- Metagenome report
- Open Access
Functional and taxonomic classification of a greenhouse water drain metagenome
Standards in Genomic Sciences volume 13, Article number: 20 (2018)
Microbiome sequencing has become the standard procedure in the study of new ecological and human-constructed niches. To our knowledge, this is the first report of a metagenome from the water of a greenhouse drain. We found that the greenhouse is not a diverse niche, mainly dominated by Rhizobiales and Rodobacterales. The analysis of the functions encoded in the metagenome showed enrichment of characteristic features of soil and root-associated bacteria such as ABC-transporters and hydrolase enzymes. Additionally, we found antibiotic resistances genes principally for spectinomycin, tetracycline, and aminoglycosides. This study aimed to identify the bacteria and functional gene composition of a greenhouse water drain sample and also provide a genomic resource to search novel proteins from a previously unexplored niche. All the metagenome proteins and their annotations are available to the scientific community via http://microbiomics.ibt.unam.mx/tools/metagreenhouse/.
All the environments in the world contain millions of microorganisms. However, most of them are uncultivable, difficulting their study under laboratory conditions using traditional culture techniques. In contrast, the rapid development of sequencing technologies and the lower of their associated costs has allowed exploring the microbial composition of almost any ecological niche using metagenomic approaches, ranging from human gut to hot springs [1,2,3]. In this regard, metagenomic approaches have been used to answer two central questions: (i) which microorganisms are present and (ii) what is their functional contribution . Metagenomic has opened the opportunity to find new microbial phyla  and novel protein families in previously unexplored niches , due to uncultivable microorganisms from there. Thus, the metagenomic resource provides the capacity of bioprospecting on the discovery of novel enzymes for research or industrial applications . According to this idea, some new challenges in functional metagenomics, phylogenomics, ecology, and biotechnology have emerged. There are numerous applications of metagenomic analysis, ranging from prevention of diseases to solve industrial problems . In recent years, the scientific community has tried to identify the role that microbial communities have in several disciplines such as human health [9, 10], and industry [11,12,13]. Metagenomics also has been applied to explore the impact of microorganisms in human-constructed niches [11, 14].
A greenhouse is an ecological niche entirely human manipulated, with the continuous exposure to pesticides, fertilizers, antibiotics and different chemicals for research purposes. Thus, subjecting the microbial communities under selective pressures. These effects can be analyzed using the high throughput sequencing methods. This allowed us the possibility to design new strategies for monitoring the microbial evolution of the structure and dynamics in particular human-constructed niches such as a greenhouse, plus comparing it to similar conditions somewhere else and eventually trace back any emerging problem. To our knowledge, this is the first report of a shotgun metagenome from a water sample of a greenhouse drain. Our work aimed to determine the microbial and functional composition of the water from a greenhouse drain. Our results indicated that this environment has low bacterial diversity, mainly dominated by Alphaproteobacteria , which is composed of Rhizobiales and Rhodobacterales orders. Interestingly, we found several antibiotic resistance genes and a functional enrichment for de novo amino acid synthesis in the metagenome.
The sampling site corresponds to the water of a greenhouse drain. The greenhouse is on the top of a building, located at the Institute of Biotechnology (IBt) of the National Autonomous University of Mexico (UNAM), in Cuernavaca City in México. The greenhouse is used for the cultivation of several plant species for research purposes.
Metagenome sequencing information
Metagenome project history
The collected sample was part of a pilot project to identify the novel bacterial composition of the water in the experimental greenhouse drain at the Institute of Biotechnology (IBt) of the National Autonomous University of Mexico (UNAM). We deposited the sequencing reads in the NCBI under the SRA accession number SRR5689218 and SRR5689219 and the Bioproject PRJNA390663. Additionally, the reads were uploaded to the MG-RAST server under the ids mgm4717011.3, mgm4717032.3, mgm4716707.3, mgm4716832.3, mgm4716680.3, mgm4716681.3, mgm4716833.3, mgm4717034.3. For more details see the study information in Table 1.
