First reported quantitative microbiota in different livestock manures used as organic fertilizers in the Northeast of Thailand

Northeastern Thailand relies on agriculture as a major economic activity, and has used high levels of agrochemicals due to low facility, and salty sandy soil. To support soil recovery and sustainable agriculture, local farmers have used organic fertilizers from farmed animal feces. However, knowledge about these animal fecal manures remains minimal restricting their optimal use. Specifically, while bacteria are important for soil and plant growth, an abundance and a diversity of bacterial composition in these animal fecal manures have not been reported to allow selection and adjustment for a more effective organic fertilizer. This study thereby utilized metagenomics combined with 16S rRNA gene quantitative PCR (qPCR) and sequencing to analyze quantitative microbiota profiles in association with nutrients (N, P, K), organic matters, and the other physiochemical properties, of the commonly used earthworm manure and other manures from livestock animals (including breed and feeding diet variations) in the region. Unlike the other manures, the earthworm manure demonstrated more favorable nutrient profiles and physiochemical properties for forming fertile soil. Despite low total microbial biomass, the microbiota were enriched with maximal OTUs and Chao richness, and no plant pathogenic bacteria were found based on the VFDB database. The microbial metabolic potentials supported functions to promote crop growth, such as C, N and P cyclings, xenobiotic degradation, and synthesis of bioactive compounds. Pearson’s correlation analyses indicated that the quantitative microbiota of the earthworm manure were clustered in the same direction as N, and conductivity, salinity, and water content were essential to control the microbiota of animal manures.


Materials and methods
Livestock fecal manure collections. In total, 13 livestock fecal manure samples (E, PC1-3, PCO, CC, DC, MB, G1, G2, D, S and R), each with three independent random samplings at sites, as representative local organic fertilizers, were collected in the Phuparn Royal Development Study Centre area, Sakon Nakhon province, Thailand, in August 2018, between 11.00 and 15.00 h. This centre is a learning source for farmers to improve their own lands, in term of not only planning of agriculture (specifically rice, mushroom, para rubber, economic vegetables, mulberry, etc.) and livestock farming, but also water source and forest rehabilitation.

Nitrogen (N), phosphorus (P), potassium (K), and fecal organic matter (FOM) compositions.
The total N, P, and K contents in the fecal samples were measured following published protocols using a Rapitest Soil Test Kit 1601 (Luster Leaf Products, Inc., Illinois, USA) 20 . In brief, the fecal sample (5 g) was diluted 1:5 (w/v) in sterile water, mixed by vortexing for a minimum of 1 min, and allowed to settle overnight. The liquid suspension was harvested: 10 mL was used for the NPK measurements while 20 mL was used for the physiochemical properties. The 10 mL for NPK measurements was split into three aliquots, and the respective N, P or K powder (0.3 g) (Luster Leaf Products, Inc., Illinois, USA) was added to one aliquot (1 mL) and mixed by shaking for 30 s. The suspension was settled for 15 min, and the developed color was used to determine the N, P, or K amount (µg) based on comparison with the Rapitest Soil Test Kit 1601 manual's color chart. In addition, the FOM content was measured in 1 g of each fecal sample using a Soil Organic Matter Test Kit (Kasetsart University Research and Development Institute, Bangkok, Thailand). Three independent measurements were performed to compute mean ± S.D.
Physiochemical properties. The  www.nature.com/scientificreports/ weighted (wet weight; WW) and then dried in a microwave oven at 105 °C for 24 h 21 to obtain the dry weight (DW). The percent water content was computed from 100 × (WW-DW)/DW. Three independent measurements were performed to compute mean ± S.D.
