A "Cul­tural" Re­nais­sance

Ge­nomics Breathes New Life into an Old Craft

This ar­ti­cle by guest au­thor Paul Carini was first pub­lished in mSys­tems (CC BY 4.0). We reprint it with the author's gen­er­ously granted per­mis­sion.

by Paul Carini

   Sooner or later, every­thing old is new again.
    —Stephen King (29)

For over a cen­tury, our un­der­stand­ing of mi­cro­bial bi­ol­ogy has been pred­i­cated on the abil­ity of sci­en­tists to cul­ti­vate and study or­gan­isms un­der con­trolled lab­o­ra­tory con­di­tions. In­deed, un­til the 1980s, our knowl­edge of mi­cro­bial di­ver­sity was con­strained to those mi­crobes that grew in the lab­o­ra­tory or ap­peared un­der the mi­cro­scope. Yet, it was also un­der­stood that the ma­jor­ity of mi­crobes were re­cal­ci­trant to growth in the lab­o­ra­tory. The in­con­gru­ence be­tween enu­mer­a­tion by di­rect mi­cro­scopic count­ing and cul­tur­able count­ing was elo­quently pre­sented by Sta­ley and Konopka as "the great plate count anom­aly" (1) and is oft at­trib­uted to the wide­spread no­tion that more than 99% of bac­te­ria are un­cul­tur­able. This anom­aly was con­firmed when Pace and col­lea­gues started in­ves­ti­gat­ing nat­ural en­vi­ron­ments with mol­e­c­u­lar tools tar­get­ing rRNA genes, lift­ing the veil mask­ing vast un­cul­tured mi­cro­bially di­verse pop­u­la­tions for the first time (2). Per­haps in­advertently, the co­in­ci­dent tim­ing of the great plate count anom­aly and Pace et al.'s early rRNA-based ap­proaches be­came the im­pe­tus for a ma­jor shift in the ways that mi­cro­bial ecol­o­gists ex­plored nat­ural mi­cro­bial as­sem­blages; a whole gen­er­a­tion of re­searchers traded piles of agar plates and growth curves for DNA se­quencers, com­puter servers, and meta­bolic re­con­struc­tions from ge­nomic data. To­day, DNA se­quenc­ing is per­ceived (or at least touted) as the way to circum­vent the prob­lem of mi­cro­bial un­cul­tur­a­bil­ity.

A "Cul­tural" Re­nais­sance

Af­ter a decade of re­fin­ing the "big-data" side of mi­cro­bi­ol­ogy, we are in the midst of a cul­tural re­naissance in which the im­por­tance of cul­ti­va­tion-fo­cused ef­forts has been re­dis­cov­ered, with ex­citing re­sults and im­pli­ca­tions. Sev­eral groups have demon­strated that a high per­cent­age of host-as­so­ci­ated mi­cro­biome mem­bers can be cul­ti­vated (3–7) and that such cul­ture col­lec­tions are in­dispensable for de­duc­ing im­por­tant as­pects of mi­cro­biome func­tion (8–10). The re­newed moti­va­tion to cul­ture cells from non­host en­vi­ron­ments, such as soil (11), aquatic en­vi­ron­ments (12), and the deep bios­phere (13), is in­spired, in part, by these suc­cesses and the enor­mous amounts of mo­lecular data that have il­lu­mi­nated seem­ingly lim­it­less mi­cro­bially di­verse pop­u­la­tions that we know very lit­tle about (14). We can now dis­sect life's blue­prints for these un­cul­ti­vated lin­eages and use this in­for­ma­tion to fa­cil­i­tate cul­ti­va­tion, in­sight not af­forded to pre­vi­ous gen­er­a­tions of resear­chers.

