Different Soil Salinity Imparts Clear Alteration In Rhizospheric Bacterial Community Dynamics In Rice And Peanut.

The rhizospheric microbiome is capable of changing the physio-chemical properties of its own micro-environment and found to be indispensable in overall health of the host-plant. The interplay between the rhizospheric environment and the microbiota residing therein tune the physiology of the associated plant. In this study, we have determined how the soil properties and the host-plant remains as an important parameter for microbial community-dynamics in the rhizosphere of rice and peanut. In addition to check the physio-chemical parameters of the rhizospheric soil, we have also prepared the metagenomic DNA from each rhizospheric soil followed by high-throughput sequencing and sequence-analysis to predict the OTUs that represents the community structure. The alpha-diversity of the bacterial community in the RRN sample was highest, while the lowest was in PRS sample. Actinobacteria is the most predominant phylum in PRN, PRS and RRN whereas Acidobacteria in RRS. We found a clear shift in bacterial community over the rice and peanut rhizosphere and also over these host-rhizospheres from normal and high saline region. The rhizospheric bacterial community composition found to be affected by the close-by environmental factors. Thus, the rhizospheric bacterial community-structure is related to both the adjoining soil characters and the type of the hosts.


Introduction
The root-adjoining soil where the plant has much higher in uence through secretion of their different rootexudates, mucilage and sloughed-cells is the rhizospheric region. The phytochemicals secreted in the rhizosphere are important for the establishment of the microbial community around it 1 . Rhizospheric region of the plant is a complicated and heterogeneous environment which dictates the type of allowed microbes therein. Molecular studies particularly the metagenomics suggest that the soil is the habitat of highest biodiversity which approximated as 1,000 Gbp of microbial genome sequence in each gram of soil 2 . The advantageous role of these rhizospheric microbes can be demonstrated through their direct role in plant-growth promotion, plant protection from phytopathogen and stress management of the plants 3 .
By virtue of having various microbes, the rhizosphere remains as the chemical factory where a complex physical, chemical and biochemical interaction are operating which in turn in uence the nutrient-cycling, nitrogen-xation, phosphate mobilization and solubilization, nutrient uptake, water uptake, production of plant growth regulators, seed germination and early plant growth promotion, development in soil structure and competing with plant pathogens 4 . Thus, the diversity of microorganism plays fundamental role in maintenance of soil-fertility and nutrient-cycling. Mostly such rhizospheric microbiome is having bene cial effect on plant, however in few cases the rhizospheric community might have representatives which are either neutral or detrimental for the host plant.
The plant root exudates chose the type of microbes it will attract or repel from the rhizospheric region, but this is not the only parameter. The physiochemical composition of the soil remains as another important criterion for building the rhizospheric community. Thus, it is in general the reciprocal action of the plant root and the soil which establishes a particular type of microbiota. The physiochemical characteristics of the soil are ever-changing mostly by various physical, chemical and biological means. The soil microbial community is in uenced by crop rotation, use of fertilizers and tillage which in turn change the soil physiochemical parameters 5 . The compounds present in root-exudates in uence the rhizospheric microbial community structure and is different at each developmental stage 6 . The microbial communityassisted nitrogen, phosphorous and other nutrients cycle are more rapid and dynamic in rhizospheric soil compared to the bulk soil 4 . Management of agriculture build up soil physicochemical properties which control microbial community composition and nutrient cycling. Furthermore, the microbial diversity and heterogeneity in bulk soil are increased by organic fertilizer 7 . Bacterial community compositions are different in organically-managed agricultural system than the conventional system 8 . Simultaneous analysis showed that the ecological interaction regulating structure, function and potential resilience of soil microbial communities 9,10 . Plant root systems are strong operator of microbial community congregation that build rhizosphere communities and are taxonomically and functionally different from bulk soil microbial community 11,12 . The rhizosphere effect and plant selection are evident in observation of microbiomes across different eld-environment 13 . Agricultural management and plant relation with microbial community structure establish taxonomy to network structure. Rhizosphere network often smaller, less densely connected and less complex than bulk soil microbial community 4 .
The abundance, diversity and composition of bacteria in the rice rhizosphere have been widely investigated. Bacterial population in rice rhizosphere were double to those of bulk soil 13,14 . The structures of bacterial communities in the rice rhizosphere are diverse and dynamic which are related to soil type, geographical location and rice genotype 15 . Metagenomic studies have indicated that the bacterial communities in rice rhizosphere are broadly inhabited by Proteobacteria, mainly Alpha-, Beta-and Deltaproteobacteria classes, Acidobacteria, Actinobacteria and Chloro exi phyla 16 . Alpha-, Beta-proteobacteria classes are abundant in rice rhizosphere and essential for ecosystem functioning 17 .