Metagenomic extraction, and quantification of total bacteria copy number, nirS and alkB. Each sample (0.25 g) was extracted for metagenomic (microbial genomic) DNA using a DNeasy Pow-erSoil Kit, following the manufacturer's instructions (Qiagen, Maryland, USA). The quantity and quality of the extracted metagenomic DNA was analysed by 1% (w/v) agarose gel electrophoresis and nanodrop spectrophotometry (A260 and A260/A280, respectively). For a quantitative count of the total bacteria in copy unit, the 16S rRNA gene qPCR was performed using the universal primers 1392F (5′-GYA CAC ACC GCC CGT-3′) and 1492R (5′-GGT TAC CTT GTT ACG ACT T-3′) 22 . The qPCR thermocycling conditions were 95 °C 5 min, followed by 40 cycles of 95 °C 30 s, 55 °C 45 s and 72 °C 45 s, and ended with a melting curve analysis to validate a single proper amplicon peak (i.e. neither primer-dimer nor non-specific amplification) 23 . The reference for copy number computation was Escherichia coli, in which the 100 base pair (bp) 1392F-1492R amplicon fragments were cloned into pGEM-T-Easy Vector (Promega, Wisconsin, USA) and the recombinant plasmids were transformed into E. coli DH5α for expression. The inserted fragments were verified by colony PCR using the primers M13F (on vector) and 1492R (inserted fragment). Ten-fold serial dilutions of the extracted plasmids (10 4 -10 8 copies/μL) were used as the references for the bacterial copy number computation based on Eq. (1) 23 : The reference DNA and metagenomic DNA (1 ng) were quantified by qPCR for bacterial copy number estimate using primers 1392F and 1492R, and iQ SYBR Green Supermix (Bio-Rad, California, USA) in a 20 µL total volume per well in a 48-well plate using the PCRmax Eco 48 real time PCR system (PCRmax, Staffordshire, UK). Three replicates were performed per reaction. For quantification of cytochrome-containing nitrate reductase (nirS) and alkane monooxygenase (alkB), established primers nirSF: 5′-GTSAACGTSAAG GAR ACSGG-3′ and nirSR: 5′-GASTTC GGR TGSGTC TTG A-3′, and alkBF: 5′-AAC TAC ATC GAG CAC TAC GG-3′ and alkBR: 5′-TGA AGA TGT GGT TGC TGT TCC-3′, along their thermal cycling parameters, were used 24,25 . 16S rRNA gene V3-V4 library preparation and MiSeq sequencing. PCR amplification of the 16S rRNA gene at the V3-V4 region was performed using the universal prokaryotic primers 515F (5′-GTG CCA GCMGCC GCG GTAA-3′) and 806R (5′-GGA CTA CHVGGG TWT CTAAT-3′) with appended adaptor and barcodes sequences as previously reported 26,27 . Briefly, each PCR reaction was comprised of 1 × EmeraldAmp GT PCR Master Mix (TaKaRa), 0.3 μM of each primer, and 50-100 ng of metagenomic DNA in a total volume of 50 μL. The PCR conditions were 94 °C 3 min, and 25 cycles of 94 °C 45 s, 50 °C 1 min and 72 °C 1 min 30 s, followed by 72 °C 10 min. A minimum of two independent PCR reactions were performed and pooled to prevent PCR stochastic bias. Then, the 381-bp amplicon was excised after agarose gel resolution and purified using a Pure-DireX PCR Clean-Up & Gel Extraction Kit (Bio-Helix, Keelung, Taiwan) prior to quantification using a Qubit 3.0 Fluorometer and Qubit dsDNA HS assay kit (Invitrogen, Waltham, USA). Finally, 180 ng of each barcoded amplicon product was pooled for sequencing using the Miseq300 platform (Illumina, San Diego, CA, USA), along with the sequencing primers and index sequence 26 . Sequencing was performed at the Omics Sciences and Bioinformatics Center, Chulalongkorn University (Bangkok, Thailand).
Bioinformatic analyses for bacterial microbiota diversity and potential metabolisms. Raw sequences were processed according to Mothur version 1.39.1′s standard operating procedures for MiSeq 28 . Processes of screening for the quality sequences included removal of (1) short read lengths of ≤ 100 nucleotides (nt) excluding primer and barcode sequences, (2) ambiguous bases ≥ 8, (3) chimera sequences and (4) homopolymer of ≥ 8 nt. The quality sequences were aligned against the 16S rRNA gene databases SILVA version 132 to remove sequences of mitochondria, chloroplast, and eukaryotic lineages 29 , and Greengenes version 13.8 for prediction of taxonomy 30 . The sequences were classified into operational taxonomic unit (OTUs) based on the naïve Bayesian taxonomic method and default parameters (sequence similarity in OTU clustering was 78% for phylum, 88% order, 91% class, 93% family, 95% genus, and 97% species) 31 . Samples were normalized to an equal sequencing depth (7,945 quality sequences per sample), and the count of total bacteria from the 16S rRNA gene qPCR data were analyzed together with the percent microbiota composition from the 16S rRNA gene sequencing to yield the quantitative microbiota [32][33][34] .
