Removal of H2S by vermicompost biofilter and analysis on bacterial community

Abstract The vermicompost collected from dewatered domestic sludge as packing material in biofilter was investigated for hydrogen sulfide (H2S) removal. No nutrients or microbial inoculation was added throughout the experiment. The corresponding bacterial community characteristics in the vermicompost biofilter of different spatial levels were evaluated by Miseq high-throughput sequencing technique. The results showed that the vermicompost biofilter performed well during operation. The H2S removal efficiency reached nearly 100% under condition of the inlet concentration <350 mg m−3 and 0.25−0.35 m3 h−1 gas flow rate. The maximum elimination capacity of 20.2 g m−3 h−1 was observed at a flow rate of 0.35 m3 h−1. Furthermore, the amounts of biodegraded products and pH varied accordingly. In addition, the results from high-throughput sequencing revealed pronouncedly spatial variation of the vermicompost, and the Rhodanobacter, Halothiobacillus, Mizugakiibacter as well as Thiobacillus, which can play an important role in removing H2S, were predominant in the final vermicompost. These results imply that the vermicompost with diverse microbial communities has a good potential for eliminating H2S.


Introduction
The odor is one of the world's seven typical public pollutions. It is harmful to people's living environment and health, and it can affect the growth of plants and animals. Furthermore, it is a potential safety concern. The hydrogen sulfide (H 2 S) which smells like rotten eggs is the most common, highly toxic, and malodorous gas. It is a colorless, corrosive, and flammable gas that is an environmental pollutant [1]. The combustion of H 2 S leads to a formation of SO x , such as SO 2 , which is an acid rain precursor and causes serious damages to ecological systems and human health [2]. H 2 S is emitted from natural and anthropogenic sources. For instance, the natural emission sources are volcanoes, sulfur springs, and undersea vents. The major anthropogenic H 2 S emission sources are petrochemical processes, paper manufacturing, food waste treatment, biogas production, composting, wastewater treatment, landfills, and livestock manure [1,3]. The olfactory threshold of H 2 S is approximately 0.00047 ppm. The American Conference of Governmental Industrial Hygienists (ACGIH) points out that people's olfactory neurons will become fatigued when exposed to H 2 S at 150−200 ppm. At approximately 1,000 ppm, people will lose consciousness and could even die in seconds owing to respiratory paralysis and cellular anoxia [4]. Therefore, H 2 S must be purified from waste gas for the sake of public health and safety as well as environmental protection.
The common methods for H 2 S removal are physical, chemical, and biological treatments. The first two methods are based on adsorption, chemical oxidation, and precipitation adsorption technologies [5,6]. However, these processes have high operating and processing costs related to energy and chemicals, and also produce hazardous byproducts that must be treated before discharge [7]. In contrast, the biological treatments (biofilters, biotrickling filters, and bioscrubbers) have advantages of low operating cost and energy consumption; high removal efficiency; less secondary pollution; and public acceptance as an environmentally friendly technology [8,9]. A biofilter has been proposed as a feasible and effective option for abatement of H 2 S in recent years [10]. The removal capability and steady operation of the biofilter depend on the synergistic effects of the microbial activity, diversity, and benign environment. The packing material acts as the carrier for the microorganisms. Its performance directly influences the environment for microbial growth and activities, mass transfer, absorption, and adsorption [11,12] and it is considered as the backbone of the biofiltration. Therefore, selection of an appropriate filter media helps to improve the efficiency and capacity of the biofilter. Many studies have shown that the packing materials, such as soil, compost, peat, volcanic rock, polyethylene pellet, compost and biochar, expanded schist polyurethane foam and activated carbon are commonly be used for H 2 S treatment [7,[13][14][15]. Morgan-Sagastume and Noyola [16] showed that mixing filter media of a compost biofilter affected the H 2 S removal efficiency. Moreover, the removal capacity has remained constant and close to 100% under the mixed compost conditions. As the treatment progressed, internal changes within the biofilter were correlated with its moisture content. However, the investment and operational costs must be considered before implementing this technology. As mentioned by Kante et al. [17] and Nor et al. [18], activated carbon had a good H 2 S removal performance; however, it alone does not provide the characteristics for microorganism survival, thus it is mostly used in combination with other packing materials that have beneficial properties.
Vermicomposting involves interaction between microorganisms and earthworms. Biological degradation and stabilization of organic solid wastes occur via digestion by earthworms that transform the waste into useful vermicompost. Vermicompost has high porosity and a large specific surface area; it is a homogenized and humic substances; it has an abundant organic carbon and bacterial community [19,20]. Just like the compost with substantial fertilizer value, the vermicomposting product is an efficient growth promoter for plants as it contains higher available nutrients and much more diverse agricultural probiotics [21,22]. Several investigations have been conducted about the use of vermicompost in abatement of waste gases such as ethylene, methane, formaldehyde, ammonia, and hydrogen, as well as odor from human feces and livestock farming [23][24][25][26]. Up to now, there is little research available on vermicompost as packing material to remove H 2 S and the microbial analysis in vermicompost.
The main objectives of this study were to assess the feasibility of removing H 2 S using vermicompost and to investigate the associated microbial community. The vermicompost obtained from dewatered sludge was chosen as a media of bioreactor, and the changes of microbial community were analyzed by Miseq highthroughput sequencing technique in this study. The results of this research may provide a basis for vermicompost as a biomedia to remove hydrogen sulfide and could be used as a guide for further design and operation of industrial-scale systems.