In this study, we have chosen the rhizospheric sample of peanut (dicotyledon) and rice (monocotyledon) from different geographical locations ( Fig. 1) in order to analyze the microbiota present therein. Here we have undertaken characterization of rhizospheric soil of rice and peanut, grown in different salinity condition, using targeted 16SrRNA genes through a metagenomic approach. We further analyzed the sequence of 16SrRNA variable region and highlighted the following prediction: the microbial community is notably different in rice rhizosphere grown in normal soil (RRN) than that of the saline soil growing rice (RRS) and rhizosphere of peanut grown in normal soil (PRN) over that of saline soil (PRS), and also from monocotyledons to dicotyledon grown in normal and saline soil growing.

Results
Physical and chemical characteristics of the rhizospheric soil The physical and chemical characteristics of all the rhizospheric soil were analysed in order to corelate with the rhizospheric microbiota (Table 1). pH analysis of the soil collected from the PRS and RRS Page 4/26 showed 7.91 ± 0.01 and 7.84 ± 0.02, respectively, whereas soil sample from PRN and RRN showed 6.48 ± 0.04 and 6.75 ± 0.03. Thus, the PRS and RRS soil sample has been found to be slight alkaline but near neutral in case of PRN and RRN. PRS and RRS showed signi cantly higher electrical conductivity (EC) than RRN and PRN. The organic carbon and organic matter present in these rhizospheric soils were found to be in a range between 0.64 to 2.79%. The salinity of PRS and RRS ranging between 1.5-1.9 mg/L and was found to be higher than PRN and RRN which were 0.2-0.3 mg/L. The rhizospheric soil samples showed higher silt and clay texture (34.19-38.25%) compared to the ne and coarse sand (5.83-15.84%). The metal analyses of the soil samples indicated prevalence of Fe and Al. The prevalence Cd in these sample was found to be the least. All the soil samples have shown presence of micro and macro nutrient (Cd, Co, Cr, Cu, Ni, Pb and Zn) which in general facilitated the growth of the microorganism.  Table S1). The number of bacteria appeared higher on LA than on ISP-2, as ISP-2 is generally supposed to be speci c for isolation of actinobacteria.
General characteristics of the amplicons and sequencing data Four amplicon samples from Illumina Miseq sequencing analysis of four different salinity regions were successfully sequenced (  Table 2). Beta-diversity indices show random distribution of bacterial population in these rhizospheres. Microbial taxonomic analysis at the phylum and class level Classi cation of the high-quality sequences also demonstrated differences in the bacterial communities among the different samples at the phylum, class, family and genus level (Fig. 2). A total of 29 phyla were identi ed in all samples out of which 7 phyla showed dominance in their relative abundance (> 1% of relative abundance in at least one sample) (p value < 0.05) (Fig. 2a). Actinobacteria was the most dominant phylum (> 20% relative abundance) across all samples, accounting for 20.4-59.23%. Acidobacteria, Chloro exi, Cyanobacteria, Fermicutes, Gemmatimonadetes and Proteobacteria were the subdominant phyla with > 1% relative abundance in at least one sample (Supplementary Table S2). The other 22 phyla had much lower abundances (less than 1% relative abundance in all samples).
A total of 85 bacterial classes were identi ed across all samples. There were twelve classes with a relative abundance of higher than 1% in at least one sample (p value < 0.05) (Fig. 2b). Among these twelve classes, the signi cantly dominant class (> 20% relative abundance) was Actinobacteria with 20.31% − 59.01% of the total high-quality sequences. The class of Acidobacteria was also dominant in RRS sample with 35.99% relative abundance. The subdominant classes were Thermoleophilia, Anaerolineae, Chloro exia, Bacilli, Clostridia, Gemmatimonadetes, Alpha-proteobacteria, Betaproteobacteria, Delta-proteobacteria, Gamma-proteobacteria with relative abundance of greater than 1% in at least one sample (Supplementary Table S3). In the sample of PRN, Alpha-proteobacteria, Bacilli, Clostridia, Gemmatimonadates, Beta-proteobacteria, Delta-proteobacteria, Gamma-proteobacteria, Acidobacteria were dominant with relative abundance ranging from 1-16%. The class Proteobacteria was co-dominantly inhabited with Actinobacteria in the sample of PRN with 16.26% relative abundance.
