10 years of Nature Protocols PERSPECTIVE The past,present and future of microbiome analyses Richard Allen White III,Stephen J Callister,Ronald J Moore,Erin S Baker Janet K Jansson Over the last decade,technical advances in nucleic acid sequencing and mass spectrometry have enabled faster and more informative metagenomic,metatranscriptomic,metaproteomic and metabolomic measurements.Here we review key improvements in multi-omic environmental and human microbiome analyses,and discuss developments required to address current measurement shortcomings. agpandCvcmuy2acaped22oThabhk2onmCcain mately 3 5 billio the planet's biosphere.Although microorganisms are known bacterial genomes in amatter of hours rth.such as carbor state of the planet's plant and animal inhabitants.greater than pyrophosphate,also known as pyrosequencing nhigh microb micoomnmicrobomes(efathe totaity of microorganisms and their collective genetic material ts.hi greatly facilitated errors in reading through the complex repeats).and surface studies of complex microbiomes and their functions.Here area loading limitations owing to bead-based DNA molecule )Genome Analyzer (GA),which ofenvironments as well as in our own bodie Nucleic acid sequencing DNA nd RNA ing At the forefront of advances in he pe tion (NextGer Gbp)haploid me ear at $00 peh equencing platforms as they have surpa sed the traditional (http://www.il mina.com/systems/hiseq equencing-system/ nated the fe eral technical ome using the Sanger dideoxynucleotide-based chain (100300-bp paired-end reads)and approach previously endeavor that took throughputs to try and address this challenge.For exampethe in(wthcompeted n9).Current P than the hisea platform which generates billions of reads per o,res NATURE PR0T0C0LS|70L.11N0.11|2016|2049
NATURE PROTOCOLS | VOL.11 NO.11 | 2016 | 2049 10 years of Nature Protocols PERSPECTIVE because of the availability of rapid and inexpensive NextGen sequencing platforms, it is now possible to sequence complete bacterial genomes in a matter of hours11. NextGen sequencing methods have used several different high-throughput platforms. The first was the Roche GS20 454 sequencer, which was based on the polymerase cleavage of pyrophosphate, also known as pyrosequencing12,13. Although 454 sequencing was a key technological advance, and 454 sequencers including the GS20 and GS FLX series machines and reagents were used for over a decade (approximately 2005 to 2016, http:// www.genomeweb.com/sequencing/roche-shutting-down-454- sequencing-business), it had several drawbacks including high cost of sequencing reagents, high homopolymer error rates (i.e., errors in reading through the complex repeats), and surface area loading limitations owing to bead-based DNA molecule deposition that restrict the throughput and number of reads obtained. The second NextGen sequencer was the Solexa (now Illumina) Genome Analyzer (GA), which was introduced in 2006 and incorporated oligonucleotide array flow cells, reversible chain terminators and bridge PCR reactions14. This technology is now routinely used to sequence DNA and RNA extracted from human and environmental microbiomes and can generate >1.8 terabases (TB) of data in a single run. However, the ultimate goal was to sequence >18,000 human genomes (~3 gigabase-pair (Gbp) haploid genome) per year at $1,000 per human genome (http://www.illumina.com/systems/hiseq-x-sequencing-system/ system.html). Illumina currently has several technical platforms including GA, MiSeq and HiSeq machines, with varying sequence read lengths (100-300-bp paired-end reads) and throughputs to try and address this challenge. For example, the maximum read length with overlapping paired reads on a MiSeq platform is ~500-550 bp, but that platform has lower throughput than the HiSeq platform, which generates billions of reads per run (Fig. 1). A relatively new approach developed by Illumina, called TruSeq synthetic long reads or Moleculo, results in longer read lengths (>8 kbp)15, and has facilitated the assembly of highly complex soil microbiomes16 and other biological samples17,18. Initial results from these technological advances are enhancing microbiome assembly into longer contigs16–18. Microbes evolved on Earth approximately 3.5 billion years ago and eventually occupied every habitable environment in the planet’s biosphere. Although microorganisms are known to be responsible for key functions on Earth, such as carbon and nutrient