Trans­lat­ing ge­nomic in­for­ma­tion into an un­der­stand­ing of or­gan­is­mal phys­i­ol­ogy is any­thing but straight­for­ward, and cul­tur­ing en­vi­ron­men­tal mi­crobes is very timein­ten­sive. This is high­lighted by the story of the abun­dant ma­rine het­erotroph Pelag­ibac­ter (SAR11). Pelag­ibac­ter was dis­cov­ered with mol­e­c­u­lar meth­ods in 1990 (15) and cul­tured on nat­ural sea­wa­ter medium 12 years later (16). Sur­pris­ingly, it was not im­me­di­ately ob­vi­ous from Pelag­ibac­ter genome se­quences (17) what nu­trients were re­quired for growth. In fact, with genomes in hand, it took an ad­di­tional 8 years plus sev­eral post­docs and grad­u­ate stu­dents to trans­late that in­for­ma­tion into a de­fined medium (re­ference 18 and ref­er­ences therein). Sub­se­quent lab­o­ra­tory stud­ies have elu­ci­dated the un­usual bi­ol­ogy that con­tributes to Pelagibacter's un­par­al­leled suc­cess in the sea and its key role in Earth's bio­geo­chem­istry (re­viewed in ref­er­ence 19). The unique char­ac­ter­is­tics of Pelag­ibac­ter cells were dis­cov­ered by care­ful test­ing in the lab­o­ra­tory af­ter scru­ti­niz­ing of genomes, tran­scrip­tomes, and pro­teomes, a back-and-forth process that re­quires a cul­ture, a genome, and pa­tience.

Ta­ble: Ex­am­ples of mic­robes that are most want­ed in cul­ture. Source. Fron­tispiece: de­tail from Source

Pelagibacter's nar­ra­tive high­lights an im­por­tant les­son. By shy­ing away from the chal­lenge of stu­dying mi­cro­bial cul­tures, we ef­fec­tively turn our backs on an en­tire world of emer­gent prop­er­ties that gov­ern mi­cro­bial ac­tiv­ity and ecosys­tem func­tion, prop­er­ties that are not al­ways pre­dic­table a pri­ori from se­quence in­for­ma­tion. To analo­gi­ze, one can­not un­der­stand the ex­pe­ri­ence of dri­ving a Fer­rari from the list of its com­po­nents; a parts list does not con­vey the han­dling, the sound, or the driver's con­nection with the ma­chine. Be­cause of the in­vest­ment in se­quenc­ing tech­nol­ogy, our ge­netic in­ven­to­ries are more ex­ten­sive than ever, yet the cul­ti­va­tion of novel mi­crobes re­mains a com­plex task, much like as­sem­bling a 3‑di­men­­sional puz­zle. Hy­pothe­ses gen­er­ated from genome-based meta­bolic re­con­struc­tions from sin­gle cells or metagenomes pro­vide a cru­cial di­mension to guide the as­sem­bly of the puz­zle, but genomes are a parts list; we seek an under­stan­d­ing of the emer­gent prin­ci­ples of the cells them­selves and the com­mu­ni­ties that they con­sti­tute.

While this cul­tural re­nais­sance is in­dica­tive of a greater aware­ness for the need to study cell cul­tures to com­pre­hen­sively un­der­stand Earth's mi­cro­biome, it is not al­ways clear how to lever­age and in­te­grate mol­e­c­u­lar data to do that. Be­low are three steps that will lay a foun­da­tion for the fu­ture of in­te­grated Earth mi­cro­biome re­search with an em­pha­sis on elu­ci­dat­ing the un­cul­tured mi­cro­bially di­verse pop­u­la­tions that we un­der­stand the least.

De­cide What to Cul­ture

Let us be clear; from a sheer num­bers per­spec­tive, it is im­pos­si­ble to cul­ture every mi­crobe on Earth no mat­ter what the en­vi­ron­ment. Be­cause of this, we should set cul­ti­va­tion goals that are spe­cific, achiev­able, and rel­e­vant. I en­vi­sion ex­plicit most­wanted lists that use mol­e­c­u­lar data to in­form ex­actly which taxa we should tar­get for cul­ti­va­tion and why they are im­por­tant. Fur­ther, this mol­e­c­u­lar data can be used to iden­tify po­ten­tial cul­ti­va­tion strate­gies through meta­bolic re­construction. The ra­tio­nales for cul­ti­va­tion of a par­tic­u­lar taxa will vary across re­search groups but may in­clude (i) an organism's high rel­a­tive abun­dance, (ii) its key role in bio­geo­chem­istry or biore­mediation, (iii) its po­ten­tial to pro­duce nat­ural prod­ucts, and (iv) its sub­stan­tial di­ver­gence from cul­tured taxa. Ta­ble 1 con­tains a short list of po­ten­tial most-wanted mi­crobes from var­i­ous eco­sys­tems that would be valu­able to ob­tain in cul­ture, based on an in­for­mal sur­vey of col­leagues and my own in­ter­ests.

