https://orcid.org/0000-0001-7472-8246
Figures
Abstract
The rhizosphere serves as a hub for a variety of microorganisms that are highly beneficial to crop production and improvement of soil health. However, intensive farming practices including utilization of agrochemicals can cause a decline in microbial diversity that could severely compromise soil health and crop productivity. Here we investigated the taxonomic abundance and functional diversity of the microbial communities of sorghum and pearl millet rhizosphere soil samples from sixteen farms in Mpumalanga and Limpopo Provinces of South Africa. Soil samples were collected at the rhizosphere of sorghum and pearl millet crops and pooled into 34 samples. The soil samples were used for 16S rRNA amplicon sequencing analysis, soil physicochemical properties, and community-level physiological profiles. The results indicated that carbon utilization was highest in the majority of soil samples from Jane Furse, which also demonstrated greater microbial richness. The 16S rRNA amplicon sequencing analysis provides insight into the relative abundance of soil microbial communities, where at phylum level Planctomycetes, Proteobacteria, and Actinobacteria were the most predominant in all farms, but their relative abundances varied. Our results revealed that physicochemical properties could affect microbial abundance and diversity. The distance-based redundancy analysis (dbRDA) explained 46.8% of the variation in the soil bacterial community structure, with Mn, Fe, NO₃⁻-N, and Ca identified as the key soil physicochemical variables shaping community composition. Thus, this study may contribute to advancing sustainable agricultural practices by providing baseline data that may inform future bioinoculant development.
Citation: Diale-Makhongela MO, Mpai T, Bopape FL, Mtsweni P, Salawu-Rotimi A, Shargie NG, et al. (2026) 16S rRNA-based metagenomics insights into the microbial diversity and functional attributes of soils from the rhizosphere of selected C4 crops of farms in Mpumalanga and Limpopo provinces, South Africa. PLoS One 21(6): e0347776. https://doi.org/10.1371/journal.pone.0347776
Editor: Ying Ma, Universidade de Coimbra, PORTUGAL
Received: September 5, 2025; Accepted: April 7, 2026; Published: June 15, 2026
Copyright: © 2026 Diale-Makhongela et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying the results presented in this study are available from the NCBI data base library using the accession number PRJNA1269557.
Funding: The funding for this project was obtained from the Department of Agriculture (DoA), Plant Production Directorate of South Africa (funding number A-134, awarded to AIH). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Nowadays, agriculture is facing a major challenge in increasing crop production due to abiotic stresses [1]. Climate change has intensified the impact of drought, salinity, and heat, making them the primary constraints on crop growth and a major threat to global food security [2]. Most arable lands are prone to one or more abiotic stresses, which affect crop yields by 60% on average [3,4]. Alternatively, extensive use of agrochemicals, irregular rainfall patterns and intensive farming, such as tillage and crop management, can alter the soil’s physicochemical properties, including pH, electrical conductivity, and organic matter content, ultimately affecting the soil microbiome and fertility [5,6]. Thus, developing more efficient and sustainable farming practices is essential to protect agricultural soils and ensure future food security.
Sorghum (Sorghum bicolor) and pearl millet (Pennisetum glaucum) are among the most important C4 cereals grown in tropical and subtropical regions [7]. The C4 cereals are staple foods that sustain the population of Sub-Saharan Africa and Asia. They serve as a significant source of energy, essential calories, proteins, and vital micronutrients [8]. Sorghum and pearl millet are among the most drought-resistant cereal crops [9] and thus referred to as climate smart crops. Due to their resilience, these crops offer valuable opportunities for enhanced household food security even under challenging climatic conditions. As a result, smallholder farmers who may lack irrigation facilities and established systems can still produce food for their households and local markets [10,11]. Although these crop species are highly resilient to climate change, low-yield remains a major production challenge [12,13]. Their yields can be significantly reduced due to high-acidity soils, uneven distribution of seasonal rain and poor soil nutrient content [13,14]. A better understanding of plant–microbiome interactions in the rhizosphere soil can unlock new prospects for sustainably enhancing pearl millet and sorghum production in various localities of South Africa.
The rhizosphere serves as a hub for a variety of microorganisms that are highly beneficial to agricultural crop production. These microbial communities are essential for maintaining ecosystem balance, facilitating nutrient cycling, synthesizing hormones that promote crop growth, and enhancing resilience to abiotic stress such as drought, salinity, and extreme temperatures [15]. Furthermore, these microorganisms work in synergy with one another and the crops to improve soil fertility and yield. On the other hand, they are essential indicators to measure soil health [16]. Such microorganisms are crucial for agricultural sustainability and alleviating the effects of climate change.
Understanding the influence, interaction and roles of microbial communities in agricultural environments is essential for promoting sustainable farming practices and its management [17]. Next-generation sequencing such as metagenomics can facilitate a better understanding of microbial communities in the soil and provide genetic information in genomes from both culturable and unculturable organisms [18]. Moreover, the composition and abundance of microbial communities in the soil vary significantly based on soil properties and environmental conditions. In this study, we aimed to use 16S rRNA amplicons to investigate microbial diversity and abundance levels, physicochemical properties and physiological levels of sorghum and pearl millet rhizosphere soil samples from 16 farms in Mpumalanga and Limpopo provinces of South Africa.
Methods and materials
Site description and soil sampling
Rhizosphere soils were collected from sixteen monocroped sorghum and pearl millet farms in Mpumalanga (Standerton) and Limpopo (Jane Furse-Maseleseleng, and Lebowakgomo) provinces, comprising of three pearl millet and thirteen sorghum farms (S1 Table, Fig 1). Sampling was conducted in December 2023. During the sampling period, the average temperature remained at 16–32 °C in Limpopo and 14–26 °C in Mpumalanga, while the average precipitation ranged from 385–800 mm in Limpopo and 570–750 in Mpumalanga [19].
Map of South Africa showing the two provinces, Mpumalanga and Limpopo, where soil samples were collected (A).Map indicating the three sampling locations within Limpopo and Mpumalanga provinces (B). Soil sampling sites in Lebowakgomo (C), Maseleseleng (D) and Standerton (E).
Soil samples were collected from subsistence farms that primarily cultivate sorghum and pearl millet, though they rotate with legumes. The farms used conventional tillage practices to prepare the soil before planting the crops. None of the farms used any fertilizers.
A stratified random sampling approach was used to collect soil samples at a depth of 0–30 cm with a soil auger. The soil samples collected from each farm were then pooled into representative samples as shown in S1 Table. The samples were used to analyze metagenomics, soil physicochemical properties, and community-level physiological profiles.
