After the completion of the filtering, denoising, and removal of chimeric sequences, the final sequence counts exhibited a range from 33,374 (T3B) to 119,809 (T1A). The percentage of sequences that remained non-chimeric, relative to the original input, varied among the samples, spanning from 64.54% (T4A) to 80.57% (C2). In terms of non-chimeric sequence percentages, the control samples, specifically C1, C2, and C4, displayed values of 68.98%, 80.57%, and 78.3%, respectively. In contrast, the treatment groups exhibited the following percentages: 73.58% (T1A), 65.05% (T1B), 66.62% (T2A), 68.37% (T2B), 67.82% (T3A), 69.72% (T3B), 64.54% (T4A), and 66.44% (T4B). The alpha diversity of the samples, characterized by multiple indices, displayed significant changes across different soil depths and between the control and treatment groups. At a depth of 10 cm, Control 1 (C1) demonstrated the highest diversity with 1264 operational taxonomic units (OTUs), a Chao1 index of 1278, a Shannon index of 9.39, a Simpson index of 0.997, and an evenness index of 0.911. In contrast, Treatment 1 samples (T1A, T1B) exhibited lower diversities after the treatment process. At 48 cm, Control 2 (C2) displayed lower diversity indices compared to Treatment 2 samples (T2A, T2B), with T2B registering the highest values. For the samples collected at 65 cm, both T3A and T3B exhibited similar levels of diversity. Lastly, at a depth of 80 cm, Control 4 (C4) had the lowest diversity among all the samples, while Treatment 4 samples (T4A, T4B) demonstrated higher diversity levels in comparison to their respective control samples (Fig. 1).
The analysis of the bacterial composition in different samples revealed statistically significant differences among the treatments and controls (p < 0.05). Control 1 (C1) exhibited a significantly higher abundance of Actinobacteria compared to other samples, making up 54.3% of the total bacterial composition. Conversely, Control 2 (C2) and Control 4 (C4) were significantly different with a dominance of Proteobacteria, constituting 71.5% and 80.6% of their respective bacterial populations (p < 0.05). Treatment 1A (T1A) also showed a significant presence of Proteobacteria, though at a relatively lower prevalence of 44.6% (p < 0.05). Furthermore, the Firmicutes phylum demonstrated the predominant bacterial group in Treatment samples, including 1B (T1B), 2A (T2A), 2B (T2B), 3A (T3A), 3B (T3B), 4A (T4A), and 4B (T4B), representing 70.8–78.7% of the bacterial communities under the treatment conditions (p < 0.05). In terms of other phyla, significant variations were observed, with Chloroflexi ranging from 0.07–12.11%, Planctomycetota from 0–11.33%, Acidobacteriota from 0.29–5.10%, Gemmatimonadota from 0.02–9.57%, Verrucomicrobiota from 0–1.31%, and Myxococcota from 0–0.65% (p < 0.05). The variations in Firmicutes from 13.34–76.97% were also statistically significant, as were the fluctuations in Bacteroidota from 0.27–8.75% (p < 0.05). Additionally, the collective group of Minor Phyla represented significant proportions ranging from 0.74–2.17% of the microbial populations across the samples (Fig. 2).
The bacterial class distribution also demonstrated significant differences across the control and treatment samples (Fig. 3). For instance, C1 showed a significantly higher proportion of Acidimicrobiia (7.18%), Actinobacteria (12.85%), and Actinobacteriota (1.55%) (p < 0.05). A shift in dominance was observed in C2, with Alphaproteobacteria taking the lead (40.20%), followed by Bacilli (14.72%), and a reduced presence of Actinobacteria and Acidimicrobiia at 3.44% and 0.53% respectively. Furthermore, there is clear dominance of Alphaproteobacteria (44.99%) and Gammaproteobacteria (35.60%) in the C3 sample (p < 0.05). Sample T1A exhibited a high prevalence of Alphaproteobacteria (25.19%), accompanied by a notable presence of Bacilli (7.73%) and Gammaproteobacteria (19.37%). Similarly, the prevalence of Clostridia (51.59%) and Bacilli (24.40%) in T1B, and the dominance of Clostridia (51.87% and 47.41%) and Bacilli (24.69% and 23.38%) in T2A and T2B, respectively, were all significant (p < 0.05). The samples T3A and T3B maintained the trend of Clostridia and Bacilli dominance, constituting 49.80% and 51.87%, as well as 24.77% and 25.21% of the respective communities. Finally, T4A and T4B continued with this pattern, harbouring a majority of Clostridia (52.07% and 52.84%) and Bacilli (24.84% and 25.80%).