In­vest in Risky Cul­ture-Based Work

A pop­u­lar per­cep­tion is that ex­ploratory cul­ture-based re­search is un­usu­ally risky and un­likely to re­sult in new iso­lates of in­ter­est. First, this is false (4,11,12,20,21). Sec­ond, all sci­ence is risky, and much is un­suc­cess­ful; how much do we want to learn? While it is tempt­ing to point the fin­ger at fund­ing agen­cies for not sup­port­ing ex­ploratory cul­ti­va­tion work, it is our col­leagues, the review­ers and pan­elists, that mar­gin­al­ize these pro­pos­als. Fund­ing agen­cies could shift this dis­course by so­lic­it­ing pro­pos­als that ex­plic­itly aim to (i) co­or­di­nate ex­ploratory cul­ti­va­tion ex­per­i­ments with cre­ative ge­nomics or metabolomics ap­proaches, (ii) de­velop new or higher-through­put cul­ti­va­tion strate­gies, and (iii) sup­port long-term projects that in­ves­ti­gate the bi­ol­ogy of slow-grow­ing or non­canonical model or­gan­isms that do not con­form to the scale of tra­di­tional fund­ing cy­cles. As J. Ca­meron Thrash ex­plores in a com­pan­ion Per­spec­tive in this is­sue, we need to con­strain the costs of in­vest­ing in cul­ti­va­tion-based work. If mi­crobes hold the an­swers to many solv­able prob­lems, at some point we will need to in­vest in cul­tur­ing them. Long-term, I en­vi­sion well-funded high-through­put cul­ti­va­tion core labs that cul­ture bac­te­ria from user sam­ples. Not only would these cen­ters pre­serve liv­ing bio­di­ver­sity that will al­most cer­tainly have util­ity in a chang­ing world, but the cul­ti­va­tion of enig­matic taxa might be a strict num­bers game: the more ex­per­i­ments that are con­ducted, the more likely they will cap­ture some­thing novel.

Build Knowl­edge Bridges with Cul­tures

The abil­ity to se­quence DNA is no longer a lim­it­ing fac­tor in un­der­stand­ing mi­cro­bial com­mu­ni­ties; the bot­tle­neck lies with trans­lat­ing these se­quence data into a func­tional con­text. The root of this lim­i­ta­tion is that most genes have poor or no func­tional an­no­ta­tion, and our in­ter­pre­ta­tion of the re­main­der is bi­ased by the phys­i­ol­ogy of a few model or­gan­isms, the so-called street­light ef­fect (22). Iso­lat­ing and study­ing or­gan­isms with un­usual bi­ol­ogy can bridge this knowl­edge gap. For ex­ample, fo­cused func­tional ge­nomics ap­proaches un­cov­ered wide­spread lipid re­mod­el­ing in mari­ne het­erotrophs and re­lated this process to poorly an­no­tated genes (23–25). Like­wise, the geno­me-fa­cil­i­tated dis­cov­ery of com­plete am­mo­nia ox­i­da­tion to ni­trate by Ni­tro­spira cul­tures funda­men­tally changed our un­der­stand­ing of the ni­tro­gen cy­cle (26). More­over, high-through­put func­tional ge­nomics ap­proaches, such as trans­po­son in­ser­tion se­quenc­ing (TnSeq), have proven to be scal­able and pow­er­ful for iden­ti­fy­ing the fit­ness land­scape of poorly an­no­tated genes (27). While these ap­proaches cer­tainly re­quire more ef­fort than se­quenc­ing alone, the in­for­ma­tion that they pro­vide in many cases un­am­bigu­ously links geno­type to phe­no­type.

Olig­otro­phy as an Emer­gent Prin­ci­ple

The chal­lenge of trans­lat­ing ge­nomic data into mi­cro­bial cul­tures and link­ing genes with func­tion is a dri­ving force for our re­search group. How­ever, an im­por­tant di­men­sion to the prob­lem of iso­lating un­cul­tured mi­crobes may be an emer­gent prin­ci­ple not eas­ily de­duced from genomes: oli­go­trophy. Oli­go­trophy de­scribes the para­dox­i­cal and enig­matic phe­nom­e­non of mi­cro­bial cells grow­ing op­ti­mally when nu­tri­ent avail­abil­ity is low. New ev­i­dence sug­gests that the nu­mer­i­cally abun­dant mi­crobes in non­host sys­tems are olig­otrophs (28). Yet, we have a mas­sively in­com­plete un­der­stand­ing of the phys­i­o­log­i­cal ba­sis for olig­otro­phy and how olig­otrophs con­tribute to micro­bial com­mu­nity sta­bil­ity and ecosys­tem func­tion.