Soil physicochemical analysis
For physiochemical analysis, soil samples were sent to the Agricultural Research Council-Soil Climate and Water Analytical Services Division. The physicochemical properties were performed using procedures described in the handbook of standard soil testing methods [20]. Exchangeable cations including calcium (Ca), magnesium (Mg), sodium (Na), and potassium (K) were extracted using 0.1 M ammonium acetate and quantified with an inductively coupled plasma optical emission spectrometer (ICP-OES). Copper (Cu), manganese (Mn), iron (Fe) and zinc (Zn) were extracted with 0.1 M HCl, while aluminium (Al) was extracted in 1.0 KCL and analyzed using ICP-OES. Soil texture (sand, silt, and clay fractions) was determined using the hydrometer method following the Bouyoucos procedure. Available phosphorus (P) was assessed using the Bray 1 extraction method and measured with a flow analyzer. Soil pH was determined in a 1:10 soil-to-water suspension using a calibrated pH meter. Nitrate (NO₃⁻) concentrations were obtained by extraction with 1 M KCl, followed by quantification using a flow analyzer.
Community-level physiological profiles of rhizosphere soil microbial communities
The microbial community level physiological profiles (CLPP) were analyzed using BIOLOG™ EcoPlates. The plates consist of 96 wells of which 31 wells contain different carbon substrates: 1) carbohydrates: Pyruvic acid Methyl Ester, D-Cellobios, α-D-Lactose, β-Methyl-D-Glucoside, D-Xylose, i-Erythritol, D-Mannitol, N-Acetyl-D-glucoseamine, Glucose-1-phosphate, D,L-αGlycerol phosphate (2) carboxylic acids: D-glucoseaminic acid, D-Galactonic acid γ-Lactone, D-Galacturonic acid, 2-Hydroxy Benzoic acid, 4-Hydroxy Benzoic acid, γ-Amino Butyric acid, Itaconic acid, α-Keto Butyric acid,D-Malic acid, (3) amino acids: L-threonine, L-phenylalanine, glycyl-L-glutamic acid, L-asparagine, L-arginine, L-serine; (4) polymers: α-cyclodextrin, glycogen, tween 80, tween 40; (5) amines/amides: phenylethylamine, putrescine and one well acting as blank control, all in triplicate. Ten grams of the pooled soil samples were suspended in sterile saline, followed by agitating for 20 minutes at 28°C. The samples were serially diluted up to 10 ⁻ ³, and a 120 µl aliquot of the 10 ⁻ ³ soil solution was inoculated into each well of the EcoPlate. The plates were incubated in the dark for 96 hours at 28°C, after which the absorbance was measured at 590 nm using a Chromate Microplate Reader. Microbial response in each microplate was expressed as average well-color development (AWCD) and Shannon–Weaver (H’) indices. The AWCD index was determined using the following formula:
Where c = absorbance of the sample, n = number of wells, and R = absorbance in the control well.
The Shannon–Weaver (H’) index was calculated using the following formula:,
Where pi is the ratio of the activity on each substrate to the sum of activities on all substrates.
DNA extraction and 16S rDNA amplicon sequencing
Genomic DNA was extracted from 34 pooled samples using the ZymoBIOMICS DNA Miniprep kit according to the manufacturer’s protocol. The quality and concentration of genomic DNA were determined using a Qubit™ 4 Fluorometer (Thermo Fisher Scientific™). The libraries were prepared following the Illumina 16S Library preparation protocol. Briefly, the V3 and V4 regions of the bacterial 16s rRNA gene were amplified using the KAPA HiFi Hot Start Master Mix PCR kit and universal primers: forward (5’
TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG) and reverse (5’ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC) [21]. The size of the PCR product was verified by using a Bioanalyzer DNA 1000 chip; and AMPure XP beads were used to purify the 16S V3 and V4 amplicon. Nextera XT Index kit was used to attach dual indices and Illumina sequencing adapters. A second clean-up of the libraries was performed using AMPure XP beads. The libraries were pooled, denatured and sequenced using the Illumina MiSeq, generating 300 bp paired-end.
16S rRNA amplicon sequence data analysis
The raw data for all samples were analyzed in R software (version 4.0.3). The quality of the raw sequencing reads was evaluated using FastQC v0.11.9 [22], and MultiQC v1.12 [23] was utilized to obtain an overview of all the quality control results from FastQC. To improve read quality, Trimmomatic v0.39 [24] was used for the removal of adapter sequences and low-quality bases. Sequenced reads were processed and filtered using the QIIME2 pipeline [25]; DADA2 [26] was utilized for read filtering, error correction, and the identification of Amplicon Sequence Variants (ASVs). To eliminate low-quality bases, the forward reads were truncated to 270 base pairs and reverse reads to 200 base pairs, at the 3’-end, while no trimming was performed at the 5’ end of either forward or reverse reads. Chimera removal was conducted using DADA2’s default settings during the denoising process. The resulting ASV feature table and representative sequences were then exported for downstream analysis using the SILVA v138 classifier [27] with a 99% similarity threshold for bacterial and archaeal species identification.
The phylum-level relative abundance plots were generated directly in phyloseq [28] using the processed ASV feature table and taxonomy table exported from QIIME2. The QIIME2 outputs were imported into R and merged into a single phyloseq object comprising the ASV table, taxonomy table, and sample metadata. Amplicon sequence variants were agglomerated to the phylum level using standard Phyloseq functions (tax_glom, transform_sample_counts) and plots were created using ggplot2 with Phyloseq-formatted output.
Genus level was visualized using a heatmap constructed on a Galaxy server (https://usegalaxy.org/). Quality control of the raw sequence reads was assessed using FastQC (version 0.12.1) [22], followed by trimming with the Trimmomatic program (Version 0.39), [24]. Relative abundance of the 21 most abundant bacterial genera was used to construct a heatmap using heatmap.2 program. Alpha diversity metrics, and beta-diversity metrics were calculated using the Phyloseq and Vegan packages in Rstudio. Alpha diversity was calculated using Observed OTUs, ACE, Chao1, Shannon, Simpson, Inverse Simpson, and Fisher indices to show both species richness and evenness.
Microbial community dissimilarities among samples were calculated using Bray–Curtis distances based on the ASV abundance table in R/RStudio using the vegan package (v2.6-4). Differences in microbial community composition among groups were assessed using permutational multivariate analysis of variance (PERMANOVA) implemented in the adonis2 function with 999 permutations. Crop type (sorghum vs pearl millet) and sampling location were included as explanatory variables to evaluate their relative contributions to microbial community structure.
To identify microbial taxa potentially associated with sorghum and pearl millet rhizospheres, a Random Forest classification analysis was performed using the randomForest package in R/RStudio. The model was trained using the ASV abundance matrix as predictors and crop type as the response variable. A total of 1000 trees were grown, and the number of variables tried at each split was determined automatically by the algorithm. Taxa importance was assessed using the mean decrease in classification accuracy, and the most influential taxa contributing to discrimination between crop types were visualized.
Correlation between environmental variables and community composition was examined using Mantel tests [29], BIO-ENV (bioenv) analysis [30], and distance-based redundancy analysis (dbRDA) [31], implemented via the vegan package [32]. Spearman correlation analysis was conducted using the Hmisc package.