We collected the sample on 14 September 2015 at 18:00 h (GMT-5) at the IBt (Latitude: 18.918611, Longitude: − 99.234167). In Table 2 the sample information according to the minimal information standards is showed .
Sample preparation, DNA extraction, library generation, and sequencing technology
Sample preparation (collection, transport, and storage)
A sample of 170 ml of water was directly collected from the greenhouse drain and immediately transported to the laboratory, located in the same building. Microbes were obtained by filtering this water through a sterilized PTFE 0.45 μm filter (Cat. 728–2045, Nalgene, NY, USA) using a vacuum pump. After filtration, we extracted the total DNA from the membranes.
DNA extraction (kits used, protocols used)
Total DNA was recovered from the filter membrane by shaking the filter for 5 min in a tube containing lysis solution and beads from ZR Soil Microbe DNA MicroPrep Kit (Cat. D6003 Zymo Research, Irvine, CA, USA). The following steps for DNA isolation were carried out following the manufacturer’s instructions for the ZR Soil Microbe DNA kit. After extraction, we assessed the DNA quality by agarose gel electrophoresis and quantity determined by the Thermo Fisher Qubit High-sensitivity fluorometric assay (Cat. Q32851, Life Technologies, Carlsbad, CA, USA).
Library generation (kits used, protocols used)
We constructed two DNA libraries containing different insert sizes: Drain-A and Drain-B with an insert size of 400 and 2000 bp, respectively (Table 3). Furthermore, different amounts of input DNA were used to construct the libraries: 1 ng for Drain-A and 25 ng for Drain-B. Both libraries were created following the manufacturer’s instructions for the Nextera XT DNA Library Preparation kit (Cat. FC-131-1024, Illumina, CA, USA). First, DNA was fragmented (tagmented) using the Nextera transposase. Second, the tagmented DNA was amplified using 12 PCR cycles to add the Index 1 (i7), Index 2 (i5), and full adapter sequences. The program on the thermal cycler was as follows: 72 °C for 3 min, 95 °C for 30 s; 12 cycles (95 °C for 10 s, 55 °C for 30 s and 72 °C for 30 s) and 72 °C for 5 min. After PCR amplification, both libraries were carefully size selected using Agencourt Ampure XP beads (Cat. A63882, Beckman Coulter, CA, USA) and the size was verified using a DNA Agilent Bioanalyzer 2100 (Cat. 5067–1504, Agilent Technologies, CA, USA).
The Illumina NextSeq 500 Mid Output cell was used for sequencing in a 2 × 150 bp paired-end format, resulting in a total of 7,378,202 of reads for a sum of 11 Gbp of DNA data. Each sample yielded 6,976,736 and 401,466 of reads for Drain-A and Drain-B libraries, respectively (Table 4).
Sequence processing, annotation, and data analysis
Pair-end raw reads were quality filtered using DynamicTrimm . To this end, we eliminated the barcodes and primers, removed the reads containing ambiguous bases and trimmed the sequences with quality >Q20 (6 bp sliding window). We mapped the raw reads against Homo sapiens genome (GRCh38) using BWA with default parameters  to remove human DNA for downstream analysis.
All the quality-filtered reads of the two libraries were used to construct two de novo metagenomic assemblies, one using IDBA-UD  with 20–125 of k-mer length range and other using MetaSpades  with a k-mer range of 21–121 with steps of 10 (Table 5). After that, we used Mummer (nucmer) with a cluster match of 80 nucleotides (contig coverage) and 99% of identity to merge all the contigs of both assemblies . The merged metagenome was selected because it contained a minor number of contigs and the best N50 and L50 among the others. The contigs were validated mapping back the reads using BWA  with default parameters, resulting 79% of the reads mapped back to the final assembly with coverage of 26X . This percentage and coverage are adequate for a metagenome assembly . All the contigs larger than 1 Kb were selected for gene prediction and functional annotation, resulting in 7003 contigs with an N50 and N75 of 4246 and 1807 bp, respectively.