In the sample of PRS Acidobacteria, Anaerolineae, Clostridia, Alpha-proteobacteria were codominant with relative abundance of greater than 1%. Actinobacteria was also dominant in PRS sample like PRN with 48.64% relative abundance. Next to Actinobacteria, Anaerolineae was dominant in the sample of PRS with 19.96% relative abundance. Relative abundance of Alpha-proteobacteria was drastically reduced from 16-4% in PRS sample than the PRN. In the sample of RRN Alpha-proteiobacteria was also codominant with 8.54% relative abundance next to Actinobacteria. Only class Gemmatimonadetes was lesser than 1% relative abundance (0.06%) in RRN. In RRS, the relative abundance of Acidobacteria was dominant over others classes with 35.99% relative abundance. RRS was the only sample where Actinobacteria (20.31%) was not a dominant taxon. Class Anerolineae and Alpha-proteobacteria were codominantly present in RRS with relative abundance of 6.27% and 8.04%, respectively. Comparative relative abundance among four samples was also represented by heatmap analysis (p value < 0.05) (Fig.   3). Phylum Acidobacteria was dominated in sample RRS. Actinobacteria was dominated in the low salt rhizospheric region of samples in both PRN and RRN. Phylum Chloro exi was dominated in PRS ( Fig.  3a). Relative abundance of rhizospheric bacteria at the class level also re ected the same abundance like phylum. The class Acidobacteria was also dominant in RRS. The class actinobacteria were also showed the dominant relative abundance in both PRN and RRN. Alpha-proteobacteria, Beta-protobacteria, Deltaprotobacteria and Gamma-protobacteria were also dominant in normal salt soil of both PRN and RRN (Fig. 3b). This dynamic behaviour of bacteria, at phylum and class level in rhizospheric soils of monocot/dicot growing in different saline-soil, was very important to maintain the plant growth promotion.

Microbial taxonomic analysis at the family and genus level
The top 20 classi ed families with greater than 1% relative abundance atleast in one sample, was represented in a bar diagram (Fig. 2c was also represented by heatmap (p value < 0.05) (Fig. 3c, d). The family Acidobacteria was dominant in rice rhizosphere growing in saline soil. Actinomycetales was also dominant in PRN, RRN than the highsalt growing rice peanut rhizosphere PRS and RRS.
Comparative fold-shift of families and genera between low and high-salt rhizosphere We have tried to highlight the dominance of different bacterial group present in the rhizosphere of both the host growing in normal and saline soil between samples in phylum, class, family and genus-level ( Fig.  4, 5, Supplementary Fig. S1, S2). We have also compared the relative abundance of bacterial families in peanut and rice rhizosphere. It is striking to note that there are many taxa which has considerable increase or decrease in relative abundance over the variation in plant and the soil they grow. In order to highlight the shift in community dynamics in family-level, we have estimated the fold increase or decrease within eight major families associated with each of the sample. The family Xanthomonadaceae has 387.46-fold higher abundance in PRN than PRS, whereas Anaerolineaceae, Chloro exi, Nocardioidaceae are dominant in PRS than in PRN (Fig. 4a, b). Family Anaerolinaceae has 44.14-fold higher abundance in PRS over PRN (Supplementary Table S6). Likewise, the abundance of Chloro exi, Nocardioidaceae, Solirubrobacteraceae, Streptomycetaceae, Thermoleophilaceae, Thermomonosporaceae families in RRN are much higher than in RRS (Fig. 4c). Thermoleophilaceae has 18.62-fold higher abundance in RRN than in RRS. Acetobacter and Gematimonadaceae have the highest fold increase in their abundance in RRS over RRN (Fig. 4d, Supplementary Table S6).