Vermicompost and hydrogen sulfide
The initial physicochemical characteristics of vermicompost are given in Table 1 [27]. The vermicompost was collected from a vermicomposting reactor of dewatered domestic sewage sludge. In detail, the vermicomposting experiment was performed at a 1/5 weight ratio of earthworms to sludge in the laboratory at 25 ± 1°C for 60 days based on the method of Fu et al. [28]. After vermicomposting, all earthworms were picked out and the obtained vermicompost was sieved using a standard sieve with a mesh range of 2.36-4.75 mm. The vermicompost was loaded in the bioreactor by hand.
Hydrogen sulfide was supplied from a concentrated standard gas cylinder (99.99% purity; 8,000 mg m −3 concentration; balanced gas N 2 ; Chengdu Keyuan Gas Co., Ltd, China) and H 2 S was diluted to different concentrations using air from a compressor.

Bioreactor set-up and operation
The bench-scale horizontal biofilter system was composed of a prehumidification system, a bioreactor, and a 71.66 ± 0.10 tail gas absorber (Figure 1b). The biofilter was a cuboid of Perspex packed with vermicompost. Its length, width, and height were 400 mm, 130 mm, and 210 mm, respectively. The vermicompost was placed in four Teflon mesh segments (80 × 12 × 200 mm in height, each), which provided an overall thickness of 320 mm and an effective packing volume of 7.37 L. Five H 2 S sampling ports were distributed within bioreactor: one each at the inlet and outlet points and three along the length of the bioreactor. Compressed air was introduced into the bioreactor system after passing through an activated carbon filter to remove impurities. The airstream was prehumidified in a distilled water container before mixing the streams. Hydrogen sulfide was safely discharged from the cylinder using a pressure-reducing regulator and a flow meter. It was diluted with prehumidified air in a mixing chamber and then entered the system from the left of the biofiltration. The H 2 S stream, after traveling throughout the bed, exited the system at the right, passed through the three-way valve, then entered into a bottle containing an absorbing solution of NaOH (10%) and, finally, into the exhaust hood. H 2 S passed through the bioreactor for removal by the microbial community in the vermicompost. The flow rates of the streams were controlled with a flow meter. The overhead view of the biofiltration bed packed with vermicompost is shown in Figure 1a.
The system was operated at room temperature (20−28°C) and no nutrients or microbial inoculation was added during the whole experimental process. The experimental conditions were as follows: gas flow rate was adjusted to 0.25, 0.35, 0.45, and 0.55 m 3 h −3 , corresponding to an empty bed gas retention time of 106, 76, 59, and 48 s, respectively. The concentrations of H 2 S varied between the range of 100−500 mg m −3 for each stage.

Sampling
H 2 S samples were periodically collected from the bioreactor using 0.5 L Tedlar bags (Cole-Parmer, USA). One fresh vermicompost sample was collected at the start of the experiment and is referred as a CK. Four samples were collected from first to fourth segments of bioreactor at the end of the experiment and are referred as T1−T4. Each sample was collected in duplicate. The collected samples were divided into two parts for different analyses. One part was immediately stored at −20°C to analyze the bacterial populations in the biofilter, while the other was air-dried, ground, and sieved using an 80 mesh to determine the physicochemical properties.