We have built our lab around the idea that many un­cul­tured soil mi­cro­bial lin­eages are oligo­trophs. In gen­eral, olig­otrophs are chal­leng­ing to cul­ture be­cause they are small, slow-grow­ing cells and do not at­tain high yields on low nu­tri­ent me­dia, ren­der­ing com­mon meth­ods of quantify­ing mi­cro­bial growth in­ef­fec­tive. To cir­cum­vent this, we in­vested in equip­ment to sep­a­rate cells from en­vi­ron­men­tal ma­tri­ces and count olig­otrophic cul­tures with high through­put and sen­si­tiv­ity. Our ini­tial cul­ti­va­tion re­sults are en­cour­ag­ing; across sev­eral ex­per­i­ments, we have ~3,000 cultur­es, some of which are rep­re­sen­ta­tives of un­cul­ti­vated lin­eages of im­por­tant soil bac­te­ria. In these ex­per­i­ments, the con­cen­tra­tion of a de­fined set of nu­tri­ents sig­nif­i­cantly in­flu­enced what grew, and many strains iso­lated on low-nu­tri­ent medium ap­pear to be ob­lig­ate olig­otrophs that are in­hi­bited by mod­est con­cen­tra­tion in­creases of the same nu­tri­ents. Our long-term goal is to in­crease the ef­fi­ciency and re­duce the per­son-hour cost of cul­ti­va­tion by au­tomat­ing cer­tain as­pects of the cul­ti­va­tion process so that we can spend more time in­ves­ti­gat­ing in­ter­est­ing iso­lates in-depth.

My hopes are not only that the cul­tural re­nais­sance is per­ceived as a ref­er­ence to a re­newed in­ter­est in cul­tur­ing cells but also that it in­vokes a cul­tural shift in the way mi­cro­bi­ol­o­gists per­ceive cells and their role in mod­ern mi­cro­bi­ol­ogy. Cells are not sim­ply "bags of bio­chem­istry" but liv­ing enti­ties that are seam­lessly in­te­grated into the phys­i­cal, eco­log­i­cal, and evo­lu­tion­ary land­scape of Earth. Be­cause of this, the meth­ods by which we choose to study them must be sim­i­larly inte­gra­ted.
 

Ac­knowl­edge­ments

I ap­pre­ci­ate crit­i­cal feed­back on this Per­spec­tive from Ryan Bartelme, Tess Brewer, Amanda Howe, and J. Cameron Thrash.