All classified reads (at different taxonomic levels) for each library were merged into a single library (Greenhouse Drain-IBt) to determine their relative abundances. After that, all quality-filtered reads were functionally and taxonomically classified using the MG-RAST server . The annotations are available under the accession numbers mgm4717011.3, mgm4717032.3, mgm4716707.3, mgm4716832.3, mgm4716680.3, mgm4716681.3, mgm4716833.3, mgm4717034.3. The taxonomic and functional classification was performed with MG-RAST server using the RefSeq and SEED subsystem databases with default parameters, respectively (Table 6). Normalized raw count was used to determine the relative abundances of reads for each taxonomic level, using an in-house developed Perl script. Additionally, the reads were also taxonomically classified by Kraken  using the RefSeq bacterial database from NCBI (ftp://ftp.ncbi.nlm.nih.gov/genomes/refseq/bacteria/). Taxonomic abundances were calculated using an in-house developed Perl script based on the number of reads for each taxonomic group. Furthermore, the reads were also functionally annotated by HUMAnN2 using the UniRef90 database. The taxonomic association in HUMAnN2  was performed with Meta PhlAn2 using ChocoPhlAn database.
Final contigs of the metagenome were used to predict 25,735 proteins by FragGeneScan (Table 7) . Out of the total of predicted proteins, 21,700 were functionally annotated by Blast2GO PRO version 2.8 , using BLASTp against NR, Gene Ontology (GOs) and InterProScan version 5.25 . Antibiotic resistance genes (ARGs) were determined using the antibiotic resistance genes database (ARDB).
Shotgun sequence data generated a total of 7,378,202 of reads that were quality processed (see Sequence processing section) and got 7,206,754 of quality reads. Next, we assigned the taxonomy classification to 1,966,261 reads using MG-RAST and to 877,380 reads using Kraken, both of them using the LCA method . Additionally, high-quality reads were used to obtain a de novo assembly consisting of 7003 contigs with an N50 and N75 of 4246 and 1807 bp, respectively. The largest contig of the assembly had 221,208 bp length (Table 5).
After using MG-RAST, we found that Proteobacteria was the most abundant phylum with 91% of the reads, followed by Actinobacteria (2.6%) and Bacteroidetes (2%) (Table 8). Additionally, at the class level, Alphaproteobacteria (74%) was highly present, and Gammaproteobacteria (10%), Betaproteobacteria (6%), and Actinobacteria (3%) showed the lowest abundances. The Rhizobiales and Rhodobacterales orders ( Alphaproteobacteria ) were the most abundant in the metagenome (Fig. 1). Next, we used Kraken to compare the taxonomy classification obtained by MG-RAST. Although Kraken assigned a lower number of reads (877,380) than MG-RAST (1,966,261), we found similar results in taxonomy abundance. Proteobacteria was the most abundant phylum (94%), followed by Actinobacteria (3%) and Bacteroidetes (0.4%) (Table 8). Also, similar to MG-RAST classification, Rhodobacterales and Rhizobiales orders were highly abundant. The most critical difference in taxonomic classification between the two algorithms was the number of orders identified, MG-RAST identified 92 and Kraken 51 orders (Fig. 1). To evaluate if this difference could impact on the microbial alpha diversity metrics, we used the relative abundance tables from MG-RAST and Kraken to measure the richness (Simpson) and evenness (Shannon) using Phyloseq . This analysis showed a Shannon index of 2.04 and 1.99 for Kraken and MG-RAST, respectively (Table 7). In contrast, the Simpson index was 0.75 and 0.69 for MG-RAST and Kraken, respectively. However, these different values between MG-RAST and Kraken were not significant, suggesting that there is no difference between both algorithms for alpha diversity classification. The observed diversity metrics indicate that the greenhouse water drain is not a diverse niche.