cycling, and determining the health and disease state of the planet’s plant and animal inhabitants, greater than 99% of the trillions of microbes thought to exist have yet to be discovered1. In addition, high microbial diversity has made it difficult to study specific functions carried out by complex microbial communities in microbiomes (defined as the totality of microorganisms and their collective genetic material present in a specific environment such as all microorganisms inhabiting the soil or human gut)2,3. Fortunately, technological advances over the last few decades have greatly facilitated studies of complex microbiomes and their functions. Here we discuss advances related to nucleic acid sequencing and mass spectrometry (MS) analyses that have enabled the exploration and understanding of microbiomes across a range of environments as well as in our own bodies3–6. Nucleic acid sequencing Next-generation sequencing. At the forefront of advances in microbiome research lie the impressive increases in the speed and throughput of nucleic acid sequencing technologies. In particular, there has been a revolution in next-generation (NextGen) sequencing platforms as they have surpassed the traditional Sanger sequencing method that dominated the field for nearly three decades (from 1977 to 2005)7. Sequencing a single bacterial genome using the Sanger dideoxynucleotide-based chaintermination approach previously was a major endeavor that took years to complete8,9. The first bacterial genome to be completely sequenced using the Sanger approach was Haemophilus influenza9 in 1995 (with Escherichia coli10 completed in 1997). Currently, The past, present and future of microbiome analyses Richard Allen White III, Stephen J Callister, Ronald J Moore, Erin S Baker & Janet K Jansson Over the last decade, technical advances in nucleic acid sequencing and mass spectrometry have enabled faster and more informative metagenomic, metatranscriptomic, metaproteomic and metabolomic measurements. Here we review key improvements in multi-omic environmental and human microbiome analyses, and discuss developments required to address current measurement shortcomings. Earth and Biological Sciences Directorate, Pacific Northwest National Laboratory, Richland, Washington, USA. Correspondence should be addressed to J.K.J. (janet.jansson@pnnl.gov) or E.S.B. (erin.baker@pnnl.gov). Received 8 June; accepted 19 July; published online 29 September 2016; doi:10.1038/nprot.2016.148 npg © 2016 Nature America, Inc. All rights reserved
PERSPECTIVE 1x100 length (bp) Over the last five years,the PacBio platform has becomea robust tec the PacBio platform needs micrograms ofigh-molecua-weigh DNA(>40 kbp)for library preparation,limiting the samples it nee 1x10 and increas environmental samples.A maior challenge with these new technologies,however,is obtaining high-quality and large molecular-weight D A in the range of hundreds ofkilob 501 and oche G520 gle-molecule sequ nces offer future promise to the microbiome community. Sequencing complex mi DNA sequ ogy 1x104 functions ofmi bes in habitats around the world and in our own bodies3 4 3.24.Most of the information obtained about microbiomes has b n from the NextGen sequencing 2004 2008 2012 2016 a phylogenetic ma teria an Year t omics)has made it p the complement of functional genes associated with specific agfhepoieonthe aph (a microbial groups in diverse environments.For example,some of s repre our own re has detine the rur tonal gene c ncin pill in the Gulf of h nafrost27.2s shown on the x axis However.a vast majority of these genes have o known function reflecting the immense diversity and biochemical potential of environmental microbiomes remaining to be discovered sequencing technolo onies that s tial for fa Metatrans ng total mrNa (ie. informative microbiome studies indude single-molecule-based metatranscriptomics)reveals which genes are exp sed DNA sequencers.One example is the single-molecule,real-time by specific organisms over spatial and temporal scales. (SMRT)technology from PacBio that relies on tethered DNA Metatranscriptomics ha ffered a wealth wledge abou polym na varie DNA sequence reads of -10-25 kbp and-300 Mbp per sFor exampe Gilbert mctatranscriptomics SMRT cell (Fig1:http://www.pacb.com).Because ofthe use of to determine the seasonal expression patterns of the marine microbiome in the English Chann Recently,we used m uch as DNA platform were highly abundant in the soil under investigation.few mRNA transcripts mapped to their genomes,suggesting g that they were I pla orm doe s not rely on ctually trans nally dormant.By contrast,the F ly active.This rev Nanopore can detect DNA modifications like the PacBio platorm has an average read length of~1-2 kbp,and the longes microbial communities?8 ngth ofle rd Nai ore sequenc its thumb- etry-based about-us/news) understanding of microbialphylogenetic and functional gene 2050 VOL.11 NO.NATURE PROTOCOLS
