Fecal microbiota cooperative metabolism of pectins derived from apple pomace: A functional metagenomic study

LWT - Food Science and Technology 187 (2023) 115362 A ailable online 29 Sep embe 2023 0023-6438/© 2023 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by- nc-nd/4.0/). Fecal mic obio a coope a i e me abolism o pec ins de i ed om apple pomace: A unc ional me agenomic s udy Ines Cal e e-To e a , b , 1 , Ca los Saba e a , b , 1 , Abela do Ma golles a , b , Lo ena Ruiz a , b , * a Depa men o Mic obiology and Biochemis y o Dai y P oduc s, Ins i u o de P oduc os L´ ac eos de As u ias-Consejo Supe io de In es igaciones Cien í icas (IPLA-CSIC), Paseo Río Lina es s/n, 33300, Villa iciosa, As u ias, Spain b Func ionali y and Ecology o Bene icial Mic obes (Mic oHeal h) G oup, Ins i u o de In es igaci´ on Sani a ia del P incipado de As u ias (ISPA), O iedo, As u ias, Spain ARTICLE INFO Keywo ds: P ebio ics Pec in Mic obio a Apple pomace ABSTRACT Complex polysaccha ides such as pec in a e ecei ing inc easing a en ion as eme gen p ebio ics which could be o in e es o human heal h. Pec in comp ises a wide a ay o highly complex and di e se polysaccha ide s uc u es, which a e abundan in ag icul u al by-p oduc s. Pec in me aboliza ion by he gu mic obio a equi es coope a i e me abolism and c oss- eeding ne wo ks among mic obial species. He ein we ha e deepened in o he gu modula o y p ope ies o apple pomaces and pec ins de i ed om his by p oduc , h ough a sho gun me agenomics app oach coupled o ad anced unc ional anno a ion and p edic ion o coope a ing ne wo ks o me agenome-assembled genomes (MAGs). Fo his pu pose, we ha e used samples om ba ch in i o ecal e men a ions pe o med wi h samples om heal hy dono s and C ohn’s disease pa ien s, in he p esence o apple pomace and pec ins exhibi ing di e en s uc u al p ope ies. Ou esul s con i m he capaci y o apple pomace/ pec ins o p omo e key axa gene ally unde ep esen ed in C ohn’s disease pa ien s bu ha ha e been a ibu ed an i-in lamma o y p ope ies, Faecalibac e ium, Ruminococcaceae membe s and Akke mansia being e y ema kable. Besides, some Ruminococcus and Akke mansia species ha e been iden i ied as key symbion s, exhibi ing complemen a y me abolic ai s o ully me abolise pec ins om apple pomace. These esul s will aid de eloping mic obiome- a ge ed p ebio ic and/o synbio ic s a egies o speci ic popula ion g oups. 1. In oduc ion The p ebio ic concep has changed o e ime, in pa , because all he ad ances in mic obiome esea ch. The In e na ional Scien i ic Associa- ion o P obio ics and P ebio ics (ISAPP) p oposes he de ini ion “a subs a e ha is selec i ely u ilized by hos mic oo ganisms con e ing a heal h bene i ” (Gibson e al., 2017). Wi h his de ini ion he doo is opened o he sea ch o new chemical compounds and subs a es wi h p ebio ic capabili y. No ewo hy, ege able by-p oduc s ha e gene ally been ecognized a good sou ce o ing edien s wi h p ebio ic p ope ies (Albuque que e al., 2021). Hence, ege able was e could ha e g ea e economic and en i onmen al a ac i eness wi hin he bio e ine y concep due o i s unc ional and/o heal h bene i s. Apple pomace is one o he mos abundan by-p oduc s ob ained ac oss se e al coun ies and consis s on he solid le o e ob ained a e apple p ocessing du ing cide manu ac u ing. I is a good sou ce o pec in, and is ich in mine als, die a y ibe and polyphenols (An onic e al., 2020; Cal e e-To e e al., 2021; Skinne e al., 2018), some o which ha e been a ibu ed heal h p omo ing p ope ies. High le els o ibe in he die ha e been epo ed o p o ec agains he de elopmen o in lamma o y bowel diseases (IBD) and a long- e m die wi h die a y ibe lowe s he isk o C ohn’s disease (CD). This e ec migh be possibly linked o he sho chain a y acids (SCFAs) ha a e p oduced upon ibe e men a ion by he in es inal mic