Next, to know the encoded functions directly from the reads we classified them using MG-RAST and HUMAnN2. According to MG-RAST, we observed that most abundant functions were related to vitamins, protein biosynthesis and central carbohydrate metabolism (Table 9). In contrast, using HUMAnN2 the de Novo nucleotide biosynthesis, vitamins, and prosthetic groups were the most abundant functions (Table 10).
We assembled the metagenome to get more insight into the protein functional composition of the water sample. In this regard, we used the contigs to predict 25,735 proteins from which only 35.8% had significant blastp match (E-value 10− 5) against the NR RefSeq proteins database. The 64.2% of unknown proteins could represent novel proteins. A total of 14,328 (55.6%) proteins were classified using Gene Ontology (GOs) using Blast2GO and Interpro. We found that the term “transport” in the Biological process category was the most abundant function encoded in the greenhouse metagenome (Fig. 2). To get insights on this observation, we found that of the 2141 transporters annotated by Interpro, the 17.23% (369 proteins) are ABC-type transporters. Interestingly, has been reported that Rhizobiales has enriched the ABC transporter genes in their genomes (Fig. 2) [31, 32]. Furthermore, we use KAAS-KEGG  to identify the pathways containing ABC-transporters. After that, we only found six complete pathways in the greenhouse metagenome such as the vitamin B12 transporter (Fig. 3). Interestingly, the ABC-transporters were also present when we used only the reads for functional analysis (Table 9 and Table 10). Additionally, the hydrolase, transferase, and oxidoreductase were the most abundant GO molecular functions (Fig. 2). This result was in agreement with reports in which rhizobial bacteria associated with the nodules and seed of plants has many genes for these molecular functions in their genomes .
Finally, we searched for Antibiotic Resistance Genes (ARGs) in the metagenome, using the ARDB database . We found a total of 31 ARGs and the most abundant genes were for resistance to spectinomycin (16%), tetracycline (12%) and aminoglycosides (9%) (Table 11). These findings are consistent with previous studies in chicken and vegetable greenhouse soil samples . Although ARG genes are common in most ecological niches more investigation is needed to explore the role that these genes could play in microbial dynamics of human-constructed niches such as the greenhouse. Furthermore, these ARGs genes could be used for further analysis in phylogenomics to aim their evolutionary history and trace the adquision of these genes .
The use of metagenomic approaches to characterize new environments such as a research-greenhouse has the potential to unveil novel bacterial dynamics, enzyme functions, and metabolic pathways. To the best of our knowledge, this is the first report of the bacterial and functional contribution of the water from a greenhouse drain. We consider it exemplifies how the utilization of a metagenomic approach provides a more comprehensive view regarding the structure and functional composition of a bacterial community. Our results indicated that soil Rhizobiales bacteria and their genome functions mainly dominate the greenhouse water drain. This study aimed to identify the bacteria and functional gene composition of a greenhouse water drain sample and also represent a genomic resource to search novel proteins from a previously unexplored niche. Interestingly, we found over 400 proteins containing unintegrated signatures, which are highly conserved domains with unknown function according to Interpro, representing potential novel enzymes. All the metagenome proteins and their annotations are available to the scientific community via http://microbiomics.ibt.unam.mx/tools/metagreenhouse/.
Antibiotic resistance genes database
Antibiotic resistance genes
Institute of Biotechnology
Bikel S, Valdez-Lara A, Cornejo-Granados F, Rico K, Canizales-Quinteros S, Soberon X, Del Pozo-Yauner L, Ochoa-Leyva A. Combining metagenomics, metatranscriptomics and viromics to explore novel microbial interactions: towards a systems-level understanding of human microbiome. Comput Struct Biotechnol J. 2015;13:390–401.