Mothur version 1.39.1 was used to estimate the sequencing coverage (Good's coverage index), rarefaction curve, alpha diversity (individual sample diversity: Chao richness and Shannon diversity), and beta diversity based on thetayc dissimilarity coefficients among the samples' quantitative microbiota and two-dimension nonmetric multidimensional scaling (NMDS) 28,35 . One-way ANOVA test was used to analyze the statistical significance for alpha diversity (p < 0.05), and Analysis of molecular variance (AMOVA) was used to analyze the statistical significance for beta diversity (p < 0.001). The frequencies of plant symbiosis bacteria and pathogenic bacteria were identified from the respective list of plant symbiosis (growth promoting) and plant pathogenic bacteria available in the Virulence Factor Database (VFDB) 36,37 , and compared among manure samples. Functional profiles of the microbial communities were predicted from the quantitative microbiota using PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States), then the communities were clustered from the predicted functional profiles by unweighted pair group with arithmetic mean (UPGMA), and the functional profiles were statistically compared using STAMP (Statistical Analysis of Metagenomic Profiles) 38 .

Results
Nutrient, FOM, and physiochemical contents of the fecal manures. For the three main fertility nutrients (N, P and K) in the fecal manures (Table 1a) the E manure contained the highest N level at 80 µg/g, 16-fold more than the other samples. In addition, it had a high P level of 100 µg/g (along with PC1 and PC3) compared with an average of 34.66 µg/g for the others, and a moderate K level of 150 µg/g (others ranged from 125.00 to 350.00 µg/g. The FOM content ranged from 1.00 to 1.50% for all the animal manures (Table 1a). For the general physiochemical properties, chicken manures showed a relatively high salinity (Table 1b: 3.52 ± 0.28 to 4.01 ± 0.64 ppt), and these ionic salts caused a high conductivity (11.27 ± 0.69 to 12.79 ± 0.08 mS/cm) and high water saturation (73.29 ± 0.40 to 80.56 ± 3.17 mS/cm), giving less available water for plant roots. A moderate level of salinity, conductivity and water content, such as those in E, are generally preferred for farming soil. The water content of the manure was lowest in G2, goats fed with P. purpureum (23.75 ± 3.07%), but increased two-fold in G1 goats fed with D. eriantha (49.35 ± 6.08%), suggesting a different type of grass feed might confer different fecal liquid and dry contents, or the chemical composition influenced the water activity in the goat's gut. The R manure also had a low water content (41.53 ± 5.57%). The visual observation of each manure was in accord with the reported water content, with a dry and hard texture (insufficient moisture) for G1, G2, and R manures, and a wet texture for the chicken, cattle, pig, and buffalo manures. Too much water in the soil can adversely affect aeration, where the plant's roots receive insufficient oxygen and rot. Moreover, most plants prefer a slightly acidic soil pH of 6.2-6.8, which was found in the R, MB Table 1. Characteristics of (a) NPK and FOM, and (b) physiochemical levels of the different animal fecal manures. Different lowercase letter superscripts within the same column denote a significant difference (ANOVA, p < 0.05). 1 Data are shown as the mean ± SD of three independent measurements. 2 Measurement could not be specified when < 5 µg/g. 3  www.nature.com/scientificreports/ and E composts (Table 1b). The manure pH can change through reactions such as organic matter decomposition that affects the availabilities of NPK 39 . From the overall comparison of the NPK nutrients and physiochemical properties, the fecal manures from the Phuparn chicken species might be less suitable as a fertilizer because of the poor N level, and high level of salts and conductivity. Indeed, the manures from most species contained a low N level, and those from goat and rabbit also had a relatively low water content (insufficient water for a plant to absorb nutrients from the soil through the roots, to the trunk, leaves and fruits). A moderate water content in the manure was appropriate. But of course, a somewhat low water level could be adjusted by adding water. The biophysiochemical quality analyses of the manures suggested that the E manure would be appropriate as a fertilizer because of the enhanced NPK levels, low salinity, low conductivity, slightly acidic-neutral pH, and a moderate water content (Table 1b).