Similarly, we found that the genus Actinocorallia, Hydrogenedens, Kaistia, Kibdelosporangium, Koribacter, Marmoricola, Nakamurella, Nitriliruptor, Plantactinospora, Prauserella, Pseudonocardiaceae, Solirubrobacterales, Sphingomonas, Sphingopyxis, Stella, Xanthomonadales are many folds higher abundant in PRN over PRS (Fig. 5a, Supplementary Table S7). Relative abundance of Xanthomonadales is higher in the PRN rhizospheric soil than in the PRS. Xanthomonadales found to have the highest fold increase (710.89). Bellilinea, Chloro exaceae, Longilinea, Nocardioides, Nocardiopsis, Pelolinea are shown to have a dominant fold increase in PRS than that of the PRN (Fig. 5b). Longilinea has the highest fold increase in abundance in PRS sample. Actinocorallia, Aminicenantes, Chloro exaceae, Marmoricola, Nocardioides, Nocardiopsis, Plantactinospora, Prauserella, Solirubrobacterales, Streptomycetaceae, Thermoleophilum, were found to have higher fold increase in RRN over RRS (Fig. 5c). Hydrogenedens, Koribacter, Paludibaculum, Telmatobacter are having a higher fold increase in RRS than the same of RRN (Fig. 5d, Supplementary Table S7). We have further con rmed the bacterial community dynamics among tested sample in phylum level through circus plot as well as venn diagram. Both these analyses not only con rm the above results but also clarify the shift in abundance of certain bacterial taxa. The circus plot represents the switching in number of OTUs in phylum-level (Fig. 6a) and family level ( Supplementary  Fig. 3) among low and high salt grown peanut and rice rhizosphere. The same pattern has also been found when a total of 61 abundant genera present in all samples were plotted in a venn diagram (Fig.  6b). There were 8.6% of the total genera present in all rhizosphere while 20%, 5.7%, 22.9% and 5.7% genera were found exclusively in PRN, PRS, RRN and RRS, respectively (Fig. 6b). The details distribution of major genera present in all kind of rhizosphere has been mentioned in Fig. 6b.

Analysis of the relationship between environmental factors and the rhizospheric microbial community
A beta-diversity analysis based on PCoA plot was performed to compare the bacterial compositions among the four different samples (p value < 0.05). PRN and RRN were clustered into same group. Although PRS and RRS were situated distantly (Fig. 7). Canonical correspondence analysis (CCA) was used to established the relationship between the environmental factors and the bacterial phylum (p value < 0.05) (Fig. 8). CCA plot was carried out using OUT S data together with environmental factors (pH, EC, salinity, coarse sand, Al, Cd, Cr, Fe, Ni. Pb and K), which might in uence the bacterial community structure, to highlight relative abundance of different phyla among these samples. According to Tukey's post-hoc test (999 permutations) the signi cant relationship between environmental variables and canonical axes were analyzed by using the paleontological statistics package version 3.01 (PAST Software 1). Based on Tukey's post-hoc test, the bacterial phylum was signi cantly linked (p ≤ 0.05) to the rhizosphere environmental factors (Fig. 8). Factors corresponding to the differences in the bacterial phylum were pH, EC, salinity of the sample, and presence of coarse sand, Al, Cd, Cr, Fe, Ni. Pb and K along with the rst axis CCA1 explaining 69.2% and the second axis CCA2 (28.57%) of the variation (Fig. 8). The plot obtained by CCA visualizing the different habitat preferences of the phyla. The pH, EC, salinity, coarse sand, Al, Cr, Fe, Ni. Pb were the most important environmental factors to in uence the rhizosphere bacterial phylum abundance, and were positively correlated with CCA1 axis (Table 3). CCA2 axis had a positive correlation with pH, EC, salinity, coarse sand, Al, Cr, Fe, Ni. Pb but was negatively correlated with Cd and K.  19,20 . The high conductivity, organic carbon content and fair distribution of the metals like Ca, Na, Mg and K for the entire soil sample gives us insight into the availability of the essential nutrient in the rhizosphere environment which helps in proliferation of microbial population in this area 19 . The salinity of the PRS and RRS was found to be higher than PRN and RRN, which indicates the presence of the soluble salt which essentially increases the conductivity of the soil. Clay content for all the soil was found to be higher with respect to the coarse sand, ne sand and silt that in turn might also provide better condition for the microbial growth in the rhizospheric region 21 . In our study, we have found total 29 phyla, 85 classes, 315 families and 639 genera from the tested samples (Supplementary Table S8-S11). The total OTUs of four different samples of PRN, PRS, RRN and RRS are 7353, 7172, 9038 and 6391, respectively. Rhizospheric bacterial alpha-diversity were higher in