Statistical analysis on community level physiological profile data
A one-way analysis of variance (ANOVA) was carried out on the CLPP data sets using the general linear model procedure (PROC GLM) in SAS version 9.4 statistical software [33]. The residuals were examined for deviations from normality and outliers causing skewness were removed. Tukey’s Honestly Significant Difference (HSD) Test was used to group means of significant effects (α = 0.05). The data was also subjected to multivariate methods like principal component analysis (PCA) and agglomerative hierarchical clustering using XLSTAT [34] software to visualize and elucidate the relationships between the samples and their attributes. Pearson correlation matrix analysis was conducted using XLSTAT to determine the relationships among soil physicochemical properties, carbon sources, and bacterial phyla [34].
Results
Soil physicochemical properties
The results of the soil physicochemical properties among the sixteen farms are shown in S2-S4 Tables. The Mineral levels varied among the farms. Soil samples from the Standerton farms showed higher concentrations of the minerals (Fe, Mn, NO3, P, Zn, Ca, and K) compared to the Lebowakgomo and Jane Furse (Maseleseleng farms). Most of the Lebowakgomo farms showed higher levels of Cu, Na, Mg, and CEC, compared to Standerton and Jane Furse farms (S2-S4 Tables). Soils at Standerton farms are sandy clay to sandy clay loam with clay content ranging from 16–36% and dominant sand content ranging from 38–72%. Maseleseleng farms soils are consistently sandy loam, characterized by high sand (74–88%) and low clay (10–18%) and silt contents (2–8%). Lebowakgomo farms soils are predominantly sandy clay to sandy clay loam, with moderate clay (22–40%) and sand (44–68%) proportions. In general, the sand content was higher in all farms. Furthermore, the content of the silt was relatively low in all three localities (S2-S4 Tables). The pH levels in all sixteen farms were comparable, ranging from 5.34 to 6.54.
Community-level physiological profiles (CLPP) of microorganisms
The metabolic activity of the soil microbial population on the five carbon sources showed significant variation across the farms (Table 1). Higher utilization of carbon substrates in amino acids, amines, carboxylic acids, and polymers was observed in the Jane Furse (Maseleseleng farms), while lower utilization of the carbon source amines was observed in most of the Lebowakgomo farms. The utilization of carbohydrates varied among the farms with Lebo-Mil3B and Stand-Mai3B soil samples showing a higher utilization of carbohydrates. To further analyze the catabolic diversity among the farms, Richness (S), Shannon-Weaver (H), Evenness (E) and AWCD were calculated (Table 1). There were statistically significant differences in the soil AWCD index among the farm sites, p < 0.0001. Stand-Mai3B recorded the highest AWCD while Stand-Sor1A recorded the lowest AWCD. However, most farms from Jane Furse have recorded statistically higher AWCD values. Statistically significant differences in the values of Shannon-Weaver diversity index and Richness were recorded among the farms. The Shannon index ranged from 1.80 to 3.28, with the highest value recorded from the Standerton farm Stand-Mai3B soil sample. The microbial richness ranged from 4.67 to 27, with Stand-Mai 3B having the highest microbial richness and Stand-Sor1A having the lowest microbial richness. There was no difference in the evenness across all farms, except for Math-Sor1C, Stand-Sor1A, Stand-Sor1D, Math-Sor1C and Mase-Sor3E. A dissimilarity dendrogram was constructed according to the Ward method. The dendrogram revealed two major clusters of soil samples, indicating substantial differences in microbial community metabolic activity profiles (S1 Fig). The first cluster (blue) shows that the soils from the farms have similar carbon utilization patterns as compared to the second cluster (red). The second cluster (red) showed distinct carbon utilization patterns among the farms.
Microbial distribution across the soil samples at phylum level
The bacterial community among the farms varies, Planctomycetes and Proteobacteria were dominant in most samples, while other phyla like Chloroflexi, Actinobacteria and TM7 were present in lower but variable proportions (Fig 2). Groups <1% represent low-abundance phyla or taxa that collectively make up less than 1% of the community in each sample. The heatmap showed that the most abundant phylum in some of the farms includes Planctomycetes, Proteobacteria, and Actinobacteria (Fig 3). Planctomycetes phyla was more observed in samples Stand-Sor1B, Stand-Sor1C, Stand-Sor1D, Stand-Sor2B, Mase-Sor1A, Math-Sor1B, Lebo-Mill2B and Lebo-Mill3A. Proteobacteria, Actinobacteria and Gemmatimonodetes were predominant in samples Mase-Sor3B, Lebo-Mill3B, Lebo-Sor7, Lebo-Mill8, Lebo-Sor9 and Mase-Sor3E.
Keynotes: TSC2-S1-Stand-Sor1A, TSC2-S2-Stand-Sor1B, TSC2-S3-Stand-Sor1C, TSC2-S4-Stand-Sor1D, TSC2-S5-Stand-Sor2A, TSC2-S6-Stand-Sor2B, TSC2-S7-Stand-Mai3A, TSC2-S8-Stand-Mai3B, TSC2-S9-Mase-Sor1A, TSC2-S10-Mase-Sor2A, TSC2-S11-Mase-Sor2B, TSC2-S12-Mase-Sor2C, TSC2-S13-Mase-Sor2D, TSC2-S14-Mase-Sor2E, TSC2-S15-Mase-Sor2F, TSC2-S16-Mase-Sor3A, TSC2-S17-Mase-Sor3B, TSC2-S18-Mase-Sor3C, TSC2-S19-Mase-Sor3D, TSC2-S20-Math-Sor1A, TSC2-S21-Math-Sor1B, TSC2-S22-Math-Sor1C, TSC2-S23-Lebo-Sor1, TSC2-S24-Lebo-Mill2A, TSC2-S25-Lebo-Mill2B, TSC2-S26-Lebo-Mill3A, TSC2-S27-Lebo-Mill3B, TSC2-S28-Lebo-Sor4, TSC2-S29-Lebo-Sor5, TSC2-S30-Lebo-Sor6, TSC2-S31-Lebo-Sor7, TSC2-S32-Lebo-Mill8, TSC2-S33-Lebo-Sor9, TSC2-S34-Mase-Sor3E.
Microbial distribution across the soil samples at genus level
The most predominant genera at the genus level were RB41, Chthoniobacter, Candidatus Udaeobacter and Pirellula in most farms (Fig 4). The RB41 genus was identified in all farms except for one farm site (Lebo-Mill3B), with higher abundance observed in Lebo-Mill3A (35.39%) and Mase-Sor2A (30.26%). The genus Chthoniobacter showed a higher abundance in soil samples from Lebo-Mill2A (18.23%), Lebo-Sor5 (16.80%), Lebo-Sor4 (14.98%) and Stand-Sor2A (13.77%) farm sites. Candidatus Udaeobacter was more predominant in sample Mase-Sor2E (20.41%) while Pirellula was more predominant in sample Mase-Sor1A (14.69%). Mycobacterium, Haliangium, and Nocardioides genus were observed in low abundance levels in all the farms.