2050 | VOL.11 NO.11 | 2016 | NATURE PROTOCOLS PERSPECTIVE Over the last five years, the PacBio platform has become a robust technology for sequencing microbiome samples for de novo and metagenomic assembly applications. The caveat is that the PacBio platform needs micrograms of high-molecular-weight DNA (>40 kbp) for library preparation, limiting the samples it can sequence. Oxford Nanopore offers an inexpensive way to potentially sequence very large scaffolds (>50 kbp) and increase genome closures and reconstruction of genomes directly from environmental samples. A major challenge with these new technologies, however, is obtaining high-quality and largemolecular-weight DNA in the range of hundreds of kilobases or even megabase lengths. Oxford and PacBio platforms are also still too low-throughput for large-scale studies, but for sequence assembly applications, these single-molecule sequences offer future promise to the microbiome community. Sequencing complex microbiomes. DNA sequencing technology improvements have enabled many discoveries of the identities and potential functions of microbes in habitats around the world and in our own bodies3,4,23,24. Most of the information obtained about microbiomes has been from the NextGen sequencing of 16S rRNA genes as a phylogenetic marker for bacteria and archaea. In addition, NextGen sequencing of total community DNA (i.e., metagenomics) has made it possible to determine the complement of functional genes associated with specific microbial groups in diverse environments. For example, some of our own research has defined the functional gene compositions in water and sediment samples after the Deepwater Horizon oil spill in the Gulf of Mexico25,26 and in thawing permafrost27,28. However, a vast majority of these genes have no known function, reflecting the immense diversity and biochemical potential of environmental microbiomes remaining to be discovered. Metatranscriptomics. Sequencing total mRNA (i.e., metatranscriptomics) reveals which genes are expressed by specific organisms over spatial and temporal scales. Metatranscriptomics has offered a wealth of knowledge about the expression of microbial genes in a variety of ecosystems, including acid-mine drainage29, human gut30, ocean25,26,31–34 and soil27,28. For example, Gilbert et al.34,35, used metatranscriptomics to determine the seasonal expression patterns of the marine microbiome in the English Channel. Recently, we used metatranscriptomics to deduce which organisms identified from soil genomes were active in soil16. Although Verrucomicrobia were highly abundant in the soil under investigation, few mRNA transcripts mapped to their genomes, suggesting that they were actually transcriptionally dormant. By contrast, the Firmicutes genomes were found to be transcriptionally active. This revealed the utility of metatranscriptomics in validating metagenomics and understanding the relative activities of different members of microbial communities28. Mass-spectrometry-based metaproteomic and metabolomic measurements Although advances in DNA sequencing have enabled a better understanding of microbial phylogenetic and functional gene Single-molecule sequencing technologies. Emerging sequencing technologies that show great potential for faster and more informative microbiome studies include single-molecule-based DNA sequencers. One example is the single-molecule, real-time (SMRT) technology from PacBio that relies on tethered DNA polymerases and zero-mode waveguides to direct light energy through small volumes of liquids19,20. PacBio currently offers long DNA sequence reads of ~10-25 kbp and ~300 Mbp per SMRT cell (Fig. 1; http://www.pacb.com). Because of the use of tethered polymerases, the PacBio platform can detect unusual or modified nucleotide bases without chemical modification during synthesis, such as DNA methylation, owing to the wobble of the polymerase21. The Oxford Nanopore sequencing platform is