogen QubitTM dsDNA HS Assay Kit (Invitrogen, Carlsbad, CA, USA) as described by the manufacturer. Then the tubes were read in the Qubit® 2.0 uorometer (Thermo Fisher Scienti c).
According to the lowest concentration, the library was prepared by mixing different volumes of the puri ed PCR products in DNA LoBind tube (Eppendorf North America, Inc., USA). The size and the quantity of the library were checked using Agilent High Sensitivity DNA Kit (Agilent Technologies) in Agilent 2100 Bioanalyzer integrated with 2100 Expert software.

Metagenomic data processing
The ITS1 raw sequencing data were obtained from the Torrent Server-Torrent Suite™ as a demultiplexed FASTQ les. All analyses were done in QIIME2 version 2019. 10 The data were then analyzed using DADA2 pipeline through the R package DADA2 version 1.16.0 (Callahan et al. 2016). The quality of the reads in the FASTQ data les was inspected, then the data were ltered and trimmed using DADA2 standard ltering parameters.
Chimeras were removed from each ltered read using removeBimeraDenovo function. Samples were checked for total number of reads to be more than 5,000 reads. To enable comparisons, the ASVs counts were standardized by transformation to relative abundance and then multiplication by the median sample read depth using phyloseq package (McMurdie and Holmes 2013). The standardized data were merged at the lowest available taxonomic level or annotation using modi ed tax_glom.kv function of the same package. The merged taxa were ltered by removing taxa that are only present at very low numbers in a small minority of samples, that are present at least 10 counts in at least 20% of samples or that have a total relative abundance of at least 1% of the total number of reads (Lennard). The ltered data were used for further analyses.

Soil fungal community composition and diversity:
To study the fungal diversity within each soil community, the observed taxa (Richness) and Shannon indices were used as alpha diversity (Within-samples) measures and they were measured on unmerged standardized data. The estimate_richness function of phyloseq package was used for this purpose.
Beta diversity (between-samples) was examined on the merged ltered taxonomies using Bray-Curtis dissimilarity measure and the non-metric multidimensional scaling (NMDS) as ordination method using phyloseq package.
Heat maps for the top 50 most abundant taxa were created using unsupervised hierarchical clustering with Bray-Curtis distances for all samples.

Functional composition prediction:
The processed ASVs were used to predict functional communities of the samples. FUNGuild tool (v1.0 Beta) was used to taxonomically parse the fungal ASVs by ecological guild (Nguyen et al. 2016). The tool was used through the python script, provided by the tool developers, that has been run from the Ubuntu 16.04 command line.

Data analysis:
All analyses were performed in R software version 3.6.3 (R Core Team 2020).
The analysis of soil properties was conducted using one-way ANOVA to compare the properties for the different land-use in the same site and two-way ANOVA to compare the properties of landuse types in different sites.
Taxonomic bar plotst o show taxonomic composition (at different levels) in all samples phyloseq package and for each site with the different land-use type were created using phyloseq R package. On the other hand, Tukey's 'Honest Signi cant Difference' (TukeyHSD) test was carried out to determine the statistical differences between different sites and land-use types' communities (alpha diversity measures). Then alpha diversity box plots were plotted using amplicon R package.
To examine how the composition of microbiome communities varies across different land-use types and the two sites, statistical test of signi cance on beta diversity was performed through permutational multivariate analysis of variance (PERMANOVA) using adonis function of vegan package version 2.5-6 (Oksanen et al. 2019).
The homogeneity of dispersion test was performed to estimate the homogeneity of each group regarding the taxonomic composition of their samples.
To determine the ASVs (taxa) that are signi cantly different between between the two sites, super. tZig.kv function modi ed from metagenomeseq's tzig and mrfulltable functions in R were used The following parameters were used to determine signi cance: the ASV should have 0.2% presence across samples, and that to keep only ASVs where at least one of the sites have 20% of the samples positive for that ASV, had a fold change (beta coe cient) of 1.25 and had adjusted p-value of 0.5.

Effect of study factors on soil fungal biomass
To study the effect of different land-use types, site and the two seasons on the fungal biomass (ITS gene copy number), two-way ANOVA was performed in R. Before that, Levene's Test for Homogeneity of Variance for different groups was performed using Levene's Test function in car package.

Results
ITS1 amplicon sequence data generated by Ion Torrent sequencing platform were used to investigate which factor affects soil mycobiome the most. Two seasons (Winter and summer), different land-use types and four sites that have different soil properties in Khartoum state were included as the factors of the study.

Soil properties:
Soil physicochemical properties that were statistically different (P-value < 0.05) between the two sites were: EC, SP%, Clay%, Sand%, Silt%, OC%, OM% and N%. Whereas, the two sites had the same pH, CaCO 3 , C/N and P ( ANOVA test revealed that there were no signi cant differences (P-value =0.885109) between the two sites Shambat and Omdurman in the amount of the fungi. Also, there were no signi cant differences between the different land-use types of the two sites and the two seasons ((Pr(>F) 0.625927 and 0.446575, respectively). However, the interaction between land-use and site indicated highly signi cantly different result (Pr(>F) 0.000102).