Microbial distribution across the location at phylum level
The soil samples were analysed at the phylum level based on their collection sites in Limpopo (Jane Furse, Lebowakgomo) and Mpumalanga (Standerton). Jane Furse recorded the highest total microbial abundance compared to Lebowakgomo and Standerton (Fig 5). Planctomycetes and Proteobacteria were more predominant in Jane Furse, while Actinobacteria, Acidobacteria and Chlorofexi also contributed significantly. In Lebowakgomo, microbial abundance were slightly lower than in Jane Furse. Moreover, Standerton soils exhibited the lowest total microbial abundance, although it shared the same dominant phyla as Jane Furse and Lebowakgomo. Acidobacteria was more dominant in Jane Furse and Lebowakgomo. Proteobacteria remained consistent across all sites but is relatively less dominant in Standerton. Variations in relative abundances indicate that environmental or geographical factors may influence the microbial ecosystems at each site.
Microbial distribution across soil samples from the two provinces
The microbial community in Limpopo has a significantly higher relative abundance compared to Mpumalanga. Planctomycetes and Proteobacteria were found to be the most dominant in both provinces. Other phyla, such as Acidobacteria, Actinobacteria and Chlorofexi also contributed noticeably to both provinces (S2 Fig). Though both provinces shared many of the same phyla, the relative abundance of each phylum varied significantly between the two regions.
Alpha diversity by province and location
Alpha diversity was analyzed in terms of observed OTUs, ACE, Chao1, Shannon, Simpson, InvSimpson and Fisher to reflect the diversity and abundance of microbial communities within the three locations (Jane Furse, Lebowakgomo and Standerton) (Fig 6). Jane Furse consistently showed lower diversity measures, and Lebowakgomo’s diversity measures were moderate and showed less variation compared to Standerton. Species richness (ACE, Chao1) of Lebowakgomo is similar to Jane Furse, but diversity metrics like Shannon and Inverse Simpson were slightly higher. Kruskal–Wallis tests indicated no statistically significant differences in alpha diversity measures between the three sites (p > 0.34). This suggests that, while descriptive patterns are visible in the boxplots, these differences are not statistically supported. S3 Fig illustrates the alpha diversity metrics for the provinces, Limpopo and Mpumalanga. Across all measures, Mpumalanga consistently exhibited greater microbial diversity than Limpopo. This may indicate a more heterogeneous environment in Mpumalanga influenced by environmental factors that promote higher species richness and evenness.
Beta diversity by location sites
Beta diversity analysis was used to explore the dissimilarity among microbial communities from soil samples collected from Jane Furse, Lebowakgomo and Standerton. The non-multidimensional scaling (NMDS) analysis revealed that samples from the same location clustered together, indicating that microbial community composition is more similar within locations (Fig 7). The overlap between the ellipses for Lebowakgomo and the other two sites suggests shared microbial communities or higher variability in this location. Standerton showed greater dispersion, indicating higher variability in microbial communities. Jane Furse and Lebowakgomo showed less spread, indicating more consistent microbial composition.
The PERMANOVA (adonis2) analysis was used to test the influence of province and location site on microbial community composition, using Bray-Curtis distances to measure dissimilarity. The PERMANOVA results indicate that both province and location significantly influence microbial community composition, with location explaining a larger portion of variance (7.38%) compared to province (3.68%). Both results were statistically significant, with p-values of 0.003 and 0.001, respectively. To determine whether host crop influenced rhizosphere microbial communities, A PERMANOVA analysis based on Bray–Curtis dissimilarity was performed. The analysis showed that crop type did not significantly influence rhizosphere bacterial community composition (R² = 0.031, p = 0.286) (S4 Fig). Random Forest analysis identified several taxa contributing to discrimination between sorghum and pearl millet rhizosphere samples, including Chloroflexi, Balneimonas, Ellin6075, TM7, and Solibacterales. However, the importance values were generally low, indicating limited discriminatory power between the two crop rhizospheres (S5 Fig).
All alpha diversity metrics were calculated on rarefied ASV tables to ensure equal sequencing depth across samples. No significant differences were detected among the three sites (Kruskal–Wallis, p > 0.34). To assess the influence of soil physicochemical properties within-sample variation in diversity, Spearman rank correlations were performed between alpha diversity indices and soil variables. Only weak associations were observed, and none remained significant after multiple-test correction, indicating that soil chemistry did not strongly influence within-sample microbial diversity (Fig 8).
However, the beta diversity analyses revealed clear relationships between physiochemical properties and community composition. The Mantel test demonstrated a significant correlation between community dissimilarity and overall soil physicochemistry (r = 0.228, p = 0.0057), demonstrating that multivariate changes in soil properties are linked to changes in bacterial community composition. BIOENV analysis identified a subset of variables (Cu, Fe, Mn, Silt, Zn, Ca) that best explained Bray–Curtis dissimilarities (p = 0.34). Using dbRDA, soil physicochemical variables collectively explained 46.8% of the total variation in community composition, indicating a substantial environmental effect. Among all the measured factors, Mn, Fe, NO₃-N and Ca were significant drivers of beta diversity (p < 0.05) (Fig 9). Other factors including pH, K, Na, Cu, Mg, Zn and soil texture were not significantly correlated with the bacteria community composition.
Correlation between bacterial phyla, soil physicochemical properties and carbon sources
The correlation analysis among the carbon sources revealed several significant relationships (P < 0.05, P < 0.001) (S1 File). Carbohydrates showed a significant negative correlation with Na, while carboxylic acids, amino acids, amines, and polymers were negatively with Mg and Na (p < 0.05). Polymers and amines negatively correlated with CEC (p < 0.001). All carbon sources, excluding carbohydrates, showed a strongly significant positive correlation with sand soil texture and a highly significant negative correlation with clay and silt (P < 0.05).
The correlation analysis revealed various significant relationships between bacterial phyla and soil physicochemical properties as well as carbon sources (p < 0.05) (S1 File). Acidobacteria showed a negative correlation with Na, while Proteobacteria showed a positive correlation with Na. Actinobacteria were negatively correlated with Mn and K, while Chlorobi showed a positive correlation with K. Armatimonadetes and WS6 negatively correlated with soil pH. Chlorobi positively correlated with K, while Gemmatimonadetes negatively correlated with K. OP11 was negatively correlated with pH but positively correlated with K. Bacteroidetes showed positive correlations with Al and NO₃-N. TM7 showed positive correlations with Cu, Mn, K, and CEC while Cyanobacteria were positively correlated with Cu, Mn, and CEC. In addition, unclassified bacterial groups were positively correlated with K but negatively correlated with Ca, Mg, and CEC.
Firmicutes showed negative correlations with carbohydrates, carboxylic acids, amino acids, and amines. In contrast, Nitrospirae were positively correlated with carbohydrates, carboxylic acids, amino acids, and amines, indicating opposite responses to these substrates. Similarly, OD1 showed negative correlations with carboxylic acids and amines. Soil texture also influenced certain phyla. Chloroflexi showed negative correlations with Mn, Ca, Mg, and CEC, but positive correlations with carboxylic acids, amino acids, polymers, and amines. Chloroflexi was also influenced by soil texture, the phyla negatively correlated with clay and silt and strongly correlated with sand.