Analysis of physicochemical properties
All measurements were determined in triplicate. The amount of H 2 S concentration was periodically determined by gas chromatography (GC-2014, Shimadzu, Japan) with a flame photometric detector (FPD) [29] and DB-5 capillary column (15 m × ∅ 0.53 mm × 1.0 µm; Hewlett Packard, USA). The temperatures of the oven, injector, and detector were fixed at 100, 50, 200°C, respectively. The flow rate was 50 mL min −1 . A six-port sampling needle valve automatically regulated the injection of product gases into the GC. Nitrogen was used as the carrier gas.
The surface area of the vermicompost was detected by a surface area and porosimetry analyzer (JW-BK200, Germany). The pH was measured with a pH meter (pH-3C, Shanghai, China), after mixing 1 g of vermicompost with 10 mL of distilled water. The moisture content of the vermicompost-based bioreactor was determined by taking the samples, then weighing and drying it for 12 h at 105°C in an oven to a constant weight. The organic matter content was measured by oven for 6 h at 550°C. The total nitrogen and phosphorus content were determined by spectrophotometry after separating alkaline potassium persulfate and sodium hydroxide. Total sulfur was analyzed by alkali fusion ion chromatography (Shimadzu, Japan) [30]. Elemental sulfur was measured with spectrophotometry. The forms of inorganic sulfur were determined by sequential extraction, based on the method described by Hu et al. [31].
Modifications to the methods were as follows: (1) Water-soluble sulfate (H 2 O-S): 1 g sieved dry sample was placed into a centrifuge tube; the sample was mixed with 10 mL ultrapure water (ratio 1 : 10); the mixture was shaken for 1 h at room temperature (20 ± 2°C); the mixture was then centrifuged at 5,000 × g for 10 min; the supernatant was filtered through a 0.45 µm membrane; and H 2 O-S was determined by ion chromatography.
(2) Adsorbed sulfate (Adsorbed-S): after centrifugation, the supernatant was discarded and then 10 mL of 0.032 mol L −1 KH 2 PO 4 was added (this was also from a 1 g sample); and determine adsorbed-S by ion chromatography.
(3) Hydrochloric acid-soluble sulfur (HCl-Soluble-S): the sample without supernatant was extracted with 10 mL of 2 M HCl. S-free activated charcoal (100 mg) was added to 10 mL of the HCl extracts to eliminate soluble organic sulfur and the associated organic S from the 2 M HCl extracts. The amount of the HCl-Soluble-S for 1 mL extracting solution was then analyzed by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7900, USA) with hexapole collision cell technology. Extraction process of adsorbed-S and HCl-Soluble-S was the same as H 2 O-S.

Microbiological analysis
The microbial community structure in the media influences the performance of the bioreactor and the operational costs. Hence studying the microbial community of odorants can help to better understand the functions of different bacterial species and is significant in optimizing the H 2 S removal process. Changes in the microbial community of different layers of vermicompost-packed bioreactor were analyzed by high-throughput sequencing.

DNA extraction and purification
Genomic DNA was extracted from vermicompost using the Power Soil® DNA Isolation kit (MOBIO QIAGEN, USA) according to the manufacturer's instructions. The purity and concentration of the total gDNA were detected by 1% agarose gels electrophoresis. The qualified DNA samples were stored at −20°C prior to use. DNA was diluted to 1 ng µL −1 using double-distilled water prior to use as a PCR template.

PCR amplification and high-throughput 16S rRNA gene sequencing
The V3−V4 hypervariable region of the 16S rRNA gene was targeted for amplification using the primers 341F Amplification cycling consisted of an initial denaturation at 98°C for 1 min; followed by 30 cycles at 98°C for 10 s, 50°C for 30 s, and 72°C for 30 s; and finally 72°C for 5 min. The PCR products were mixed with the same volume of 1× loading buffer (contained SYB green) and then detected by 2% agarose gel electrophoresis. PCR products were mixed in equidensity ratios. The mixed PCR products were purified using the Gene JETTM Gel Extraction Kit (Thermo Scientific, USA), as specified by the manufacturer. Sequencing libraries were prepared with the Ion Plus Fragment Library Kit 48 rxns (Thermo Scientific, USA) and quantified by using a Qubit ® 2.0 Fluorometer (Thermo Scientific, USA). Amplicon sequencing was performed by Novogene Bioinformatics Technology Co., Ltd (Beijing, China) on the Ion Torrent S5™ platform.