Fungal community:
3.3.1 Metagenomic raw data and data processing: The total number of raw reads in all 42 samples (Samples from the two sites, four land-use types and two seasons) was 1436320 reads decreased to 880097 after ltering and denoising and almost to the half (705829 reads) after removing chimeric sequences. Shambat had the highest number of reads, and mango soils had the highest number of reads among all land-use types, followed by sorghum, bare land and onion. These reads were then classi ed into 4403 ASVs based on unite database and they were all classi ed as fungi. Among nonchimeric sequences, the minimum number of reads per sample was 7202 (Recorded for sample Omdurman_G2_L3 collected from Sorghum soil in Omdurman), and the maximum number was 45580 reads (Sample Shambat_G2_L1 collected from Sorghum soil in Shambat). With 16805.45 average reads number and median of 16661.5. All the samples were used in the analysis (after standardization) and no samples were excluded because of the low reads number. Merging the ASVs at the lowest available taxonomic annotation resulted in 677 ASVs, 231 out of them were remained after ltering (Present at 10 counts, in 20% of the samples or at 1% of the total abundance). These ltered data were used in all statistical tests.

Soil Fungal community structure and taxonomic composition:
In all soil samples, the taxonomic composition revealed 15 different fungal phyla that include a total of 39 classes, before merging and ltering. However, 1406 (31.9%) out of the 4403 ASVs could not be identi ed to phylum level. Among those identi ed to phylum level, 2017 ASVs did not assign to certain class. This represented 45.8% of the total number and 67.3% of the ASVs that assigned to certain phylum.
After merging ASVs that had similar taxonomy and then ltering based on abundance and/or frequency, there were 11 fungal phyla (Included 24 classes, 47 orders, 75 families, 123 genera and 117 Species), with number  (Table 3.2). The remaining nine phyla represented 5.17% of the total abundance and included 16.45% of the ASVs (Fig. 3.1). On the other hand, the dominant class of the phylum Ascomycota and among all classes was Sordariomycetes (41.02%), followed by Dothideomycetes (19.80%), while 12.41% of the total abundance recorded by ASVs that did not identi ed to class level (Fig. 3.2).
The most abundant ten taxa belonged to the phylum Ascomycota; ve of them to the class Sordariomycetes, and the other ve belong to: Pezizomycetes (two), Dothideomycetes (one) and Eurotiomycetes (one). The most abundant ASV among them represents sequences that identi ed to phylum level only and it was present in high abundance in all study samples.
Heat map of the top 50 abundant taxa (Fig. 3.4) showed the occurrence of those taxa in different samples was made.
The taxa that appear in hot colors indicates high abundance and those have blue colors indicates low level of abundance. It is clear that, most of these taxa not identi ed to genera and species levels.  (Fig 3.8). Season had no effect on alpha diversity of soil fungal communities. Also, the interaction between site and land use had no effect on the diversity (Observed Pr(>F) 0.258, Shannon Pr(>F)0.080834 ). and site of collection. The interaction between the two factors, land-use and site, indicated signi cant differences (P=0.001, R2 0.14893). The differential abundance between the two sites revealed that there were 83 ASVs signi cantly different between the two sites that met threshold criteria mentioned previously (Fig. 3.10). 21 taxa were found only in Shambat, and only one taxon was found only in Omdurman.

Functional composition:
FunGuild tool made assignments on 156 ASVs out of 231 (75 ASVs were unassigned). The con dence of the assignments was varied. Of these 10.9% were highly probable to be assigned to the correct guild, 60.26% were probable, 28.85 % were possible to be assigned correctly.