Discussion
Microorganisms play a crucial role in the agricultural ecosystem and productivity. They perform various functions such as regulating soil properties and fertility, nutrient cycling, inhibiting pathogens, enhancing plant growth and resistance to abiotic stresses, such as drought and heavy metal pollution [35,36]. In this study, the soil’s physicochemical properties, microbial community level, and 16S rRNA metagenomics provided insight into the health and microbial diversity of the soil among farms in Standerton, Jane Furse, and Lebowakgomo.
The soil physicochemical properties varied among the farms, with pH ranging from 5.3 to 6.4. While the metabolic activity of the soil microbial population on the five carbon sources showed significant variation across the farms, with some farms showing higher microbial richness compared to others. Previous studies reported that acidic and alkaline soils usually exhibit lower microbial richness compared to soils with slightly acidic to neutral pH [37,38]. Our study revealed that farms with slightly acidic soil pH (5.67–6.0), including Stand-Mai3B, Mase-Sor2A, Mase-Sor3D, Lebo-Mill3B, Math-Sor1A, and Stand-Sor2B, exhibited the highest microbial richness. However, some farms within this pH range (Stand-Sor1A, Stand-Sor1B, Math-Sor1B and Math-Sor1C) showed lower microbial richness levels. This variation may be influenced by other physicochemical factors such as total soil organic carbon, nitrogen, phosphorus, and soil type [39]. The soil microbial communities utilized various carbon substrates at different rates. This variation may be due to differences in their structural composition and associated functional guilds within each community. However, carbohydrates were the most utilized carbon source within the samples. Similar results on the utilization of carbohydrates by soil microorganisms were reported by other studies [40–42]. In addition, Jane Furse (Maseleseleng) soils showed higher utilization of amino acids, amines, carboxylic acids, and polymers. This suggests that the microbial communities in these soils are actively involved in nutrient cycling and the decomposition of organic matter through processes such as protein degradation, deamination, and the breakdown of complex organic compounds [43,44]. This functional activity aligns with the higher AWCD values and higher microbial richness observed in most Jane Furse farms, suggesting that these soils hosts a diverse community of microorganisms which can utilize a wide range of carbon sources. A rich microbial community can improve soil functions such as water retention, organic matter decomposition and nutrient cycling, making the soil more resilient to environmental stresses like drought and diseases [45].
Microbial abundance and diversity
It was observed that the rhizosphere soils of all sorghum and pearl millet farms were dominated by the same microbial phylum. The most dominant phyla in the farms were Planctomycetes and Proteobacteria, Actinobacteria and Acidobacteria, which is in agreement with previous studies. A study by Kumar et al. [46] reported Proteobacteria as the most predominant phyla in sorghum rhizosphere soil. Proteobacteria include a wide variety of bacterial groups with versatile metabolic capabilities, enabling them to survive and adapt in various environmental conditions [46]. A previous study reporting on different breed lines of pearl millet rhizosphere diversity found Actinobacteria and Proteobacteria to be the most dominant phyla [13]. Planctomycetes was found to be one of the dominant phyla in the maize rhizosphere [47].
Proteobacteria have been reported to play a key role in numerous processes, including nitrogen, carbon and sulfur cycling, which are essential for nutrient availability and uptake in plants. Based on known literature, Acitinobacteria are potential bioinoculants/biofertilizers due to their ability to solubilize phosphorus, zinc, and potassium, as well as their production of iron-chelating agents and plant-growth-promoting hormones such as indole acetic acid [45,47]. In addition, Planctomycetes are reported to have potential in contributing significantly to soil nutrient degradation, such as nitrogen, potassium and phosphorus, and are involved in facilitating anaerobic ammonium oxidation [48]. At the genus level, RB41, belonging to the phylum Acidobacteria, was the predominant genus across all soil samples, although its relative abundance varied (Fig 4). The RB41 genus is recognized as a rhizosphere-associated bacterium; which, based on literature is known to play a significant role in maintaining soil metabolic stability and biogeochemical cycling under hostile environmental conditions [49]. These microorganisms have a potential of utilizing carbon sources and nutrients in the soil and participate in the soil carbon cycle [50].
The alpha diversity analysis across the three locations (Jane Furse, Lebowakgomo, and Standerton) revealed no significant differences in microbial richness or diversity. Although Standerton displayed slightly higher values, the large variance within samples resulted in non-significant statistical outcomes (P > 0.34). This indicates that microbial communities within the three locations share similar diversity and the observed differences may reflect normal environmental variability. Additionally, the amount of organic matter, and its composition may influence both the microbial abundance and their diversity [51,52]. The spearman correlation analysis further supported these results, showing that the number of bacterial richness and evenness were similar across all three study sites. Diversity within samples was not driven by the soil physicochemical properties. Previous studies have shown that when soil samples have similar environmental properties, bacterial richness and evenness frequently show limited variation in bacterial diversity [53].
The PERMANOVA results demonstrated that both province and location significantly influenced microbial community composition, with location explaining a greater proportion of the variance (7.38%) than province (3.68%). Despite the significant effects of province and location, a substantial proportion of the variance remained unexplained, suggesting that additional environmental and agronomic variables may contribute to microbial community dynamics. These may include factors such as soil moisture at the time of sampling, crop rotation practices, and tillage intensity [54,55].
The PERMANOVA analysis showed that sorghum vs. pearl millet did not significantly influence rhizosphere bacterial community composition. Sorghum and pearl millet are both C4 grasses belonging to the Poaceae family and they share similar root architectures and patterns of rhizodeposition [56–58]. Root exudates are known to play an important role in shaping rhizosphere microbial communities by providing carbon sources and signalling compounds that influence microbial recruitment [59]. Due to these physiological similarities, the two crops may recruit comparable rhizosphere microbial assemblages, resulting in the limited host-associated differences observed in this study. In contrast, site-associated environmental factors appeared to exert a stronger influence on microbial community composition, as demonstrated by the significant PERMANOVA effect of location. This suggests that soil physicochemical conditions and local environmental factors may play a more dominant role than host identity in structuring rhizosphere bacterial communities in these agricultural systems.
Random Forest analysis identified several taxa contributing to discrimination between sorghum and pearl millet rhizosphere samples, including members of Chloroflexi, Balneimonas, Ellin6075, TM7, and Solibacterales. However, the importance values were generally low, indicating limited discriminatory power between the two crop rhizospheres, which is consistent with the PERMANOVA results showing no significant crop-type effect on overall microbial community composition.
The dbRDA indicated that Mn, Fe, NO₃-N and Ca were the key factors driving the variations in the bacterial community structure which is similar to previous studies [60,61]. The observed variation in Mn, Fe, NO₃-N, and Ca concentrations may result from both crop rotation and inherent soil properties such as soil texture, parent material and mineral composition. Sorghum and Pearl millet is reported to be rich in iron [62] therefore, the release of iron-enriched root exudates may contribute to the elevated soil iron concentrations observed in this study. The farms rotate the C4 crops with legumes. The rotation of legumes may contribute to the increased NO₃-N levels through biological nitrogen fixation process which explains the elevated NO₃-N levels observed in the soils. Furthermore, decomposition of the rotational legume residues can increase nitrate availability in the soil [63,64]. In addition, incorporating crop residues through conventional tillage can facilitate the gradual decomposition of organic matter, which mobilizes Fe and enhances its availability in the soil [65].