Sequence analyses
The raw reads were optimized by length, quality, primer, and barcode mismatches. Subsequently, high-quality tagged sequences were gained after chimera detection and removal using the QIIME (Version 1.7.0). The sequences were then clustered into an operational taxonomic unit (OTUs) at 97% similarity using UPARSE package. The representative sequence for each OUT, which was the most abundant sequence, was used for taxonomic classification. Using the RDP classifier (version 2.2), the taxonomic information of the sequences was annotated separately. To indicate the microbial diversity in vermicompost, the α-diversity indices were quantified in the light of OUT richness.

Statistical analysis
Significant differences were analyzed by one-way analysis of variance (ANOVA) at 95% confidence level using the SPSS 22.0 statistical software. The Shannon index, Chao1 species richness estimators, and beta diversities of the bacterial communities were calculated using QIIME (Version 1.9.1, http://qiime.org/index.html/) [32]. Cluster analysis of the unweighted and weighted UniFrac distance metrics of the OTUs was performed by principal coordinate analysis (PCoA).
Ethical approval: The conducted research is not related to either human or animal use. respectively. The standard deviation for lower gas flow rates verifies the stability in the biofilter operation [16]. An increase in elimination capacity of the biofilter is due to the elevated inlet concentration or flow rate and plays a role in the treatment efficiency. The loading rate and the elimination capacity are used to reflect the performance of bioreactor. Researches have shown that the removal efficiency decreases in high loading rates, because H 2 S molecules do not have enough time to diffuse inside the biofilm, and the biofilm cannot fully degrade the pollutant [13,33]. The H 2 S elimination capacity and removal efficiency as a function of the inlet H 2 S loading was investigated and is shown in Figure 2b. The H 2 S removal efficiency abated gradually which was affected by the inlet H 2 S loading rate of 5.6-23.3 g m −3 h −1 at a constant flow rate of 0.35 m 3 h −1 , Moreover, there existed a near-linear relation between the elimination capacity and the inlet H 2 S loading rate. In addition, the maximum observed elimination capacity was 20.2 g m −3 h −1 , and bioreactors reached above 92.4% when the inlet H 2 S loading was less than 15.1 g m −3 h −1 . However, the elimination capacity of the vermicompost-mediated bioreactor slowly improved and lower efficiencies were obtained when a higher concentration of H 2 S and an increased flow rate were used. The decreased removal efficiency at the high loading rate was probably due to effective biomass and gas-liquid mass transfer limitations in the bioreactor since biological activity could be inhibited when the substrate conversion ability of the biomass was near the maximum [33]. In addition, exposure to high concentrations of H 2 S could result in toxicity for the vermicompost microbial community. Hence, the vermicompost-mediated bioreactor has an increased ability to remove H 2 S from contaminated air and may be more suitable for practical application.
The removal rate of contaminants from waste gas in a biofilter was correlated with concentration [34]. Hydrogen sulfide samples from different bioreactor thickness (1 # -80 mm, 2 # -160 mm, 3 # -240 mm, and 4 # -320 mm) were periodically investigated to determine the distribution of H 2 S. The influence of thickness on the removal efficiency of H 2 S and changes in concentration are demonstrated in Figure 3. The results showed that the removal efficiency of H 2 S could be notably enhanced with increasing vermicompost thickness but the concentration of hydrogen sulfide reduced gradually. The removal efficiency increased because the path length and the contact area that hydrogen sulfide passed through in the vermicompost increased with the greater thickness. Thus, H 2 S molecules had enough time to diffuse inside the biofilm and accelerate biodegradation. Different packing material thicknesses contributed differently to the H 2 S removal efficiency. For instance, the removal efficiency at 80 mm made up over 50% and was the main area of deodorization. Furthermore, the removal efficiency of H 2 S decreased at 80 mm and increased for the rest of the thickness. The H 2 S contaminated air stream first entered the bioreactor segment #1 (left side of the bioreactor) through the inlet and then successively passed through segments 2 to 4 (right side). When the gas flow rate remained constant, the H 2 S loading rate relied on the concentration. Consequently, the bioreactor achieved the highest H 2 S load in segment #1 when compared to the other segments. The different loads along the bioreactor resulted in differences in the elimination capacity for different segments.
Similar observations have been achieved in other studies [29,35].