both monocot and dicot grown in normal soil than the saline soil. Beta-diversity shows random distribution of bacterial population in all the four samples. We have identi ed total 29 phyla in all four samples. Among them Actinobacteria, Acidobacteria, Chloro exi, Cyanobacteria, Firmicutes, Gemmatimonadetes, and Proteobacteria were higher in abundance than other 22 phylum. Previous research has shown that Proteobacteria, Acidobacteria, Actinobacteria, Bacteroidetes and Firmicutes were dominant phyla in the rhizosphere soil of Arabidopsis and cotton 27,28 . Total 85 bacterial classes were identi ed across four samples. Actinobacteria, Acidobacteria, Acidobacteria, Thermoleophilia, Anaerolineae, Chloro exia, Bacilli, Clostridia, Gemmatimonadetes, Alpha-proteobacteria, Betaproteobacteria, Deltaproteobacteria, Gamm-aproteobacteria were the higher relative abundance classes in four sampling sites. Different species of Actinobacteria were responsible for recycling of nutrient in a great extent of rhizosphere soil 29,17,30 and this Actinobacteria is the most abundant phylum in the rhizospheric soil PRN, PRS and RRN. However, the rhizospheric soil of RRS showed Acidobacteria as most abundant but they have quite high abundance of Actinobacteria as well. Previous study showed that Acidobacteria could grow in environment which is nutrient less and also showing higher abundance in poor soil 31 . Acidobacteria and Actinobacteria were found to be abundant in disease-suppressive soils and reported to be responsible to suppress disease-causing microbes and trigger enhancement of bene cial microbes that have potential to promote crop health 32  Thermoleophilaceae, Thermomonosporaceae were dominant fold increase values in normal rhizospheric soil than the saline condition, whereas Acetobacter and Gematimonadaceae were the highest fold increase in saline rhizosphere soil compare to normal soil. The genus Actinocorallia, Aminicenantes, Chloro exaceae, Marmoricola, Nocardioides, Nocardiopsis, Plantactinospora, Prauserella, Solirubrobacterales, Streptomycetaceae, Thermoleophilum, were the higher fold increase in normal rhizosphere than saline soil. Hydrogenedens, Koribacter, Paludibaculum, Telmatobacter were showed a higher fold increase in saline rhizosphere than the normal soil. Several studies reported Sphingomonadaceae, Chitinophagaceae, Nocardioidaceae, Solibacteraceae, Bacillaceace, Cytophagaceae and Methylobacteriaceae were predominant families (> 2% relative abundance) of T. aestivum L rhizosphere region and different bacterial genera like Sphingomonas, Microvirga, Bacillus, Nocardioides, Marmoricola, Bryobacter, Flavisolibacter were the dominant (> 1% relative abundance) 35,36 . Whereas Bacillus nealsonii, Rhodospirillales_ bacterium_WX36 and Bacillus niacini were prominent (> 0.5% relative abundance) bacterial species 36 . Sphingomonas, Kaistia, Xanthomonadales were unique genus belong to phylum Proteobacteria and Nakamurella, Plantactinospora, Thermomonospora, from Actinobacteria; genus Stella from Bacteroidetes present only in the rhizosphere of PRN, whereas genus Pelolinea, Longilinea belong to phylum Chloro exi were found only in PRS. It was interesting that phylum Chloro exi has been found only in the rhizosphere of PRS. Similarly, the genus Thermoleophilum, Prausere, Solirubrobacterales, Rhodococcus, Marmoricola from phylum Actinobacteria; Methyloceanibacter, Rhodospirillales from Proteobacteria and Aminicenantes has been found distinctively only in the rhizosphere RRN, whereas genus Paludibaculum, Koribacter from Acidobacteria were found only in the rhizosphere of RRS.
The in uence of different environmental factors describing the microbial community is like the nding of beta diversity analysis due to differences of rhizospheric community to the monocot and dicot types (Fig. 7). Factors corresponding to the differences in the bacterial communities were pH, EC, Salinity, Coarse Sand, Al, Cd, Cr, Fe, Ni. Pb and K along with the rst axis explaining 69.2% and the second axis (28.57%) of the variation (Fig. 8). CCA plot shows that pH, EC, Salinity, Coarse Sand, Al, Cr, Fe, Ni. Pb were the most important environmental factors to signi cantly in uence the rhizosphere bacterial phylum abundance (Fig. 8). Previous study has shown bacterial community structure were in uenced by pH, Cr, Sb, As, Zn and moisture content, but pH was the most dominant factor 37 .
The clustering of PCoA and CCA analysis signify PRN and RRN were clustered into the same group, while PRS and RRS were situated distantly. The CCA result in our study indicate pH, EC, Salinity, Coarse Sand, Al, Cr, Fe, Ni. Pb is positively correlated with rhizospheric community structure. Thus, pH, EC, Salinity, Coarse Sand, Al, Cr, Fe, Ni. Pb appeared to be an important factor that in uence microbial communities and dynamics.