Soil pH is considered a key driver of soil bacterial community structure. However, in this study, pH was not significantly correlated with the bacterial community composition, which is consistent with previous studies [66]. Similarly, Zhang et al. [67] reported that soil bacterial community composition is altered more by soil nutrient availability than pH.
Correlation analysis of bacterial phyla, soil physicochemical properties and carbon sources
The correlation analysis revealed that several bacterial phyla were strongly associated with soil physicochemical properties and available carbon sources, highlighting the role of nutrient availability and soil chemistry in shaping microbial community structure [68, 69]. Acidobacteria negatively correlated with Na, whereas proteobacteria showed a positive correlation which is consistent with other studies [70–72]. Chloroflexi showed a negative correlation with clay and silt but a positive correlation with sand, contrasting with other studies that reported a positive correlation with clay and silt [73].
Most phyla were positively correlated with specific carbon sources while other phyla, such as Firmicutes and OD1 revealed a negative correlation. This further emphasize that carbon availability significantly influences the bacterial community composition in agricultural soils [68]. Similarly, previous studies have shown that carbon sources such as carbohydrates, amino acids, and carboxylic acids are key drivers of microbial metabolic activity and community structure [74].
Conclusion
This study investigated the bacterial microbiome diversity and abundance in 34 soil samples pooled from sixteen farms in Limpopo and Mpumalanga. Overall, results revealed that physicochemical properties play a key role in shaping the microbial community systems in the soil. A variation in microbial richness was observed throughout the farms. The 16S rRNA amplicon sequencing results revealed that the most predominant phylum in all localities are Planctomycetes, Proteobacteria, Actinobacteria, Acidobacteria and Chlorofexi, which are known to improve soil fertility and crop production. Carbon utilization, soil physicochemical properties and soil microbial community structure are crucial indicators of a healthy soil. Our study reveals new insight into microbial carbon utilization, physicochemical properties, microbial richness, and diversity from sorghum and pearl millet farms in Limpopo and Mpumalanga provinces, which might help improve the sustainability of agriculture in these and similar areas.
Supporting information
S1 File. Pearson correlation matrix relationship among soil physicochemical, carbon sources and bacterial phyla.
https://doi.org/10.1371/journal.pone.0347776.s001
(XLSX)
S1 Fig. Dendrogram illustrating the dissimilarity of the carbon utilization patterns between 34 soil sample.
https://doi.org/10.1371/journal.pone.0347776.s002
(DOCX)
S2 Fig. Comparison of microbial phylum-Level distribution between Province (Limpopo and Mpumalanga).
https://doi.org/10.1371/journal.pone.0347776.s003
(TIFF)
S3 Fig. Alpha diversity measures between Limpopo and Mpumalanga.
https://doi.org/10.1371/journal.pone.0347776.s004
(TIFF)
S4 Fig. Non-metric multidimensional scaling (NMDS) ordination of rhizosphere bacterial communities based on Bray– Curtis dissimilarity, colored by crop type.
https://doi.org/10.1371/journal.pone.0347776.s005
(TIFF)
S5 Fig. Random Forest importance plot showing microbial taxa contributing to discrimination between sorghum and pearl millet rhizosphere samples.
Importance values represent the mean decrease in classification accuracy when each taxon is removed from the model.
https://doi.org/10.1371/journal.pone.0347776.s006
(TIFF)
S1 Table. Farm sites where soil samples were collected in the two provinces.
https://doi.org/10.1371/journal.pone.0347776.s007
(DOCX)
S2 Table. Physical and chemical properties of soil samples from Standerton farms.
https://doi.org/10.1371/journal.pone.0347776.s008
(DOCX)
S3 Table. Physical and chemical properties of soil samples from Jane Furse farms.
https://doi.org/10.1371/journal.pone.0347776.s009
(DOCX)
S4 Table. Physical and chemical properties of soil samples from Lebowakgomo farms.
https://doi.org/10.1371/journal.pone.0347776.s010
(DOCX)
Acknowledgments
Due acknowledgment goes to the National Research Foundation (NRF) of South Africa for granting Postdoctoral Fellowships to Mamonokane Olga Diale-Makhongela and Tiisetso Mpai under the ARC-PDP block grant for postgraduate programme (PDP/2023/04/03/02). We also extend our acknowledgement to the extension officers and farmers from Mpumalanga and Limpopo, for allowing the survey and the opportunity to sample soils from the farms.
References
- 1. Palmgren M, Shabala S. Adapting crops for climate change: regaining lost abiotic stress tolerance in crops. Front Sci. 2024;2:1416023.
- 2. Fahad S, Bajwa AA, Nazir U, Anjum SA, Farooq A, Zohaib A, et al. Crop Production under Drought and Heat Stress: Plant Responses and Management Options. Front Plant Sci. 2017;8:1147. pmid:28706531
- 3. Kopecká R, Kameniarová M, Černý M, Brzobohatý B, Novák J. Abiotic Stress in Crop Production. Int J Mol Sci. 2023;24(7):6603.
- 4. Waqas MA, Kaya C, Riaz A, Farooq M, Nawaz I, Wilkes A, et al. Potential Mechanisms of Abiotic Stress Tolerance in Crop Plants Induced by Thiourea. Front Plant Sci. 2019;10:1336. pmid:31736993
- 5. Chávez-Romero Y, Navarro-Noya YE, Reynoso-Martínez SC, Sarria-Guzmán Y, Govaerts B, Verhulst N, et al. 16S metagenomics reveals changes in the soil bacterial community driven by soil organic C, N-fertilizer and tillage-crop residue management. Soil Tillage Res. 2016;159:1–8.
- 6. Al-Shammary AAG, Al-Shihmani LSS, Fernández-Gálvez J, Caballero-Calvo A. Optimizing sustainable agriculture: A comprehensive review of agronomic practices and their impacts on soil attributes. J Environ Manage. 2024;364:121487. pmid:38889650
- 7. Choudhary S, Guha A, Kholova J, Pandravada A, Messina CD, Cooper M, et al. Maize, sorghum, and pearl millet have highly contrasting species strategies to adapt to water stress and climate change-like conditions. Plant Sci. 2020;295:110297. pmid:32534623
- 8. Sanjay Kumar T, Nageswari R, Somasundaram S, Anantharaju P, Baskar M, Ramesh T, et al. Significance of millets for food and nutritional security—an overview. Discov Food. 2024;4(1).
- 9. Chaturvedi P, Govindaraj M, Govindan V, Weckwerth W. Editorial: Sorghum and Pearl Millet as Climate Resilient Crops for Food and Nutrition Security. Front Plant Sci. 2022;13:851970. pmid:35360320
- 10. Satyavathi CT, Ambawat S, Khandelwal V, Srivastava RK. Pearl Millet: A Climate-Resilient Nutricereal for Mitigating Hidden Hunger and Provide Nutritional Security. Front Plant Sci. 2021;12:659938. pmid:34589092
- 11. Dunjana N, Dube E, Chauke P, Motsepe M, Madikiza S, Kgakatsi I. Sorghum as a household food and livelihood security crop under climate change in South Africa: A review. S Afr J Sci. 2022;118(9–10):35–40.