another emerging and promising single-molecule sequencer. Unlike current platforms, the Oxford platform does not rely on sequencing by synthesis but instead directly sequences nucleic acid molecules by threading the strands through a Nanopore22. Oxford Nanopore can detect DNA modifications like the PacBio platorm, has an average read length of ~1-2 kbp, and the longest maximum read length offered by any sequencer (>90 kbp) (Fig. 1)22. The main benefit of the Oxford Nanopore sequencer is its thumb-drivesized format that can be analyzed on a personal labtop computer in real time using wireless technology (https://nanoporetech.com/ about-us/news). Read length (bp) 100 1,000 10,000 Illumina HiSeq X Illumina GAIIx (2010) Illumina GAIIx (2007) Illumina MiSeq v3 Illumina Moleculo PacBio RS II Oxford MinION Illumina HiSeq v4 Roche GS FLX+ Roche GS FLX Titanium Roche GS FLX Roche GS20 1 x 1010 1 x 108 1 x 106 1 x 104 Number of sequences per lane, cell or plate Year 2004 2008 2012 2016 Figure 1 | Advances in sequencing technologies over the last decade. Average read length in base pairs is log10-transformed and is the fill factor of the points on the graph (larger points equal longer read lengths). Throughput (number of sequence reads) is represented on the y axis as follows: Roche 454 technology is per plate (i.e., picotiter plate), Illumina technology is per lane, PacBio and Oxford MinION is per cell. The year of introduction of each sequencing machine represented is shown on the x axis. npg © 2016 Nature America, Inc. All rights reserved
PERSPECTIVE compositions in microbiomes,it is also desirable to know which are pro ow per metabolites prpduced by microbiom s in differ des ha mainly been achieved using MS In MSanalyses of microbiome ttheir mass-to- a& The introduction of electrospray ionization(ESD)in1984 asurements for assessment lenge with ESI is that it is performed 10 torr This p sure difference accounts for more than nine orders of ge ion lo Hesigns using ion ith bette vesing the perform ution is s re in the last while decre asing the pressure The improvements in these areas resulted on thex axis. ctively.The instrument and year it was intro eo ar cult nd tho ms (Fig.3).Fo challenges required the improvement of the ample in the early 2000s Banfield and co-workers were the respect to resolution,accuracy and MS/MS speed.Quadrupole, first to use a combination of sequencing and MS approaches to on trap,tim -of-fligl ance (ICR)m iomes in acid min drainage with lo hezsioedhig,esolnicnsndesnaselctnmber thehuan女edumet ce of laboratories in the 1990s and into the 2000s,the orbitrap -35.permafrost 48 and native prairie soill high-o capabilities more affordable for equencing or MS logies,or 83 can rates in both the linear ion trap (low resolution)and orbitrap microbial community phenotypes,or phenomes (high resolution).Aplot of the maximum resolution and MS There are still many ch enges that ne eed tobe addressed in rapping been ontimized to obtain the best no ssible biomolecule coverage mmunity analyses are the bioinformatics and computational and accuracy in each measurement.However,owing to the bottlenecks.Examples of these include building gene catalogs complexity of microbiome samples,additi nal technologies such to ameliorate reads and peptide assignments,biome-specific as one a tw onal liquid chromatography sed to increase the number of p lexity and analysis timeline identified.These separation technologies reduce the complexity of the sample before detection,allowing less suppression in the ion ng to Microbiomes of increasing complexity:limitationsand uture directi acid and MS technologi over the last decade have made it possible to analyze microbome gueAproieineineiheapeioficreioayope NATURE PROTOCOLS I VOL.11 NO.11 1 2016 I 2051