Detailed examination of the bacterial genus distribution (Fig. 4) across different sample groups revealed the Methylorubrum genus to be ubiquitous, with substantial variations in its relative abundance. While it achieved peak prevalence in the Treatment 2A group at 24.47%, it displayed relatively lower concentrations in the other groups, varying from 2.15–2.44%. This noticeable variation suggests the potential selectivity of Treatment 2A in promoting the proliferation of Methylorubrum, a trait absent in the other treatments and control groups. "Escherichia-Shigella" and Bradyrhizobium genera were also present in significant proportions across the different treatments, again with evident variation. The Comamonadaceae genus showed an interesting pattern, with a significant rise in its relative abundance across all treatment groups, surpassing 2%. Its maximum abundance, however, was found in the Control 4 group at 7.91%. This might imply that the conditions in the Control 4 group are especially favourable to the proliferation of this bacterial genus. Meanwhile, Bacillus and Paenibacillus genera showed relatively limited occurrence, suggesting that these genera were less influenced by the applied treatments. In Treatment 3, notable abundance was observed for the Lactobacillus and Clostridium_sensu_stricto_1 genus. The Lactobacillus genus exhibited an evident increase in Treatment 3A and 3B, peaking at 3.91% and 3.42% respectively. On the other hand, Clostridium_sensu_stricto_1 displayed a significantly improved occurrence, achieving 28.58% in Treatment 3A and 25.18% in Treatment 3B. This strongly high representation in these specific treatment groups hints towards a selective stimulatory effect of Treatment 3A and 3B on Clostridium_sensu_stricto_1. Romboutsia also showed similar shifts, with high prevalence observed solely in Treatment 3A and 3B groups.
The Mann-Whitney U test was used to identify significant differences in the bacterial community between the control and treatment groups. Notably, a significant change was observed between Control 1 and Treatment 1, while other control-treatment pairs did not show any significant differences. In terms of bacterial composition, Treatment 3 exhibited differentiation from Treatments 1 and 2 but not from Treatment 4 (Fig. 5). The most significant changes in bacterial populations from Control 1 to Treatment 1 were observed in several genera. The genus 67 − 14 (Thermoleophilia class) and an uncultured genus (Acidimicrobiia class) exhibited a significant reduction of approximately 84.09% and 72.86%, respectively. In contrast, Methylobacterium-Methylorubrum exhibited a significant increase of about 1987.62%, indicating a highly favourable response to the treatment conditions. The genus Escherichia-Shigella displayed a notable increase of approximately 1244.86%. In Treatment 2, Methylobacterium-Methylorubrum exhibited a marked decrease of 864.58%, and Escherichia-Shigella demonstrated a significant decline of 680.91%. Bradyrhizobium and Ralstonia encountered considerable reductions of 909.66% and 1081.32%, respectively, while Bacillus showed a significant reduction of 3590.48%. In contrast, Rubrobacter displayed an increase of 25.51%. Comparison of Treatment 4 with Control 4 revealed considerable variations, with Methylobacterium-Methylorubrum decreasing by approximately 642.85%, and Escherichia-Shigella experiencing a significant drop of 591.46%. Bradyrhizobium exhibited a loss of nearly 694.04%, and Ralstonia suffered a remarkable drop of 890.70%. Conversely, Lactobacillus demonstrated a notable increase of 58.11%, and Bacillus displayed a significant rise of 692.61%. Clostridium_sensu_stricto_1 (Clostridia class) displayed the most significant increase, with a shift of approximately 27.07%, followed by Romboutsia (Clostridia class) with an increase of about 15.67%. Turicibacter (Bacilli class) exhibited an increase of 10.69%, while the genus Muribaculaceae (Bacteroidota phylum) showed a smaller but notable increase of 4.74%. Additionally, the well-known genus Lactobacillus from the Bacilli class showed an increase of 3.75%. Interestingly, Methylobacterium-Methylorubrum, despite decreasing in other treatments, exhibited a modest increase of 2.25% in Treatment 3. These findings highlight the diverse microbial responses and shifts to different treatments, potentially driven by selective pressures in response to oil exposure.
The co-network analysis of the top 30 bacterial genera across various control and experimental conditions provides some valuable insight into the complex interactions within the microbial community (Fig. 6). For instance, the genus Vicinamibacteraceae demonstrates a notable ability to maintain positive relationships across multiple conditions, with significant positive correlation indicating strong associations with other genera. While g_67 presents a contrasting dynamic, with a prevalence of negative correlation, especially notable in experimental conditions. Genera like Methylobacterium and Escherichia show a network of positive correlations, that are particularly strong in experimental conditions implying that these genera may thrive or respond similarly under the conditions tested in the experiments. Romboutsia and Turicibacter have a combination of positive and negative correlations across both control and experimental conditions. Notably, Uncultured, Burkholderia, and Providencia are characterized by several significant positive correlations that are higher in experimental conditions compared to control, suggesting that the experimental manipulations may favour their association with other genera. The absence of significant negative correlations for genera such as Rubrobacter and Alcaligenes, and for several uncultured genera, across both control and experimental conditions, is indicative of either a non-competitive stance or a broad ecological niche that allows for coexistence without direct antagonism. while others exhibit a balance of positive and negative correlations across control and experimental conditions. This pattern suggests that experimental manipulations can either promote cooperation or competition among genera, highlighting the dynamic nature of microbial interactions in response to environmental changes.