Ref­er­ences

1. Sta­ley JT, Konopka A. 1985. Mea­sure­ment of in situ ac­tiv­i­ties of non­pho­to­syn­thetic micro­or­gan­isms in aqua­tic and ter­res­trial habi­tats. Annu Rev Mi­cro­biol 39:321–346 [doi]
2. Pace NR, Stahl DA, Lane DJ, Olsen GJ. 1985. An­a­lyz­ing nat­ural mi­cro­bial pop­u­la­tions by rRNA se­quences. ASM News 51:4 –12
3. Bai Y, Müller DB, Srini­vas G, Gar­rido-Oter R, Pot­thoff E, Rott M, Dom­browski N, Münch PC, Spae­pen S, Re­­mus-Em­ser­mann M, Hüt­tel B, McHardy AC, Vorholt JA, Schulze-Lefert P. 2015. Func­tional over­lap of the Ara­bidopsis leaf and root mi­cro­biota. Na­ture 528:364 –369 [doi]
4. Browne HP, Forster SC, Anonye BO, Ku­mar N, Neville BA, Stares MD, Gould­ing D, Law­ley TD. 2016. Cultur­ing of "un­cul­tur­able" hu­man mi­cro­biota re­veals novel taxa and ex­ten­sive sporu­la­tion. Na­ture 533:543–546 [doi]
5. Forster SC, Ku­mar N, Anonye BO, Almeida A, Vi­ciani E, Stares MD, Dunn M, Mkan­dawire TT, Zhu A, Shao Y, Pike LJ, Louie T, Browne HP, Mitchell AL, Neville BA, Finn RD, Law­ley TD. 2019. A hu­man gut bac­te­r­ial genome and cul­ture col­lec­tion for im­proved metage­nomic analy­ses. Nat Biotech­nol 37:1–12 [doi]
6. Nunes da Rocha U, Cadillo-Quiroz H, Karaoz U, Ra­jeev L, Kl­it­gord N, Dunn S, Truong V, Buenros­tro M, Bo­wen BP, Gar­cia-Pichel F, Mukhopad­hyay A, Northen TR, Brodie EL. 2015. Iso­la­tion of a sig­nif­i­cant frac­tion of non­pho­totroph di­ver­sity from a desert Bi­o­log­i­cal Soil Crust. Front Mi­cro­biol 6:1–15 [doi]
7. Lagk­ou­var­dos I, Pukall R, Abt B, Foe­sel BU, Meier-Kolthoff JP, Ku­mar N, Bres­ciani A, Martínez I, Just S, Zieg­ler C, Bru­giroux S, Garzetti D, Wen­ning M, Bui TPN, Wang J, Hugen­holtz F, Plugge CM, Pe­ter­son DA, Hornef MW, Baines JF, Smidt H, Wal­ter J, Kris­tiansen K, Nielsen HB, Haller D, Over­mann J, Stecher B, Clavel T. 2016. The Mouse In­testi­nal Bac­te­r­ial Col­lec­tion (miBC) pro­vides host-spe­cific in­sight into cul­tured di­ver­sity and func­tional po­ten­tial of the gut mi­cro­biota. Nat Mi­cro­biol 1:16131 [doi]
8. Swen­son TL, Karaoz U, Swen­son JM, Bowen BP, Northen TR. 2017. Link­ing soil bi­ol­ogy and che­mis­try in bio­logical soil crust us­ing iso­late ex­ometabolomics. Nat Com­mun 9:19 [doi]
9. Zhal­nina K, Louie KB, Hao Z, Man­soori N, da Rocha UN, Shi S, Cho H, Karaoz U, Lo­qué D, Bowen BP, Fire­sto­ne MK, Northen TR, Brodie EL. 2018. Dy­namic root ex­u­date chem­istry and mi­cro­bial sub­strate pref­er­ences drive pat­terns in rhi­zos­phere mi­cro­bial com­mu­nity as­sem­bly. Nat Mi­cro­biol 3:470–480 [doi]
10. Cas­trillo G, Teix­eira P, Pare­des SH, Law TF, de Lorenzo L, Feltcher ME, Finkel OM, Break­field NW, Miecz­kow­ski P, Jones CD, Paz-Ares J, Dangl JL. 2017. Root mi­cro­biota drive di­rect in­te­gra­tion of phos­phate stress and im­mu­nity. Na­ture 543:513–518 [doi]
11. Pold G, Billings AF, Blan­chard JL, Burkhardt DB, Frey SD, Melillo JM, Schn­abel J, van Diepen LTA, DeAn­ge­lis KM. 2016. Long-term warm­ing al­ters car­bo­hy­drate degra­da­tion po­ten­tial in tem­per­ate for­est soils. Appl En­viron Mi­cro­biol 82:6518–6530 [doi]
12. Hen­son MW, Pitre DM, Weck­horst JL, Lan­c­los VC, Web­ber AT, Thrash JC. 2016. Ar­ti­fi­cial seawa­ter me­dia fa­cil­i­tate cul­ti­vat­ing mem­bers of the mi­cro­bial ma­jor­ity from the Gulf of Mex­ico. mSphere 1:e00028-16 [doi]
13. Parkes RJ, Sellek G, Web­ster G, Mar­tin D, An­ders E, Weight­man AJ, Sass H. 2009. Cul­tur­able pro­karyotic di­ver­sity of deep, gas hy­drate sed­i­ments: first use of a con­tin­u­ous high-pres­sure, an­aero­b­ic, en­rich­ment and iso­la­tion sys­tem for sub­seafloor sed­i­ments (Deep­IsoBUG). En­v­i­ron Mi­cro­biol 11:3140 –3153 [doi]