NATURE PROTOCOLS | VOL.11 NO.11 | 2016 | 2051 PERSPECTIVE from a variety of increasingly complex ecosystems (Fig. 3). For example, in the early 2000s, Banfield and co-workers were the first to use a combination of sequencing and MS approaches to study microbiomes in acid mine drainage with low microbial diversity41. Subsequently, microbiomes of varying diversity and complexity, including leaf-cutter ant colonies42–44, the termite gut45, the human gut46, sediments26,47, ocean water samples25,31–35, permafrost soil27,28,48, and native prairie soil16 have been investigated using either advanced sequencing or MS technologies, or both48 (Fig. 3). As the technologies continue to improve, we expect information on new and already studied microbial communities to multiply, providing greater insight into microbial community phenotypes, or phenomes49. There are still many challenges that need to be addressed in order to gain a deeper understanding of the molecular functions of microbiomes5,49. One of the biggest obstacles in microbial community analyses are the bioinformatics and computational bottlenecks. Examples of these include building gene catalogs to ameliorate reads and peptide assignments, biome-specific compositions in microbiomes, it is also desirable to know which proteins (i.e., metaproteomics) and metabolites (i.e., metabolomics) are produced under specific conditions, and how perturbations impact microbial functions. The measurement of proteins and metabolites produced by microbiomes in different samples has mainly been achieved using MS. In MS analyses of microbiome samples, the biomolecules of interest are ionized in the source, separated according to their mass-to-charge ratio in the mass analyzer and finally detected, providing metaproteomic or metabolomic measurements with high sensitivity, resolution and throughput. The introduction of electrospray ionization (ESI) in 1984 greatly increased the utility of MS measurements for assessment of biomolecules36. One challenge with ESI is that it is performed at atmospheric pressure (760 torr), whereas mass analyzers normally operate in pressure regimes between 10-6 and 10-11 torr. This pressure difference accounts for more than nine orders of magnitude, and if the source and mass analyzer regions are not well coupled, huge ion losses occur36,37. Thus, interface designs using ion funnels and transmission quadrupole regions have been optimized in the last two decades to refocus the ions while continually decreasing the pressure38. The improvements in these areas resulted in a decrease in ion losses and higher sensitivity measurements38. However, analyzing biomolecules in complex microbiomes by MS is still very difficult owing to the high dynamic range and thousands of components present in a given sample. These challenges required the improvement of the mass analyzer with respect to resolution, accuracy and MS/MS speed. Quadrupole, ion trap, time-of-flight and ion cyclotron resonance (ICR) mass analyzers constituted the majority of mass analyzers used until the introduction of the orbitrap37 in 2005. Whereas ICR mass analyzers allowed high-resolution studies in a select number of laboratories in the 1990s and into the 2000s, the orbitrap technology made high-resolution capabilities more affordable for additional laboratories39,40. Advances in orbitrap technology over the past decade have included higher-resolution measurements and faster MS/MS scan rates in both the linear ion trap (low resolution) and orbitrap (high resolution). A plot of the maximum resolution and MS/ MS scan speeds for MS trapping instruments introduced over the last 15 years (Fig. 2), reveals that both of these features have been optimized to obtain the best possible biomolecule coverage and accuracy in each measurement. However, owing to the complexity of microbiome samples, additional technologies such as one-dimensional and two-dimensional liquid chromatography separations and gas-phase ion mobility spectrometry are also being used to increase the number of proteins and metabolites identified. These separation technologies reduce the complexity of the sample before detection, allowing less suppression in the ion trap and detector owing to the many molecules present in each microbiome sample, and enabling higher coverage of the proteins and metabolites in a given sample39,40. Microbiomes of increasing complexity: limitations and future directions Developments in nucleic acid sequencing and MS technologies over the last decade have made it possible to analyze microbiomes LCQ LTQ LTQ-FT LTQ-Orbitrap LTQ-FT Ultra Exactive Velos Orbitrap Velos Orbitrap Elite Q-Exactive Orbitrap Fusion Q-Exactive HF Orbitrap Fusion Lumos 0 4 8 12 16 20 24 2000 2002 2004 2006 2008 2010 2012 2014 2016 0 100 200 300 400 500 600 700 800 MS/MS (low resolution) Instrument resolution Instrument resolution (x1,000) Scan rate (spectra/s) Year introduced MS/MS (high resolution) Soil and sediment Permafrost Ocean Human gut Leaf cutter ant colony Termite hindgut Extreme environments Microbiome complexity and multi-omics analysis timeline 2000 2016 