Conclusion
In summary, this study illustrated the rhizospheric bacterial community of rice and peanut growing in normal soil and saline soil of West Bengal, India. Along with comment on the relative abundance of bacterial taxa present in different rhizospheric soil of peanut and rice, we also highlighted the shift in community dynamics in phylum, class, family and genus level. The normal rhizospheric soil in both monocot and dicot shows more diverse compare to saline rhizospheric soil in monocot and dicot. Alphadiversity like Shannon Index was higher in both monocot and dicot normal rhizospheric soil with high diversity in respect to saline rhizospheric soil. Once again it has been proved that the rice being a monocot and the peanut being a dicot has their own choice to contain preferable bacterial representative around their roots. The decreased abundance and variation of bacterial population in the rice and peanut rhizosphere of high saline zone might be one of the reasons behind the signi cant reduction in the vigor and yield of the crop cultivated in high salt containing region. Lastly the shift in community dynamics has been shown as a factor of soil parameters as well as the choice of the associated plants.

Site description
The study was performed with samples of rhizospheric region associated with peanut and rice plant grown in normal and high saline soil condition collected from places of West Bengal, India (Fig. 1).
Geographical location of the sample collection sites is Galsi (23°19'48.00"N, 87°42'.00"E) for rice rhzospheric soil with low salinity area (RRN). This is central plain areas of the district Burdwan of West Bengal, sometimes encounter heavy oods during rainy seasons. Next sample collection site is Egra (21°53'58.09"N, 87°32'16.58"E) for the peanut rhizospheric soil with low salinity area (PRN). Third and fourth sites are Gosaba (22°09'36.00"N, 88°47'60.00"E) both for rice and peanut rhizospheric soil with high salinity (RRS and PRS, respectively). This is one of the main deltaic islands in the Sundarban region.

Sampling
Four samples of the peanut and rice rhizosphere having different soil-salinity were collected in May 2017.
The owner of the all four sampling sites granted all necessary permits to access. On spot, surface soils around the plants were cleared of debris and unwanted pebbles and the soil sample were collected to an 8-10 cm depth from the surface on soil using auger with 10 cm diameter around the roots. Roots are gently shaken to remove the non-rhizospheric soil. The rhizospheric soil samples were collected using brush. Each of the soil sample that is free of debris kept into a labeled zipper storage bag in cool place for further analysis.
Analysis of the physio-chemical parameters of the sample grade water) were used to get the total bacterial cfu/gm (colony forming unit/microliter) soil sample. The media pH was maintained between 7-8. Cycloheximide (50 µg/ml) was added in the media used as the anti-fungal agent to avoid the undesired growth of the fungi. This soil-suspensions were then serially diluted and 100 µl from different selected dilutions from each sample were then used as inoculum and spread onto LA and ISP-2 agar plates. LA plates were used for estimating the cfu of all general types of bacteria; while, the ISP-2 media were used for the estimation of cfu of actinobacteria. The LA and ISP-2 agar plates were then incubated at 37ºC and 30ºC for 2 and 5 days, respectively. Colonies were then counted to determine the cfu/ml from each sample. were created for all taxonomic levels from phylum to genus, as well as for OTUs. QIIME was used to determine the alpha diversity and rarefaction curve that respectively represents community diversity (Inverse Simpson and Shannon) and species richness (observed OTUs and Chao1). Beta diversity was calculated by weighted UniFrac distance matrix 46 and was visualized using the Principal Coordinate Analysis (PCoA) plot. To assess the stability of the PCoA plot was performed on the OTU table. Based on the relative abundance of genera from each sample Venny 2.1 software was used to construct the venn diagram 48 whereas the Circos-0.67-7 software has been used to construct the Circos diagram 49 .
Associations among relative abundances of phylum in each sample and measured environmental factors (pH, EC, Salinity, Coarse Sand, Al, Cd, Cr, Fe, Ni. Pb and K) were determined using canonical correspondence analysis (CCA) 50 .

Statistical analysis
Statistical analyses were performed using the Paleontological Statistics package version 3.01 (PAST software1). All the data were subjected to one way ANOVA analysis followed by Tukey's post-hoc test.
Differences were considered signi cant if p ≤ 0.05 with multiple comparisons using 999 permutations.  SIIslametalSciRep.pdf