- 12. Rai KN, Murty DS, Andrews DJ, Bramel-Cox PJ. Genetic enhancement of pearl millet and sorghum for the semi-arid tropics of Asia and Africa. Genome. 1999;42(4):617–28.
- 13. Ndour PMS, Barry CM, Tine D, De la Fuente Cantó C, Gueye M, Barakat M, et al. Pearl millet genotype impacts microbial diversity and enzymatic activities in relation to root-adhering soil aggregation. Plant Soil. 2021;464(1–2):109–29.
- 14. Makuchete L, Hove A, Nezomba H, Rurinda J, Mbanyele V, Mlambo S, et al. Productivity of sorghum and millets under different in-field rainwater management options on soils of varying fertility status in Zimbabwe. Front Agron. 2024;6:1378339.
- 15. Hartmann M, Six J. Soil structure and microbiome functions in agroecosystems. Nat Rev Earth Environ. 2022;4(1):4–18.
- 16. Chen Y, Du J, Li Y, Tang H, Yin Z, Yang L, et al. Evolutions and Managements of Soil Microbial Community Structure Drove by Continuous Cropping. Front Microbiol. 2022;13:839494. pmid:35295291
- 17. Wang Y, Yu P, Liu Y-X. Metagenomic Analysis for Unveiling Agricultural Microbiome. Agronomy. 2024;14(5):981.
- 18. Nwachukwu BC, Babalola OO. Metagenomics: A Tool for Exploring Key Microbiome With the Potentials for Improving Sustainable Agriculture. Front Sustain Food Syst. 2022;6.
- 19.
South African Weather Service. Annual State of the Climate of South Africa 2023. Pretoria, South Africa: South African Weather Service; 2024.
- 20.
Non-Affiliated Soil Analysis Work Committee. Handbook of standard soil testing methods for advisory purposes. Soil Science Society of South Africa; 1990.
- 21. Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res. 2013;41(1):e1. pmid:22933715
- 22.
Simon Andrews FK, Segonds-Pichon A, Biggins L, Krueger C, Wingett S. FastQC: a quality control tool for high throughput sequence data [Internet]. 2010
- 23. Ewels P, Magnusson M, Lundin S, Käller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016;32(19):3047–8. pmid:27312411
- 24. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30(15):2114–20. pmid:24695404
- 25. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37(8):852–7. pmid:31341288
- 26. Callahan BJ, Wong J, Heiner C, Oh S, Theriot CM, Gulati AS, et al. High-throughput amplicon sequencing of the full-length 16S rRNA gene with single-nucleotide resolution. Nucleic Acids Res. 2019;47(18):e103. pmid:31269198
- 27. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41(Database issue):D590–6. pmid:23193283
- 28. McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013;8(4):e61217. pmid:23630581
- 29. Mantel N. The detection of disease clustering and a generalized regression approach. Cancer Res. 1967;27(2):209–20. pmid:6018555
- 30. Clarke KR, Ainsworth M. A method of linking multivariate community structure to environmental variables. Mar Ecol Prog Ser. 1993;92:205–19.
- 31. Legendre P, Anderson MJ. Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol Monogr. 1999;69(1):1–24.
- 32.
Oksanen J, Simpson GL, Blanchet FG, Kindt R, Legendre P, Minchin PR, et al. vegan: Community Ecology Package. R package version 2.6-4. 2022.
- 33.
SAS Institute Inc. Statistical analysis software.
- 34.
Addinsoft A. XLSTAT statistical and data analysis solution. Long Island (NY): Addinsoft; 2019.
- 35. Mpai T, Diale MO, Shargie N, Gerrano AS, Mtsweni PN, Bopape FL, et al. Functional and taxonomic profiles of soil microbial communities of tropical legume soils from smallholder farmers’ fields in Tzaneen, Limpopo province, South Africa. BMC Microbiol. 2025;25(1):601. pmid:41039197
- 36. Marzouk SH, Kwaslema DR, Omar MM, Mohamed SH. “Harnessing the power of soil microbes: Their dual impact in integrated nutrient management and mediating climate stress for sustainable rice crop production” A systematic review. Heliyon. 2024;11(1):e41158. pmid:39758363
- 37. Fierer N, Jackson RB. The diversity and biogeography of soil bacterial communities. Proc Natl Acad Sci U S A. 2006;103(3):626–31. pmid:16407148
- 38. Zhou X, Tahvanainen T, Malard L, Chen L, Pérez-Pérez J, Berninger F. Global analysis of soil bacterial genera and diversity in response to pH. Soil Biol Biochem. 2024;198:109552.
- 39. Sadeghi S, Petermann BJ, Steffan JJ, Brevik EC, Gedeon C. Predicting microbial responses to changes in soil physical and chemical properties under different land management. Appl Soil Ecol. 2023;188:104878.
- 40. Furtak K, Gawryjołek K, Gajda AM, Gałązka A. Effects of maize and winter wheat grown under different cultivation techniques on biological activity of soil. Plant Soil Environ. 2017;63(10):449–54.
- 41. Jezierska-Tys S, Joniec J, Mocek-Płóciniak A, Gałązka A, Bednarz J, Furtak K. Microbial activity and community level physiological profiles (CLPP) of soil under the cultivation of spring rape with the Roundup 360 SL herbicide. J Environ Health Sci Eng. 2021;19(2):2013–26. pmid:34917389
- 42. Chinthalapudi DPM, Pokhrel S, Kingery WL, Shankle MW, Ganapathi Shanmugam S. Exploring the synergistic impacts of cover crops and fertilization on soil microbial metabolic diversity in dryland soybean production systems using Biolog EcoPlates. Appl Biosci. 2023;2.
- 43. Burns RG, DeForest JL, Marxsen J, Sinsabaugh RL, Stromberger ME, Wallenstein MD, et al. Soil enzymes in a changing environment: Current knowledge and future directions. Soil Biol Biochem. 2013;58:216–34.
- 44. Geisseler D, Horwath WR. Investigating amino acid utilization by soil microorganisms using compound specific stable isotope analysis. Soil Biol Biochem. 2014;74:100–5.
- 45. Chen Q, Song Y, An Y, Lu Y, Zhong G. Soil Microorganisms: Their Role in Enhancing Crop Nutrition and Health. Diversity. 2024;16(12):734.
- 46. Lalid Kumar SP, Latha MR, Janaki P, Parameswari E, Kalaiselvi T, Senthamilselvi D, et al. Sustainable farming practices enhance bacterial diversity and nutrient levels in sorghum rhizosphere soil. Rhizosphere. 2024;32:100967.
- 47. Omotayo O, Igiehon O, Babalola O. Metagenomic Study of the Community Structure and Functional Potentials in Maize Rhizosphere Microbiome: Elucidation of Mechanisms behind the Improvement in Plants under Normal and Stress Conditions. Sustainability. 2021;13(14):8079.