Lopez-Lopez O, Cerdan ME, Gonzalez-Siso MI. Hot spring metagenomics. Life (Basel). 2013;3:308–20.
Wang WL, Xu SY, Ren ZG, Tao L, Jiang JW, Zheng SS. Application of metagenomics in the human gut microbiome. World J Gastroenterol. 2015;21:803–14.
Oulas A, Pavloudi C, Polymenakou P, Pavlopoulos GA, Papanikolaou N, Kotoulas G, Arvanitidis C, Iliopoulos I. Metagenomics: tools and insights for analyzing next-generation sequencing data derived from biodiversity studies. Bioinform Biol Insights. 2015;9:75–88.
Eloe-Fadrosh EA, Paez-Espino D, Jarett J, Dunfield PF, Hedlund BP, Dekas AE, Grasby SE, Brady AL, Dong H, Briggs BR, et al. Global metagenomic survey reveals a new bacterial candidate phylum in geothermal springs. Nat Commun. 2016;7:10476.
Ufarte L, Potocki-Veronese G, Laville E. Discovery of new protein families and functions: new challenges in functional metagenomics for biotechnologies and microbial ecology. Front Microbiol. 2015;6:563.
Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, Huttenhower C. Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12:R60.
Coughlan LM, Cotter PD, Hill C, Alvarez-Ordonez A. Biotechnological applications of functional metagenomics in the food and pharmaceutical industries. Front Microbiol. 2015;6:672.
Critchfield JW, van Hemert S, Ash M, Mulder L, Ashwood P. The potential role of probiotics in the management of childhood autism spectrum disorders. Gastroenterol Res Pract. 2011;2011:161358.
Major G, Spiller R. Irritable bowel syndrome, inflammatory bowel disease and the microbiome. Curr Opin Endocrinol Diabetes Obes. 2014;21:15–21.
Bonifay V, Wawrik B, Sunner J, Snodgrass EC, Aydin E, Duncan KE, Callaghan AV, Oldham A, Liengen T, Beech I. Metabolomic and metagenomic analysis of two crude oil production pipelines experiencing differential rates of corrosion. Front Microbiol. 2017;8:99.
Buie T, Campbell DB, Fuchs GJ, 3rd, Furuta GT, Levy J, Vandewater J, Whitaker AH, Atkins D, Bauman ML, Beaudet AL, et al: Evaluation, diagnosis, and treatment of gastrointestinal disorders in individuals with ASDs: a consensus report. Pediatrics 2010, 125 Suppl 1:S1–18.
Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature. 2006;444:1027–31.
Gomez-Alvarez V, Revetta RP, Santo Domingo JW. Metagenome analyses of corroded concrete wastewater pipe biofilms reveal a complex microbial system. BMC Microbiol. 2012;12:122.
Yilmaz P, Kottmann R, Field D, Knight R, Cole JR, Amaral-Zettler L, Gilbert JA, Karsch-Mizrachi I, Johnston A, Cochrane G, et al. Minimum information about a marker gene sequence (MIMARKS) and minimum information about any (x) sequence (MIxS) specifications. Nat Biotech. 2011;29:415–20.
Cox MP, Peterson DA, Biggs PJ. SolexaQA: at-a-glance quality assessment of Illumina second-generation sequencing data. BMC Bioinformatics. 2010;11:485.
Li H, Durbin R. Fast and accurate short read alignment with burrows-wheeler transform. Bioinformatics. 2009;25:1754–60.
Peng Y, Leung HC, Yiu SM, Chin FY. IDBA-UD: a de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth. Bioinformatics. 2012;28:1420–8.
Nurk S, Meleshko D, Korobeynikov A, Pevzner PA. metaSPAdes: a new versatile metagenomic assembler. Genome Res. 2017;27:824–34.
Delcher AL, Phillippy A, Carlton J, Salzberg SL. Fast algorithms for large-scale genome alignment and comparison. Nucleic Acids Res. 2002;30:2478–83.