Effect of pH changes
The removal of H 2 S using biological methods mainly depends on the metabolic activity of the sulfur-oxidizing microorganisms that are present in the packing material and biological metabolism is inseparable from the enzymatic catalysis [36]. Furthermore, enzymes can only work at certain pH values. Hence, pH also plays an important role in biofilter performance. Results showed that the pH decreased from 7.15 to 5.18 with increasing feed gas flux during the entire experimental period without additional treatment and there was a significant drop in the pH of T1 (Figure 4). However, no significant difference was obtained from T4. The decrease in pH might have been caused by sulfide oxidizing bacteria in the vermicompost that adsorbed and degraded the hydrogen sulfide, which was biologically oxidized to H + , sulfate, and sulfuric acid as presented in equations (1)-(4) [37]. Furthermore, this indicated that the pH change resulted in decreased biological activity and a change of transformation in metabolism. Omri et al. [33] showed that a pH drop resulted from oxidation by SOB and the biodegradation of H 2 S gas was achieved in the bottom and middle layers. Neisi et al. [38]. reported that H 2 S was efficiently removed from a septic tank with a vermicompost biofilter in the pH range of 4.7−7.8 when the H 2 S concentration varied between 33 and 54 ppm.

Analysis of metabolic products
Sulfur fertilizer has become the fourth fertilizer following N, P, and K fertilizers applied in agriculture. Sulfur exists in soils in varieties of inorganic and organic species. Inorganic sulfur is more mobile, it can be divided into SO 4  ) [40]. Hence, the determination of inorganic sulfur in different vermicompost samples was performed before and after the experiment. Compared with CK, the H 2 O-S and adsorbed-S concentration in vermicompost samples increased significantly and accumulated the most in T1, whereas HCl-Soluble-S decreased slowly ( Figure 5). In addition, adsorbed-S is inorganic sulfur that is kept on the surface of soil colloids by anion exchange and coordination adsorption, which has strong stability. The order of adsorbed-S content in the different vermicompost samples was T1 > T2 > T3 > T4. This was mainly because the pH of T1 was lower than the others and this less-acidic environment was attributed to SO 4 2− accumulation; therefore, T1 was the main area of deodorization. The increase of sulfate indicated that H 2 S was degraded by the vermicompost and that H 2 S removal by the vermicompost in the bioreactor was due to a combination of adsorption and biodegradation mechanisms. Vermicompost is rich in organic carbon, mineral nutrients, and inorganic impurity catalysts (iron oxide, calcium oxide, and magnesia) ( Table 1), which could facilitate H 2 S oxidation and convert it to S 0 in the presence of oxygen [37]. Meanwhile, S 0 would be oxidized slowly to sulfate as shown in equation (4). Hence, a yellow substance, which was deposited on the vermicompost after certain running time, was determined to be elemental sulfur (10.85−16.58 mg g −1 ) through analysis. These results are in correspondence with the literature reports [29,41]. Thus, the vermicompost with H 2 S removed can become a sulfur fertilizer resource, offering nutrients to the growth of plants and enriching and improving the soils as well.

Bacterial communities developed in vermicompost for removal hydrogen sulfide
3.4.1 α-diversity and β-diversity of the bacterial communities Diversity indices of vermicompost bacterial communities and the control during hydrogen sulfide treatment are presented in Table 2 [27]. Five samples were sequenced and subjected to quality filtering. Each sample had an average of 847 high-quality OTUs with coverage rates above 99% ( Table 2). This indicated that the sequencing depth was reasonable and truly reflected the diversity of the bacterial community. Chao 1 indices simply indicate the number of species in a community, and the higher values reflect the more rich species in an environmental system. Shannon diversity index indicates some relation between the number of species and number of individuals, and truly reflects the diversity of bacterial community [42]. As shown in Table 2, in contrast to the control treatment, vermicompost showed relatively low Shannon and Chao 1 indices after treating hydrogen sulfide and suggested that these samples had lower diversity and richness than the control. Furthermore, this indicated that the bacterial community in different segments significantly varied. These results were closely related to gas admission. The contaminated air stream was  directed and the H 2 S concentration decreased horizontally in the bioreactor from left to right; therefore, microorganisms of segment #1 were first exposed to higher concentrations of H 2 S, while the higher H 2 S concentration and sulfuric acid produced by the bacteria had a great toxic effect on the microorganisms. Thus, species richness and community diversity were the lowest in T1.
The similarity/dissimilarity of the 16S rRNA gene sequences among the different samples was determined by the weighted UniFrac clustering method. PCoA results based on fast UniFrac analysis showed that the bacterial communities in different vermicompost samples were significantly different (Figure 6a). However, T2 and CK had similar microbial communities.