Figure 3 | Approximate timeline with examples of increasingly complex microbiomes analyzed by sequencing and/or other omics technologies. Figure 2 | The evolution of ion-trap-based MS instruments. Illustration of how instrument resolution and MS/MS scan rates have increased with time to provide users with better measurements. However, using the highest values for both resolution and MS/MS acquisition is not usually advised for best instrument performance. Maximum instrument resolution is shown as red bars (values according to left y axis), with MS/MS low-resolution and highresolution scan rates indicated by yellow and blue bars (values according to right y axis), respectively. The instrument and year it was introduced are shown on the x axis. npg © 2016 Nature America, Inc. All rights reserved
PERSPECTIVE omiotation plator omingnh 5. 6 of biomolecules from highly diverse and complex sample matrices. such as soil,sediments and the human gut;assembly of complete genomes rather er speed,throug 9. 7 ly o meeting the high computational RAMrequirements ford assembly of large metagenomes and metatranscriptomes 10. 1. throughput loping sta Is to data cl and m 28136.36519 ofabricated high-density research.Recently,the White House Office of Science and ing using even faster 034543 lations and inaddition to thedevelopment of bioinformatics programs tha obiom oma varicty do not cu ucher-resoutio o 19. separations of peptides and metabolites and new databases for ase molecules molecular assignm ents.These technological and computational mproven nts w be vital o the decade 1114622012 the influe olex inte of microbial communities on ecosystem sustainability and health 23. of the Such in-depth knowledge of microbial 6-74(200 ommunity phenomes e a better und anding of w perturbations s ab d 25 improved ccoystem sustainability and human health strategics 26 ACKNOWL thank N.Jo and C.Brisla assista er Hon 8,146414752014 et ol onolpil5 earch (Ge inD 208-212 5NE1.9.101 1023(2015) 0. AUTHOR CONTRIBUTIONS R.A.W.S.J.C.R.J.M.E.S.B.andJ.K.J.all contributed to this work and commented on the manuscript at all stages ING FINANCIAL INTERESTS The authors declare ompeting hinar 32. 38102008) 33. Repiatantpemtsioiitiomatioaisavalabteonineathtip:/wmnatue 59.326-269(209 1 975(2016 2. 3. Lamendella.F 23915002012. 36.Yamashita, 2052 VOL.11 NO.11|2016 NATURE PROTOCOLS
2052 | VOL.11 NO.11 | 2016 | NATURE PROTOCOLS PERSPECTIVE 5. Alivisatos, A.P. et al. Microbiome. A unified initiative to harness Earth’s microbiomes. Science 350, 507–508 (2015). 6. Lozupone, C.A., Stombaugh, J.I., Gordon, J.I., Jansson, J.K. & Knight, R. Diversity, stability and resilience of the human gut microbiota. Nature 489, 220–230 (2012). 7. Heather, J.M. & Chain, B. The sequence of sequencers: The history of sequencing DNA. Genomics 107, 1–8 (2016). 8. Sanger, F., Nicklen, S. & Coulson, A.R. DNA sequencing with chainterminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463–5467 (1977). 9. Fleischmann, R.D. et al. Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269, 496–512 (1995). 10. Blattner, F.R. et al. The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1462 (1997). 11. Loman, N.J. et al. Performance comparison of benchtop high-throughput sequencing platforms. Nat. Biotechnol. 30, 434–439 (2012). 12. Ronaghi, M., Uhlén, M & Nyrén, P. A sequencing method based on real-time pyrophosphate. Science 281, 363, 365 (1998). 13. Margulies, M. et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437, 376–380 (2005). 14. Bentley, D.R. et al. Accurate whole human genome sequencing using reversible terminator chemistry. Nature 456, 53–59 (2008). 15. Voskoboynik, A. et al. The genome sequence of the colonial chordate, Botryllus schlosseri. eLife 2, e00569 (2013). 16. White, R.A., III et al. Moleculo long-read sequencing facilitates assembly and genomic binning from complex soil metagenomes. mSystems 1, e00045-16 (2016). 17. Sharon, I. et al. Accurate, multi-kb reads resolve complex populations and detect rare microorganisms. Genome Res. 25, 534–543 (2015). 18. Kuleshov, V. et al. Synthetic long-read sequencing reveals intraspecies diversity in the human microbiome. Nat. Biotechnol. 34, 64–69 (2016). 19. Levene, M.J. et al. Zero-mode waveguides for single-molecule analysis at high concentrations. Science 299, 682–686 (2003). 20. Eid, J. et al. Real-time DNA sequencing from single polymerase molecules. Science 323, 133–138 (2009). 