14. Lloyd KG, Steen AD, Ladau J, Yin J, Crosby L. 2018. Phy­lo­ge­net­i­cally novel un­cul­tured mi­cro­bial cells domi­nate Earth mi­cro­bio­mes. mSys­tems 3:e00055-18 [doi]
15. Gio­van­noni SJ, Britschgi TB, Moyer CL, Field KG. 1990. Ge­netic di­ver­sity in Sar­gasso Sea bacte­rio­plankton. Na­ture 345:60–63 [doi]
16. Rappé MS, Con­non SA, Ver­gin KL, Gio­van­noni SJ. 2002. Cul­ti­va­tion of the ubiq­ui­tous SAR11 ma­rine bacte­rioplankton clade. Na­ture 418:630–633 [doi]
17. Gio­van­noni SJ, Tripp HJ, Gi­van S, Po­dar M, Ver­gin KL, Bap­tista D, Bibbs L, Eads J, Richard­son TH, Noorde­wier M, Rappé MS, Short JM, Car­ring­ton JC, Mathur EJ. 2005. Genome stream­lin­ing in a cos­mopoli­tan oceanic bac­terium. Sci­ence 309:1242–1245 [doi]
18. Carini P, Steindler L, Besz­teri S, Gio­van­noni SJ. 2013. Nu­tri­ent re­quire­ments for growth of the ex­treme olig­otroph "Can­di­da­tus Pelag­ibac­ter ubique" HTCC1062 on a de­fined medium. ISME J 7:592– 602 [doi]
19. Gio­van­noni SJ. 2017. SAR11 bac­te­ria: the most abun­dant plank­ton in the oceans. Annu Rev Mar Sci 9:231–255 [doi]
20. San­toro AE, Cas­ciotti KL. 2011. En­rich­ment and char­ac­ter­i­za­tion of am­mo­nia-ox­i­diz­ing ar­chaea from the open ocean: phy­logeny, phys­i­ol­ogy and sta­ble iso­tope frac­tion­a­tion. ISME J 5:1796 –1808 [doi]
21. Kim S, Kang I, Seo J‑H, Cho J‑C. 2018. Cul­tur­ing the ubiq­ui­tous fresh­wa­ter acti­nobac­te­r­ial acI lin­eage by sup­ply­ing a bio­chem­i­cal 'helper'catalase. bioRxiv [doi]
22. Wikipedia. 30 Sep­tem­ber 2018. Street­light ef­fect, on Wikipedia, The Free En­cy­clo­pe­dia. Ac­cessed 1 Feb­ru­ary 2019.
23. Carini P, Van Mooy BAS, Thrash JC, White A, Zhao Y, Camp­bell EO, Fredricks HF, Gio­van­noni SJ. 2015. SAR11 lipid ren­o­va­tion in re­sponse to phos­phate star­va­tion. Proc Natl Acad Sci U S A 112:7767–7772 [doi]
24. Se­bastián M, Smith AF, González JM, Fredricks HF, Van Mooy B, Koblížek M, Brandsma J, Koster G, Mestre M, Mosta­jir B, Pitta P, Pos­tle AD, Sánchez P, Gasol JM, Scan­lan DJ, Chen Y. 2016. Lipid re­modelling is a wide­spread strat­egy in ma­rine het­erotrophic bac­te­ria upon phos­pho­rus de­fi­ciency. ISME J 10:968 –978 [doi]
25. Smith AF, Ri­ht­man B, Stir­rup R, Sil­vano E, Mausz MA, Scan­lan DJ, Chen Y. 2019. Elu­ci­da­tion of glu­t­a­mine lipid biosyn­the­sis in ma­rine bac­te­ria re­veals its im­por­tance un­der phos­pho­rus de­plete growth in Rhodo­bac­teraceae. ISME J 13:39–49 [doi]
26. Daims H, Lebe­deva EV, Pje­vac P, Han P, Her­bold C, Al­bert­sen M, Jehm­lich N, Palatin­szky M, Vier­heilig J, Bu­laev A, Kirkegaard RH, Bergen von M, Rat­tei T, Bendinger B, Nielsen PH, Wag­ner M. 2015. Com­plete nitri­fi­ca­tion by Ni­tro­spira bac­te­ria. Na­ture 528:504 –509 [doi]
27. Price MN, Wet­more KM, Wa­ters RJ, Callaghan M, Ray J, Liu H, Kuehl JV, Mel­nyk RA, Lam­son JS, Suh Y, Carl­son HK, Es­quivel Z, Sadeeshku­mar H, Chakraborty R, Zane GM, Ru­bin BE, Wall JD, Visel A, Bris­tow J, Blow MJ, Arkin AP, Deutschbauer AM. 2018. Mu­tant phe­no­types for thou­sands of bac­terial genes of un­known func­tion. Na­ture 557:503–509 [doi]
28. Gao Y, Wu M. 2018. Free-liv­ing bac­te­r­ial com­mu­ni­ties are mostly dom­i­nated by olig­otrophs. bioRxiv [doi]
29. King S. 2005. The Col­orado kid. Dorch­ester Pub­lish­ing Co., New York, NY.

 

Paul Carini is As­sis­tant Pro­fes­sor for Mi­cro­bial Eco­phys­i­ol­ogy at the Uni­ver­sity of Ari­zona. He's on twit­ter: @Paul_Carini.