Quantification of total bacteria in the fecal manures. The total number of bacteria in copies/g DW fecal manure was derived from the 16S rRNA gene qPCR. Figure 1a showed the data were consistent between independent triplicate samples, and that the E and R manures had the lowest bacterial counts (at 2.1 × 10 9 and 2.5 × 10 9 copies/g DW, respectively). The other animal fecal manures contained more than 5 × 10 9 copies/g DW, and the greatest bacterial load was found in PC2 and MB at 2.6 × 10 10 and 1.9 × 10 10 copies/g DW, respectively, while PC1 and PC3 had a significantly lower level (p < 0.001 [PC1:PC2] and p = 0.001 [PC3:PC2]) (Fig. 1b).
Bacterial taxonomic profiles by 16S rRNA gene sequences. The 16S rRNA gene V3-V4 library preparation and next generation sequencing were successful in that the number of raw and quality reads allowed data normalization (N = 7945 quality reads per sample) that covered > 99% and > 98% of the sequencing coverage of taxonomic compositions at the genus and species levels, respectively ( Table 2). The average Good's coverage indices were 99.54% and 99.43% for the genus and species levels, respectively (Supplemental Table 1). This was consistent with the plateau rarefaction curves, showing the frequencies of OTUs become constant despite increasing sequencing reads, meaning a sufficient sequencing coverage was obtained (Supplemental Fig. 1).
Combining the count of total 16S rRNA gene copies with the OTU percent compositions gave the quantitative number of copies of each OTU in a community. These quantitative microbiota data were then used to compute the alpha and beta diversity measurements. The alpha diversity revealed that the microbiota of E had the relatively most diverse OTUs at both the genus and species levels (Chao richness, Fig. 2a,c), regardless of having the lowest total bacterial count (Fig. 1). The low total bacterial abundance but high diversity in E underlined that the various OTUs of bacteria in E might be present in small numbers when compared to the copy numbers of OTUs in the other animal manures. Note that the alpha diversity obtained by considering the distribution of general OTUs among the different animal manures were similar (Shannon diversity, Fig. 2b,d).  www.nature.com/scientificreports/ Figure 3a described the quantitative number of bacterial phylum OTUs. Proteobacteria, Firmicutes and Bacteroidetes were common in all manure samples. PC2 and CC were predominantly Proteobacteria (at 9.41 × 10 9 and 7.24 × 10 9 copies/g DW, respectively), and MB and also PC2 were predominantly Bacteroidetes (at 8.22 × 10 9 and 9.03 × 10 9 copies/g DW, respectively). Firmicutes were also moderately common in all the manures except for E. The percentage abundance of genera above 1% were demonstrated in Fig. 3b, and was comprised of a total of 23 genera. The genus Ignatzschineria were responsible for the high Proteobacteria levels in PC2, MB and PC1, while the genus Acinetobacter were responsible for the high Proteobacteria levels in the other animal manures except for the E. Manures PC1, PC2 and MB contained all Bacteroides genera. Streptococcus was only found in PCO and was relatively abundant (20.04%). Escherichia were relatively low in R, G2, E, DC and MB, while Treponema were responsible for the high Spirochaetes levels in PC3, PCO and R at 3.63, 3.20 and 6.75%, respectively. Noted that the diversified OTUs in E belonged to Bacteroidetes at 6.28%. Moreover, a proportionate percentage of unclassified OTUs were presented in G1, G2, DC and E at 19.62, 41.86, 21.71 and 27.22%, respectively.