Hunt M, Newbold C, Berriman M, Otto TD. A comprehensive evaluation of assembly scaffolding tools. Genome Biol. 2014;15:R42.
Van Belleghem SM, Roelofs D, Van Houdt J, Hendrickx F. De novo transcriptome assembly and SNP discovery in the wing polymorphic salt marsh beetle Pogonus chalceus (Coleoptera, Carabidae). PLoS One. 2012;7:e42605.
Keegan KP, Glass EM, Meyer F. MG-RAST, a metagenomics Service for Analysis of microbial community structure and function. Methods Mol Biol. 2016;1399:207–33.
Wood DE, Salzberg SL. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol. 2014;15:R46.
Abubucker S, Segata N, Goll J, Schubert AM, Izard J, Cantarel BL, Rodriguez-Mueller B, Zucker J, Thiagarajan M, Henrissat B, et al. Metabolic reconstruction for metagenomic data and its application to the human microbiome. PLoS Comput Biol. 2012;8:e1002358.
Rho M, Tang H, Ye Y. FragGeneScan: predicting genes in short and error-prone reads. Nucleic Acids Res. 2010;38:e191.
Conesa A, Gotz S. Blast2GO: a comprehensive suite for functional analysis in plant genomics. Int J Plant Genomics. 2008;2008:619832.
Quevillon E, Silventoinen V, Pillai S, Harte N, Mulder N, Apweiler R, Lopez R. InterProScan: protein domains identifier. Nucleic Acids Res. 2005;33:W116–20.
Huson DH, Auch AF, Qi J, Schuster SC. MEGAN analysis of metagenomic data. Genome Res. 2007;17:377–86.
McMurdie PJ, Holmes S. Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013;8:e61217.
Konstantinidis KT, Tiedje JM. Trends between gene content and genome size in prokaryotic species with larger genomes. Proc Natl Acad Sci U S A. 2004;101:3160–5.
Lopez-Leal G, Tabche ML, Castillo-Ramirez S, Mendoza-Vargas A, Ramirez-Romero MA, Davila G. RNA-Seq analysis of the multipartite genome of rhizobium etli CE3 shows different replicon contributions under heat and saline shock. BMC Genomics. 2014;15:770.
Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M. KAAS: an automatic genome annotation and pathway reconstruction server. Nucleic Acids Res. 2007;35:W182–5.
Peralta H, Aguilar A, Diaz R, Mora Y, Martinez-Batallar G, Salazar E, Vargas-Lagunas C, Martinez E, Encarnacion S, Girard L, Mora J. Genomic studies of nitrogen-fixing rhizobial strains from Phaseolus vulgaris seeds and nodules. BMC Genomics. 2016;17:711.
Liu B, Pop M. ARDB--antibiotic resistance genes database. Nucleic Acids Res. 2009;37:D443–7.
Fang H, Wang H, Cai L, Yu Y. Prevalence of antibiotic resistance genes and bacterial pathogens in long-term manured greenhouse soils as revealed by metagenomic survey. Environ Sci Technol. 2015;49:1095–104.
Davies J, Davies D. Origins and evolution of antibiotic resistance. Microbiol Mol Biol Rev. 2010;74:417–33.
This work was supported by Consejo Nacional de Ciencia y Tecnología (CONACyT) from México with grant SALUD-2014-C01–234188 and by Universidad Nacional Autónoma de México (UNAM) with grant DGAPA PAPPIT UNAM (IA203118). GLL acknowledges the support of CONACyT as a postdoctoral fellow.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
López-Leal, G., Cornejo-Granados, F., Hurtado-Ramírez, J.M. et al. Functional and taxonomic classification of a greenhouse water drain metagenome. Stand in Genomic Sci 13, 20 (2018) doi:10.1186/s40793-018-0326-y
- Shotgun sequencing
- Environmental sample
- Water drain