Discussion
We used metabarcoding approach to investigate the soil mycobiome of four land-use types in the two sites: Shambat and Omdurman in Khartoum State, Sudan. ITS1 was used as the fungal DNA barcode and sequencing was done using Ion Torrent platform. The total number of ASVs recorded in this study was 4403 ASVs which were represented 677 taxa merged at their lowest available taxonomic annotation.
Of them 231 were abundant and/or frequent taxa. This result suggests high diversity in soil mycobiome of the study area compared to the number of investigated samples.
The signi cance was checked using the following parameters: perc = the ASV should have 0.2% percent presence across samples, and that to keep only ASVs where at least one of the sites have 20% of the samples positive for that ASV, had a fold change (beta coe cient) of 1.25 and had adjusted p-value of 0.5. This study is the rst to determine the whole mycobiome content in Sudan soils using the advanced metagenomic sequencing techniques. To the best of our knowledge and as mentioned above, very few mycological survey studies were conducted previously to survey soil fungi in Sudan as general and in Khartoum state in particular (Nour 1955 investigate myco ora in two soil types of Clay and Sandy that collected from north west Sennar sugar cane plantation. They isolated 24 species from sugar cane rhizosphere, non-rhizosphere and root surfaces. Another study in White Nile state also, 23 species of 16 genera were recorded by El-Amin and Saadabi (2007) from sugarcane rhizosphere in twenty sites covering Kenana sugar Estate.
The number of recorded fungal taxa was varied between these studies and it was very low compared to the present study. Amin and Abdalla (1980) reported that Sudan's soils are generally poor in soil fungi based on the previous investigations. However, we assume that these low numbers of fungal taxa due to the method used in the investigation. As mentioned before, all previous studies were based on conventional culture-based methods that could only detect cultivable fungi which represents very small fraction of soil fungi. On the other hand, both cultivatable and uncultivable fungi could be detected by using metabarcoding approach which may be the reason behind the higher number of fungal taxa recorded in this study.
Also, the primers pair used in this study is characterized by high-coverage across diverse fungal taxa. It has been developed previously by Toju et al. (2012) to improve the coverage across fungi kingdom and they revealed that the designed primers covered 99% of fungal taxa at the species level. Furthermore, it could amplify the sequences of the two phyla Ascomycota and Basidiomycota) without signi cant taxonomic biases and could also cover an ecologically important clade of mycorrhizal fungi, Glomeromycota.
Of the whole ASVs, 31.9% identifed as Fungi only and did not assigned to certain phylum. Also, at class level, 45.8% of the total number of sequences could not be identi ed to class and 67.3% of the ASVs that assigned to phylum failed to be identi ed to class as well. The percentage of unidenti ed ASVs is increasing toward the lowest taxonomic rank such as genera and species. The high percentage of unidenti ed fungi in the soil mycobiome at different levels con rmed that there is a large number of fungi that still not explored. For example, the most abundant ASV (after merging) accounted for 10.67% which represented by unknown sequences that belong to Ascomycota, and they may include different taxa of different importance. Therefore, more studies should be conducted to explore these unknown fungi.
Results of this study also indicated that the dominant fungal phylum in soil community is Ascomycota  Mango had highest and signi cantly different clay content which means high nutrients and minerals.
Therefore, high diversity of mango community may be due to the high clay content. Bare land is lacking nutrients and other factors that related to plant cover such as root exudate which may enhance the microbial growth. Also, and although it is not signi cant from the others, mango and sorghum recorded the highest organic matter content. This may refer to the continuous falling leaves in mango land and the animal feces in sorghum. Other factors related to each land-use type such as root exudates may be responsible for the difference in their diversity.
Beta diversity analysis result revealed that both land-use and site affect the fungal community structure with land-use having the greatest effect that responsible for 20.14% of the total variance, while 10.59% of the variance due to site. The two seasons had similar fungal community structure. Also, the interaction between the two factors site and land-use had affected the fungal community structure as well as interaction between the three factors. This is consistent with our hypothesis that different land-use types and different soil types or properties have different soil communities and consistent with previous studies (Tian et al. 2017). Also, previous studies reported that seasonality had no effect or had small effect on fungal community diversity (Tian et al. 2017).
The quantity and composition of the microbial biomass depend on soil characteristics and the abundance of carbon (C) for energy and cell metabolism. Soil organic carbon (SOC) is the backbone of organic matter, which is the source of energy for most of the soil biota (Hesham et al. 2021). However, it has been reported previously that the quantity and composition of the microbial biomass depend on soil characteristics and the abundance of carbon (C) for energy and cell metabolism ).
The fungal biomass was statistically not signi cant in the two sites and the different land-use types. But Shambat had higher value than Omdurman, and sorghum and mango had also higher values than other land-use types. This is obviously related to the organic matter which is highly needed for microbial growth. Shambat had high organic matter and carbon. Also, sorghum land receives organic matter from grazing in sorghum after cut. While Mango, the permanent land-use type received the organic matter from

Conclusion
Based on soil samples collected from 42 representative sites in two agricultural sites within Khartoum State, Sudan, this study demonstrated the high diversity of soil fungal communities in these ecosystems.
We found that variations in soil fungal community diversity and composition were mainly regulated by soil composition and land-use while season had no effect. Different sites and different land-use types had different alpha diversity, whereas the two seasons had the same diversity.

Declarations
Funding: This study was partly funded by the Joint Research Program of Arid Land Research Center, Tottori University (Grant No. 31GR1001).
Con icts of interest/Competing interests: The authors declare that this research has any potential con ict of interest.
Availability of data and material: Not applicable.
Code availability: Not applicable.
Authors' contributions: All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Marwa H. E. Elnaiem. The research project and the draft of the manuscript were designed by Marmar A. El Siddig and revised by Takeshi Taniguchi. All authors commented on previous versions of the manuscript. All authors read and approved the nal manuscript.     3.9 Beta diversity was studied using Bray-Curtis dissimilarity distance. NMDS (Fig. 3.9, Stress value0.2116419) indicated that samples from each site tended to cluster together irrespectively of the land-use type except for one or two samples in each site. However, fungal communities according to different land-use types in Shambat were separated from each other's, particularly those from bare land soils which clustered apart from others and were grouped with Omdurman samples. The homogeneity of dispersion test revealed signi cant results of fungal composition between land-use groups (F =4.056, P =0.019) which indicates the heterogeneity of each land-use community.