obio a, a well-known con ibu o o he in e play be ween die and heal h (Ishisono e al., 2019). Hence, unde s anding how he mic obio a coope a i ely me abolise p ebio ic ibe s is c i ical o de ine hei heal h e ec s. In i o ecal e men a ions ha e been conduc ed o in es iga e he p ebi- o ic po en ial o he pec in o apple pomaces (Cal e e-To e e al., 2021). The s udies wi h pec in, ha is a he e opolysaccha ide composed o pa ly me hyl-es e i ied galac u onic acid (GalA) and ami ied domains wi h neu al suga s, ha e demons a ed ha pec in has good * Co esponding au ho . Depa men o Mic obiology and Biochemis y o Dai y P oduc s, Ins i u o de P oduc os L´ ac eos de As u ias-Consejo Supe io de In es igaciones Cien í icas (IPLA-CSIC), Paseo Río Lina es s/n, 33300, Villa iciosa, As u ias, Spain. E-mail add ess: [email p o ec ed] (L. Ruiz). 1 These au ho s ha e equally con ibu ed o his wo k. Con en s lis s a ailable a ScienceDi ec LWT jou nal homepage: www.else ie .com/loca e/lw h ps://doi.o g/10.1016/j.lw .2023.115362 Recei ed 31 July 2023; Recei ed in e ised o m 26 Sep embe 2023; Accep ed 28 Sep embe 2023 LWT 187 (2023) 115362 2 e men a ion cha ac e is ics and ca dio ascula p o ec ion ac i i y, as well as an i-in lamma o y and an ioxidan p ope ies (Cal e e-To e e al., 2021; Gull´ on e al., 2013; Ribei o e al., 2022). Ne e heless, pec in e men a ion abili y is no widely p esen in all mic obial axa ound in he in es inal mic obio a (Saba e , Cal e e-To e, e al., 2021), hence hei e men a ion pa e n in a ge popula ions migh be s ongly condi ioned by he me abolic capabili ies o he axa p esen in hei gu mic obio a, as well as by hei capaci y o es ablish c oss- eeding e- la ionships. Unde s anding he me abolic ai s and ne wo ks equi ed in a gu mic obio a communi y o achie ing e icien e men a ion o pec in- ich ma e ials, and p omo ion o heal hy communi ies, is essen- ial o a ionally design pec in-based p ebio ic ing edien s. La es ad ances in nex -gene a ion sequencing echniques like me - agenomics in combina ion wi h ad anced bioin o ma ic me hods ha e e ealed use ul o unde s and he e ec o die a y ing edien s on he in es inal mic obio a, and ha e been used o in es iga e he ibe me abolism o heal h-associa ed mic obial gene a (Saba e , Cal e e-To e, e al., 2021). These me hodologies could be o u mos in e es o ailo nex gene a ion p obio ics and p ebio ics ha could a ge speci ic disease condi ions such as CD (Lin e al., 2019; Lo dan e al., 2020). In his ega d, sho gun me agenomics allows he iden i i- ca ion o he mic oo ganisms a species and s ain le els and he e- co e y o whole me agenomic assembled genomes (MAG) om complex mic obial communi ies o complex unc ional analysis o hei coop- e a i e me abolism (Culp & Goodman, 2023; F ioux e al., 2020; Tomioka e al., 2022). Howe e unc ional analysis o MAGs om mic obiome samples ha e been sca cely used o he pu pose o p e- bio ics cha ac e iza ion and design. To his end, ad anced me abolic modelling so wa e ools ha e been de eloped o cha ac e ise he me abolic po en ial and o simula e syn- e gis ic in e ac ions be ween MAGs (Belcou e al., 2020; Saba e e al., 2022). Me abolic coope a ion o gu mic obes, o ins ance h ough es ablishmen o c oss- eeding ne wo ks, educes compe i ion be ween coexis ing axa enhancing hei g ow h. These mechanisms can be elucida ed in silico o de e mine he minimal se o key mic oo ganisms equi ed in a communi y o coope a i ely exe a gi en unc ion, iden- i ying essen ial and al e na i e symbion s. In his sense, essen ial symbion s comp ise key mic oo ganisms ha a e equi ed in e e y minimal communi y o mic oo ganisms o conduc one speci ic unc ion h ough me abolic coope a ion, such as me aboliza ion o die a y ibe and po en ially p ebio ic polysaccha ides like pec in (Saba e e al., 2022). Al e na i e symbion s, on he