For the plant pathogenic bacteria, high levels were found in the manures that also contained high levels of plant symbiosis bacteria (PC1-3, G1, G2 and CC), except for E, plus the other manures, such as S, D and MB. The E had none of VFDB-listed plant pathogens. Thus, all manures except for E contained plant pathogenic bacteria, and these were from Escherichia, Shigella, Enterococcus, Clostridium, Acinetobacter, Treponema, Staphylococcus and Bacteroides (Fig. 4b). This finding correlated with the percent abundance of genera (Fig. 3b) where, examples, Treponema were relatively high in PCO and PC3 at 2.03 × 10 8 and 1.31 × 10 8 cells/g DW, respectively. Escherichia were relatively low in R and DC at 1.15 × 10 5 and 2.06 × 10 5 cells/g DW, respectively. In contrast, only E did not contain any VFDB-listed plant pathogenic bacteria suggesting that E offers a plant pathogen-free organic fertilizer. www.nature.com/scientificreports/

Relationship among bacterial communities, and statistical correlation with the biophysiochemical properties.
The NMDS demonstrated both the reproducibility of the data between the independent triplicate samples, except for R and PCO. The larger variation in bacterial communities were found generally between the manures from different animal species, while the minor variation in bacterial communities were found between the breeds, or the feeding diets. For examples, the variation in quantitative microbiota profiles between G1, G2 and E, compared with PC1-3 (p = 0.085) and MB (p = 0.59) (Supplemental Fig. 2). Indeed, the quantitative microbiota structures belonging to the six animal manures that had all the VFDB's categorized plant pathogens (PC, S, R, D and MB, except PCO) were rather distant from E (p = 0.101, 0.052, 0.089, 0.104 and 0.099, respectively). Seven parameters (NPK and four physiochemical properties) were analyzed for possible Pearson's correlation with any of the quantitative microbiota structures. The N level was strongly correlated (p = 0.01) to the structures and in the same direction of the E (Fig. 5). The percent water, salinity and conductivity characteristics were significant and associated in the direction opposite to E, and also to the G1, G2 and PCO microbiota structures. On the other hand, the PC1, PC2 and MB microbiota structures were strongly associated with the water content, salinity and conductivity. These parameters are suggested to be important in controlling the diversity of microbiota structures.
Predicted functional profiles from quantitative microbiota. Clustering by metabolic profiles separated most of the PC communities, then PCO and CC, from the E clusters (Fig. 6a). The quantitative microbiota in E demonstrated enhanced levels of COGs defined as metabolism, for instances, carbon fixation, oxidative phosphorylation and photosynthesis pathways compared to the rests (Fig. 6b: p = 3.26 × 10 −3 , 0.01 and 0.012, in order). The degradation functions of toxic compounds (xenobiotics) including styrene, caprolactam, aminobenzoate, nitrotoluene, polycyclic aromatic hydrocarbon and benzoate were also high in E compared to the rest (p = 2.81 × 10 −5 , 8.73 × 10 −4 , 1.38 × 10 −3 , 1.65 × 10 −3 , 0.027 and 0.03, respectively), indicating that the earthworm manure had a greater potential potency to degrade toxic compounds contaminated in the soil.
The KEGG categories of bacterial toxin, Staphylococcus aureus infection, and plant-pathogen interaction were dominated statistically higher in the other manures than E (p = 1.6 × 10 −3 , 2.14 × 10 −3 and 0.011, respectively), in which the findings were correlated with Fig. 4b. Within the same metabolic profile clusters with E ( Fig. 6a) but carried all VFDB listed plant pathogens (Fig. 4b), the metabolic profile of D was selected for comparison www.nature.com/scientificreports/ with that of E to perhaps capture some metabolic differences that might involve pathogens. Consistently, antibacterial compounds (i.e. biosynthesis of penicillin and cephalosporin, and antibiotic secondary metabolites) www.nature.com/scientificreports/ were higher in E (Fig. 6c), supporting the ability of the manure's bacterial community to potentially protect against plant pathogens40. Related plant pathogen functions (bacterial toxin, Staphylococcus aureus infection, and plant-pathogen interactions) were statistically higher in D (Fig. 6c), resembling those in the other manures ( Fig. 6b: all other samples). Moreover, E had enhanced levels for carbon fixation and photosynthesis, which are essential for plant growth (p = 2.93 × 10 −3 and 7.77 × 10 −3 , respectively).