21. Murray, I.A. et al. The methylomes of six bacteria. Nucleic Acids Res. 40, 11450–11462 (2012). 22. Laver, T. et al. Assessing the performance of the Oxford Nanopore Technologies MinION. Biomol. Detect. Quantif. 3, 1–8 (2015). 23. Venter, J.C. et al. Environmental genome shotgun sequencing of the Sargasso Sea. Science 304, 66–74 (2004). 24. Willing BP, et al. A pyrosequencing study in twins shows that gastrointestinal microbial profiles vary with inflammatory bowel disease phenotypes. Gastroenterology 139, 1844–1854.e1 (2010). 25. Mason, O.U. et al. Metagenome, metatranscriptome and single-cell sequencing reveal microbial response to Deepwater Horizon oil spill. ISME J. 6, 1715–1727 (2012). 26. Mason, O.U. et al. Metagenomics reveals sediment microbial community response to Deepwater Horizon oil spill. ISME J. 8, 1464–1475 (2014). 27. Mackelprang, R. et al. Metagenomic analysis of a permafrost microbial community reveals a rapid response to thaw. Nature 480, 368–371 (2011). 28. Hultman, J. et al. Multi-omics of permafrost, active layer and thermokarst bog soil microbiomes. Nature 521, 208–212 (2015). 29. Goltsman, D.S.A., Comolli, L.R., Thomas, B.C. & Banfield, J.F. Community transcriptomics reveals unexpected high microbial diversity in acidophilic biofilm communities. ISME J. 9, 1014–1023 (2015). 30. Zoetendal, E.G. et al. The human small intestinal microbiota is driven by rapid uptake and conversion of simple carbohydrates. ISME J. 6, 1415–1426 (2012). 31. Shi, Y., Tyson, G.W., Eppley, J.M. & DeLong, E.F. Integrated metatranscriptomic and metagenomic analyses of stratified microbial assemblages in the open ocean. ISME J. 5, 999–1013 (2011). 32. Frias-Lopez, J. et al. Microbial community gene expression in ocean surface waters. Proc. Natl. Acad. Sci. USA 105, 3805–3810 (2008). 33. Shi, Y., Tyson, G.W. & DeLong, E.F. Metatranscriptomics reveals unique microbial small RNAs in the ocean’s water column. Nature 459, 266–269 (2009). 34. Gilbert, J.A. et al. Detection of large numbers of novel sequences in the metatranscriptomes of complex marine microbial communities. PLoS One 3, e3042 (2008). 35. Gilbert, J.A. et al. Metagenomes and metatranscriptomes from the L4 longterm coastal monitoring station in the Western English Channel. Stand. Genomic Sci. 3, 183–193 (2010). 36. Yamashita, M. & Fenn, J.B. Negative ion source production with electrospray ion source. J. Phys. Chem. 88, 4671–4675 (1984). 37. Hu, Q., Noll, R.J., Li, H., Makarov, A., Hardman, M. & Graham Cooks R. The Orbitrap: a new mass spectrometer. J. Mass Spectrom. 40, 430–443 (2005). annotation platforms for improving interpretation and multiomics integration algorithms. Several other challenges that need to be addressed for better microbiome analyses include: extraction of biomolecules from highly diverse and complex sample matrices, such as soil, sediments and the human gut; assembly of complete genomes rather than sequence fragments directly from complex ecosystems; higher speed, throughput and dynamic range of MS technologies for metaproteomic and metabolomic measurements; meeting the high computational RAM requirements for de novo assembly of large metagenomes and metatranscriptomes16; sufficient storage and analysis options for terabytes to petabytes of data; and developing statistical and mathematical models to integrate the data and provide meaningful biological insights. Challenges aside, the future looks very bright for microbiome research. Recently, the White House Office of Science and Technology Policy (OSTP) announced a National Microbiome Initiative with funding from several federal agencies, industries and foundations focusing on improved technologies for understanding microbiomes5,45. We believe this initiative will prompt even faster and more informative nucleic acid sequencing and MS analyses, in addition to the development of bioinformatics programs that can quickly analyze microbiomes from a variety of ecosystems. We also expect the initiative will enable new technologies that do not currently exist, such as higher-resolution ion-mobility separations of peptides and metabolites and new databases for molecular assignments. These technological and computational improvements will be vital over the next decade for deciphering the roles of microbes in their natural habitats and determining the influence of the complex interplay between members of microbial communities on ecosystem sustainability and health. Such in-depth