- 48. Zuo Y-W, Zhang J-H, Ning D-H, Zeng Y-L, Li W-Q, Xia C-Y, et al. Comparative Analyses of Rhizosphere Bacteria Along an Elevational Gradient of Thuja sutchuenensis. Front Microbiol. 2022;13:881921. pmid:35591985
- 49. Duan R, Lin Y, Zhang J, Huang M, Du Y, Yang L, et al. Changes in diversity and composition of rhizosphere bacterial community during natural restoration stages in antimony mine. PeerJ. 2021;9:e12302. pmid:34721985
- 50. Ma Y, Liu M, Hong Y, Wang Y, Chang X, Shi G, et al. Influence of Soil Physicochemical Properties and Inter-Root Microbial Communities on the Inhibition of Anthracnose in Peppers. Microorganisms. 2025;13(3):661. pmid:40142554
- 51. Yang L, Han D, Jin D, Zhang J, Shan Y, Wan M, et al. Soil physiochemical properties and bacterial community changes under long-term polycyclic aromatic hydrocarbon stress insitu steel plant soils. Chemosphere. 2023;334:138926. pmid:37182712
- 52. Estrada R, Porras T, Romero Y, Pérez WE, Vilcara EA, Cruz J, et al. Soil depth and physicochemical properties influence microbial dynamics in the rhizosphere of two Peruvian superfood trees, cherimoya and lucuma, as shown by PacBio-HiFi sequencing. Sci Rep. 2024;14(1):19508. pmid:39174594
- 53. Plassart P, Prévost-Bouré NC, Uroz S, Dequiedt S, Stone D, Creamer R, et al. Soil parameters, land use, and geographical distance drive soil bacterial communities along a European transect. Sci Rep. 2019;9(1):605. pmid:30679566
- 54. Yang T, Lupwayi N, Marc S-A, Siddique KHM, Bainard LD. Anthropogenic drivers of soil microbial communities and impacts on soil biological functions in agroecosystems. Glob Ecol Conserv. 2021;27:e01521.
- 55. Bana RS, Grover M, Singh D, Bamboriya SD, Godara S, Kumar M, et al. Enhanced pearl millet yield stability, water use efficiency and soil microbial activity using superabsorbent polymers and crop residue recycling across diverse ecologies. Eur J Agron. 2023;148:126876.
- 56. Chaturvedi P, Govindaraj M, Sehgal D, Weckwerth W. Editorial: Sorghum and pearl millet as climate resilient crops for food and nutrition security, volume II. Front Plant Sci. 2023;14:1170103.
- 57. Afzal MR, Naz M, Ullah R, Du D. Persistence of Root Exudates of Sorghum bicolor and Solidago canadensis: Impacts on Invasive and Native Species. Plants (Basel). 2023;13(1):58. pmid:38202366
- 58. Al-Khaleel RI, Chellapilla T, Siddaiah CN. Pearl Millet: A Sustainable Source of Food and Nutrition in the Age of Climate Change. Anthr Sci. 2025;4(1–2):1–13.
- 59. Hemelda NM, Noutoshi Y. Root-exuded sugars as drivers of rhizosphere microbiome assembly. Plant Biotechnol (Tokyo). 2025;42(3):215–27. pmid:41181071
- 60. Dai Z, Guo X, Lin J, Wang X, He D, Zeng R, et al. Metallic micronutrients are associated with the structure and function of the soil microbiome. Nat Commun. 2023;14(1):8456. pmid:38114499
- 61. Wang L, Tang M, Gong J, Malik K, Liu J, Kong X, et al. Variations of soil metal content, soil enzyme activity and soil bacterial community in Rhododendron delavayi natural shrub forest at different elevations. BMC Microbiol. 2024;24(1):300. pmid:39135165
- 62. B K B, Dahal S, Koirala M, Poudel R, Kandel BP. Role of Millets for Food Security Under Climate Change. Plant Environ Interact. 2026;7(1):e70128. pmid:41737834
- 63. Kebede E. Contribution, utilization, and improvement of legumes-driven biological nitrogen fixation in agricultural systems. Front Sustain Food Syst. 2021;5:767998.
- 64. Lengwati DM, Mathews C, Dakora FD. Rotation Benefits From N2-Fixing Grain Legumes to Cereals: From Increases in Seed Yield and Quality to Greater Household Cash-Income by a Following Maize Crop. Front Sustain Food Syst. 2020;4.
- 65. Preeti, Sheoran S, Prakash D, Ankit, Todarmal, Rani S, et al. Soil carbon, micronutrients and microbiological dynamics under cash crop-based cropping systems in semi-arid National Capital Region of India. Sci Rep. 2026;16(1):4855. pmid:41507271
- 66. Liang B, Ma C, Fan L, Wang Y, Yuan Y. Soil amendment alters soil physicochemical properties and bacterial community structure of a replanted apple orchard. Microbiol Res. 2018;216:1–11. pmid:30269849
- 67. Zhang H, Jiang N, Zhang S, Zhu X, Wang H, Xiu W, et al. Soil bacterial community composition is altered more by soil nutrient availability than pH following long-term nutrient addition in a temperate steppe. Front Microbiol. 2024;15:1455891. pmid:39345260
- 68. Zheng B, Xiao Z, Liu J, Zhu Y, Shuai K, Chen X, et al. Vertical differences in carbon metabolic diversity and dominant flora of soil bacterial communities in farmlands. Sci Rep. 2024;14(1):9445. pmid:38658691
- 69. Ejaz MR, Badr K, Shabani F, Ul Hassan Z, Zouari N, Al-Thani R, et al. Comprehensive Investigation of Qatar Soil Bacterial Diversity and Its Correlation with Soil Nutrients. Microbiol Res. 2025;16(9):196.
- 70. Yang C, Li K, Lv D, Jiang S, Sun J, Lin H, et al. Inconsistent response of bacterial phyla diversity and abundance to soil salinity in a Chinese delta. Sci Rep. 2021;11(1):12870. pmid:34145370
- 71. Mucsi M, Borsodi AK, Megyes M, Szili-Kovács T. Response of the metabolic activity and taxonomic composition of bacterial communities to mosaically varying soil salinity and alkalinity. Sci Rep. 2024;14(1):7460. pmid:38553497
- 72. Hou Z, Chen W, Zhang X, Zhang D, Xing J, Ba Y, et al. Differentiated response mechanisms of soil microbial communities to nitrogen deposition driven by tree species variations in subtropical planted forests. Front Microbiol. 2025;16:1534028. pmid:40143871
- 73. Xia Q, Rufty T, Shi W. Soil microbial diversity and composition: Links to soil texture and associated properties. Soil Biol Biochem. 2020;149:107953.
- 74. Huang Z, Qin Y, He X, Zhang M, Ren X, Yu W, et al. Analysis on metabolic functions of rhizosphere microbial communities of Pinus massoniana provenances with different carbon storage by Biolog Eco microplates. Front Microbiol. 2024;15:1365111. pmid:38511000