Bacterial community composition
As illustrated in Figure 6b, the bacterial community compositions differed significantly in each sample. A total of 41 phyla were identified in the samples. The five most dominant phyla were Proteobacteria, Gemmatimonadetes, Bacteroidetes, Actinobacteria, and Firmicutes, accounting for 82.1-96.6% of the bacterial rRNA gene sequence abundance. Proteobacteria was identified as a prominent member representing sulfur-oxidizing bacteria, which were mainly responsible for the degradation of hydrogen sulfide [43]. This phylum was the single largest group in the vermicompost, accounting for 85.0%, 46.1%, 62.0%, and 44.0% in samples CK, T1, T2, T3, and T4, respectively, after 106 days of operation. In contrast, compared to the control treatment, Gemmatimonadetes decreased in different vermicompost samples and only accounted for 2% in T1. The metabolic capacity of phylogenetically related microbial species is similar [8]. Thus, community structures at the genus level were analyzed to determine the main reasons for the differences in vermicompost from different spatial distributions. The results showed that the genus Rhodanobacter had the highest abundance in the four samples used for hydrogen sulfide treatment, accounting for an average of 28.7% of the bacterial 16S rRNA gene abundance (Figure 7). The genus Rhodanobacter is a Gram-negative, aerobic, yellow-colored bacteria, belonging to the family Xanthomonadaceae of the phylum Proteobacteria [44], and has been shown to remove hydrogen sulfide [45]. Additionally, the abundance of Halothiobacillus (mean 5.0%), Thiobacillus (mean 3.5%), and Mizugakiibacter (mean 2.7%) was significantly higher to those in the control group. However, the abundance of Bacillus reached its maximum value in the control treatment and then decreased distinctly in the vermicompost. The genus Halothiobacillus belongs to the γ-subclass of the Proteobacteria (family Halothiobacillaceae) and it is a mesophilic, obligately chemolithoautotrophic, sulfur-oxidizing bacterium. Halothiobacillus cells grow chemolithoautotrophically using sulfur, sulfide, sulfite, thiosulfate, and tetrathionate as electron donors and carbon dioxide as a carbon source [46]. Elemental sulfur production obtained from the bioreactor packed with vermicompost was higher, which indicated that vermicompost contributed to the growth of Halothiobacillus sp. Previous reports have found that the genus Halothiobacillus was the dominant bacterial community member in high salinity wastewater containing sulfide and a biotechnology-based deodorizing system containing high concentrations of hydrogen sulfide [46,47]. Thiobacillus is a colorless, strictly aerobic, chemolithoautotrophic sulfuroxidizing bacteria that being able to oxidize sulfide to elemental sulfur, sulfite, and thiosulfate. Furthermore, it has been identified as the dominant bacterial community member in desulfurization systems [48,49]. The abundance of Thiobacillus increased with increasing vermicompost thickness. The results also indicated that segment 4# was more suitable for the growth of this genus during the late period. This could be connected with the fact that Thiobacillus species was able to effectively degrade H 2 S [50]. In addition, the different H 2 S concentrations resulted in variations in the bacterial community and abundance as the thickness of the bioreactor increased. The relatively high bacterial diversity that was observed in the vermicompost contributed to a good H 2 S removal performance.

Conclusions
A vermicompost-packed biofilter effectively removed hydrogen sulfide in contaminated air. The H 2 S removal performance varied with the gas flow rate, concentration, thickness, and microbial community composition. Accordingly, different forms of sulfate and sulfur were identified as biodegradation products. It is conceivable that the vermicompost with H 2 S removed contained relatively high sulfur, which can improve its fertilizer value. Moreover, high-throughput sequencing results indicated that the different spatial levels of vermicompost exhibited differences in bacterial diversity. Rhodanobacter, Halothiobacillus, Thiobacillus, and Mizugakiibacter were identified as predominant bacterial genera in the final vermicompost. Studying the microbial communities associated with different biofilter layers can help to better understand the functions of different bacterial species and is significant in optimizing the H 2 S removal process. Removal of H 2 S by vermicompost biofilter and analysis on bacterial community  729