knowledge of microbial community phenomes will facilitate a better understanding of how perturbations such as climate change and disease affect microbiome functions, enabling better predictions of the impacts of these changes and facilitating improved ecosystem sustainability and human health strategies. ACKNOWLEDGMENTS We thank N. Johnson and C. Brislawn for their assistance in preparing the figures. This research was supported by the Pan-omics Program that is funded by the US Department of Energy’s Office of Biological and Environmental Research (Genomic Science Program) and the Microbiomes in Transition (MinT) Laboratory Directed Research and Development Initiative at the Pacific Northwest National Laboratory. Pacific Northwest National Laboratory is a multi-program national laboratory operated by Battelle for the Department of Energy under contract DE-AC06-76RL01830. AUTHOR CONTRIBUTIONS R.A.W., S.J.C., R.J.M., E.S.B. and J.K.J. all contributed to this work and commented on the manuscript at all stages. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Reprints and permissions information is available online at http://www.nature. com/reprints/index.html. 1. Locey, K.J. & Lennon, J.T. Scaling laws predict global microbial diversity. Proc. Natl. Acad. Sci. USA 113, 5970–5975 (2016). 2. Jansson, J.K. Towards “Tera-Terra”: Terabase sequencing of terrestrial metagenomes. eScholarship http://escholarship.org/uc/item/04p1x29k. 2011. 3. Lamendella, R., VerBerkmoes, N. & Jansson, J.K. ‘Omics’ of the mammalian gut—new insights into function. Curr. Opin. Biotechnol. 23, 491–500 (2012). 4. Gilbert, J.A., Jansson, J.K. & Knight, R. The Earth Microbiome project: successes and aspirations. BMC Biol. 12, 69 (2014). npg © 2016 Nature America, Inc. All rights reserved
PERSPECTIVE e0134752(2015) 39. 21) .toL.Adandngtehighhroughptieninationoftie 389-39 (2014 .Mol 45. 308,191 920(2005 g3tcusgn9yoponuiptrhetnge 6L79.3770-377812013 48.Wilm f plant biomas 49- 5016092o16 for m 1) EDITOR SUMMARY lanet lar n and collea eview the development of microbiome the past decade.and comment on the future potential of this fast-moving field 台 NATURE PROTOCOLS VOL.11 NO.11 2016 2053
NATURE PROTOCOLS | VOL.11 NO.11 | 2016 | 2053 PERSPECTIVE e0134752 (2015). 44. Huang, E.L. et al. The fungus gardens of leaf-cutter ants undergo a distinct physiological transition during biomass degradation. Environ. Microbiol. Rep. 6, 389–395 (2014). 45. Burnum, K.E. et al. Proteome insights into the symbiotic relationship between a captive colony of Nasutitermes corniger and its hindgut microbiome. ISME J. 5, 161–164 (2011). 46. Verberkmoes, N.C. et al. Shotgun metaproteomics of the human distal gut microbiota. ISME J. 3, 179–189 (2009). 47. Handley, K.M. et al. Biostimulation induces syntrophic interactions that impact C, S and N cycling in a sediment microbial community. ISME J. 7, 800–816 (2013). 48. Wilmes, P., Heintz-Buschart, A. & Bond, P.L. A decade of metaproteomics: where we stand and what the future holds. Proteomics 15, 3409–3417 (2015). 49. Jansson, J.K. & Baker, E.S. A multi-omic future for microbiome studies. Nat. Microbiol. 1, 16049 (2016). 38. Kelly, R.T., Tolmachev, A.V., Page, J.S., Tang, K. & Smith, R.D. The ion funnel: theory, implementations, and applications. Mass Spectrom. Rev. 29, 294–312 (2010). 39. Baker, E.S. et al. Advancing the high throughput identification of liver fibrosis protein signatures using multiplexed ion mobility spectrometry. Mol. Cell. Proteomics 13, 1119–1127 (2014). 40. Baker, E.S. et al. Enhancing bottom-up and top-down proteomic measurements with ion mobility separations. Proteomics 15, 2766–2776 (2015). 41. Ram, R.J. et al. Community proteomics of a natural microbial biofilm. Science 308, 1915–1920 (2005). 42. Aylward, F.O. et al. Leucoagaricus gongylophorus produces diverse enzymes for the degradation of recalcitrant plant polymers in leaf-cutter ant fungus gardens. Appl. Environ. Microbiol. 79, 3770–3778 (2013). 43. Aylward, F.O. et al. Enrichment and broad representation of plant biomassdegrading enzymes in the specialized hyphal swellings of Leucoagaricus gongylophorus, the fungal symbiont of leaf-cutter ants. PLoS One 10, EDITOR SUMMARY In this Perspective, Janet Jansson and colleagues review the development of microbiome analysis technologies over the past decade, and comment on the future potential of this fast-moving field. npg © 2016 Nature America, Inc. All rights reserved