Discussion
Organic manures have been valued as one approach to improve soil quality for planting among Thai local farmers. This study contributes a novel understanding of the assorted range of different local livestock manures by their quantitative microbiota compositions using the advanced next generation sequencing, associated with the manures' nutrients and biophysiochemical properties. The NPK measurements revealed that the E manure had high organic N and P levels, consistent with previous reports that earthworms in soil increased N, P and C nutrients to support soil quality for plant growth 41,42 . The elevation of N and P levels in soil could also shift the soil bacterial communities in a way to support common and copiotrophic bacteria (e.g. Actinobacteria and Alphaproteobacteria), while limiting Acidobacteria and Planctomycetes 19,43 . This is consistently related to our study where, for example, the E manure showed relatively higher Actinobacteria and Alphaproteobacteria levels (4.89 × 10 9 and 2.55 × 10 9 cells/g DW, respectively) than those in the manures of the other animals. Although, the NPK and FOM contents in E might be lower than agrochemical fertilizers, the organic minerals in E are safe for use in the soil and our environmental ecosystems, and can add bacterial diversity to the chemically damaged soil. Selection of the appropriate physiochemical properties (pH, salt, and ions) is essential for soil fertility. The high salinity and electrical conductivity in the PC1-3 manures compared to the others, suggested that chicken feces might not be an appropriate manure because the high ionic salts could interfere with the plant's water absorption and also cause an imbalance in nutrient ions. A dramatically low soil pH (< 5.5) was caused from the long-term chemical fertilizer treatment. Here, none of the tested animal feces (manure) exhibited an acidic pH. Rather, they exhibited a near neutral range pH of 6.2-6.8 in E, R and MB which was optimal for a fertile soil, given the slight decrease of pH in E with its high N nutrient content could be due to an incomplete N cycling activity leading to acidic derivative 44 .
The lower total bacterial biomass, yet higher alpha diversity (Chao, OTUs' richness index) in E compared to the other manures supported the fact that the earthworm gut lacks sufficient enzymes to digest OMs, but rather the earthworm gut selects and stimulates microorganisms (thus the reduced total microbial biomass in the E) to assist in the digestion of OMs. These include denitrifying bacteria that are able to survive in the reduced oxygen condition of the earthworm gut 41,45,46 . Further, the relatively high level of available N and P nutrients in the E manure would increase the species richness by promoting copiotrophic microorganisms 41 . www.nature.com/scientificreports/ Overall, Proteobacteria, Firmicutes and Bacteroidetes were found to be common phyla in all the manures (avg. 16.7, 31.3 and 15.9%, respectively), which was consistent with previous reports. For instances, the fecal microbiota of horse, minipig and conventional pig were mainly comprised of the phyla Firmicutes and Bacteroidetes 23,47 . The low (3.3%) content of Firmicutes, gram-positive bacteria that can survive in desiccated and extreme conditions, found in E supported an amiable fecal compost along with the NPK availabilities, low salinity, and conductivity; hence the low presence of Firmicutes were present. The presence of Proteobacteria might be due to available nutrients (NPK and C) in animal feces to support their growth 19 , where Proteobacteria promote nutrient cycling to support soil bacterial growth.
In addition, the microbiota in E were found to be plant pathogen-free, and were comprised of every genus listed as plant growth promoting bacteria, for examples, Pseudomonas, Streptomyces and Flavobacterium. These bacteria function to assist soil nutrient cycling and plant growth, such as P. fluorescens and P. putida process nitrogen fixation and phosphate solubilization, and Flavobacterium processes (solubilizes) complex organic substrates to simple forms for other species usages 37,48 . Subsequently, the presence of plant growth-promoting bacteria was linked to the ability to support crop growth. Manures PC1-3 also contained high numbers of plant-growth promoting bacteria (except for Streptomyces), but had a number of plant pathogens similar to the other animal manures (except for E). Plant-pathogenic bacteria, such as Escherichia (e.g. E. coli), Shigella and Acinetobacter, come from soil and animal fecal contaminations. Many species of Escherichia and Shigella synthesize toxins, and some have been reported to be antibiotic-resistant (e.g. Acinetobacter baumannii), reducing the success of plant treatment by antibiotics 49 . Treponema cause chronic venereal disease (syphilis). Thus, even the minor abundance of some plant pathogens in fecal composts might be hazardous as an organic manure to enrich soil fertility for crop growth. Consistently, previous studies reported that E enhance plant growth and soil biodiversity by promoting beneficial bacteria, which could directly benefit plants via production of plant growth-regulating hormones and enzymes, or indirectly via secondary metabolites that control plant pathogens, nematodes and other pests 40 . However, our study analysis was limited to bacteria, while plant pathogens may be fungi.
In support of E, we observed rice and lemon planting results in our sampling local farms, and found that rice kernels supplemented E were relatively healthy with fewer leaves, while those supplemented with CC, for example, were withering with many leaves. Furthermore, many leaves could pose a risk for one of the most rice pest named brown planthopper that directly damages rice cultivars by feeding and transmitting two viruses, ragged stunt virus and rice grassy stunt virus. For lemon cultivars, using E was found relatively high in flower stalking and fruiting (Chanabun, personal communication). Studies reported that earthworms' guts (including gut microbial communities) can digest various types of organic and even agro-industrial wastes into the reduced organic carbon forms, which represent the more ready-to-use N, P and K forms for plants and soil microbiota 50 . The vermicompost was thereby considered a powerful biofertilizer in sustainable agriculture, containing not only nutrients but also potentially beneficial microbial communities 41 . Indeed, the presence of earthworms represents one biological indicator of naturally arable soil, because its living could accelerate a soil bioconversion process by 2-5 times faster compared to traditional composting (Chanabun, personal communication). Subsequently, we analyzed two key microbial functional gene expressions in E that involve a soil bioconversion process, a cytochrome-containing nitrite reductase (nirS) and an alkane monooxygenase (alkB). The nirS functions in denitrification in nitrogen cycling and alkB functions in alkane biodegradation, respectively. Comparing among E, D and MB, E had the relatively greatest nirS and alkB copies while the total bacterial copy number (16S rRNA) was lowest (Supplemental Fig. 3). This finding was consistent with the high metabolisms of "nitrogen metabolism" and "alkene degradation" predicted from the E's quantitative microbiota using PICRUSt (Fig. 6b,c). The PICRUSt algorithm has been used to predict the functional features of the bacterial community in environmental samples, including soil and fecal manure. For instances, Wang et al. 51 analyzed metabolism functions of bacterial community from microbiota data using PICRUSt in different phases of a swine composting system. In 2019, Meng et al. utilized PICRUSt to analyze metabolic potentials from daily livestock manure microbiota and found that metabolisms of amino acids, lipids and carbohydrates were similar to the previous reports 52 . Therefore, our E quantitative microbiota and functional analyses revealed the understanding of the E. Different earthworm species have previously been reported to contain different bacterial richness and diversity 41 . This study reported the influences that affect the microbiota diversity and microbial metabolic potentials in livestock manures that were found different between animal genera and species (e.g. PCO vs. CC), and animal feeds (G1 vs. G2). Analyses of statistical correlations suggested the importance of available N, conductivity, salinity, and water content, in regulating the quantitative microbiota structures and the microbial metabolic functions. Worthy for our further ongoing studies are effects of (1) these organic manure farmlands compared with long-term chemical farmlands, (2) different species of earthworms, and (3) the appropriate feed diets and beddings for earthworm manures (fecal and urea composts).

Conclusion
Total microbial biomass and microbial community structures were characterized from the manures of different livestock species and breeds, and livestock diets. These fecal manures showed differences in available nutrients and physiochemical properties that could affect the soil fertility for crop growth. The most appropriate livestock manures among the different animal species and feeds analyzed in this study was the E manure due to its appropriate fecal composition, as well as the bacterial communities, which consisted of several plant promoters but no plant pathogens, and metabolic potentials that involved many nutrient cyclings, complex phosphate or other xenobiotics degradation, and bioactive compounds. The favorable microbial community in this E manure is expected to support soil fertility and crop health.