o he hand, occu only in some o hese minimal communi ies o in e ac ing mic oo ganisms. The e o e, hese ools can be used o design minimal mic obial conso ia equi ed o coope a i ely me abolise a gi en subs a e, such a complex p ebio ic s uc u e. I should be no ed ha his in o ma ion canno be ob ained by assembly- ee me hods o analysis o me agenomic da a. Thei applica ion o in silico analyze me aboliza ion o complex ca bohyd a e s uc u es equi es de ailed anno a ion o ca bohyd a e ac i e enzymes (CAZy) in MAGs eco e ed om mic obio a samples, and p edic ion o syne gis ic in e ac ion among coexis ing MAGs, wha may help elucida e p ebio ic e men a ion pa e ns in he mic obio a. Fo ins ance, in he con ex o pec in e men a ion, such app oaches ha e enabled in es iga ion o he po en ial o pec in e men a ion o amelio a e CD symp oms (Saba e e al., 2022). In his ega d, pec in is a e y complex polysaccha ide and equi es many di e en enzymes o i s enzyma ic deg ada ion including: glycoside-hyd olases (GH) (Nguyen e al., 2019), polysaccha ide-lyases (PL) (Su he land, 1995), ca bohyd a e-es e ases (CE) (A menda iz-Ruiz e al., 2018), and no el ca bohyd a e-binding modules (CBM) (Bo as on e al., 2004), as e iewed elsewhe e (Bonin & Pelloux, 2020; Ndeh e al., 2017). How- e e , mos s udies epo ed o da e in his ega d comp ise in silico ap- p oaches using publicly a ailable sequences and ew s udies epo MAG eco e y om in i o ecal e men a ion expe imen s using samples om he a ge popula ion ha is pa ien s o CD. Besides, o ou knowledge, no p e ious a emp s o de ine mic obial conso ia o gu commensal mic oo ganisms necessa y o coope a i ely me abolise pec in ha e been epo ed. In p io in es iga ions, we ha e p edic ed he po en ial o pec in s uc u es o modula e he gu mic obio a o CD pa ien s (Saba e e al., 2022) and ha e demons a ed in i o, h ough ecal e men a ions and 16S RNA sequencing he capabili y o apple pomaces and pec ins de i ed he e om o modula e key commensal axa in he mic obio a o CD pa ien s (Cal e e-To e e al., 2021). In his wo k we ha e u he del ed in o he e men a i e p ope ies o apple pomaces de i ed om p oduc ion o mono a ie al cide s ha ha e been epo ed o ep esen a p omising sou ce o pec in wi h a ied s uc u al cha ac e is ics depending on he apple a ie ies hey o igina e om: Sola ina a ie y ich in homogalac u onan (HG) and Pe ico and Regona a ie ies ich in hamnogalac u onan (RG)-I (Cal e e-To e e al., 2021). To his aim we ha e cha ac e ized me abolic capabili ies om MAGs eco e ed om ba ch ecal e men a ions pe o med wi h apple pomace and wi h hei pec ins, by using ecal samples om dono s wi h C ohn’s disease (CD) and heal hy dono s. Mic obial in e ac ions wi hin he modula ed com- muni ies we e simula ed using bioin o ma ic me hods o de ine mic o- bial conso ia o gu commensal species equi ed o me abolise pec in in he con ex o CD as a a ional p ebio ic design s a egy. 2. Ma e ials and me hods 2.1. Fecal e men a ion expe imen s and sho gun me agenomics sequencing Fecal e men a ion expe imen s analyzed in his epo we e p e i- ously desc ibed pa ially (Cal e e-To e e al. (2022)). B ie ly, ba ch ecal e men a ions we e pe o med wi h ecal samples p o ided by h ee heal hy male dono s (aged be ween 28 and 50 yea s) and h ee male CD pa ien s (aged be ween 24 and 60 yea s), by using apple pomace and pec in de i ed om h ee apple a ie ies (Pe ico, Regona and Sola ina). In Cal e e-To e e al. (2022), 16S RNA p o iling o he ecal e men a ions a se e al imepoin s was conduc ed and epo ed. He ein, we u he deepen in o he unc ional capaci y o he mic obial popula ions modula ed in he e men a ions by pe o ming sho gun me agenomics sequencing in a selec ion o samples, coupled o unc- ional analysis o MAGs and p edic ion o me abolic syne gies among he modula ed popula ions. To al samples (n =42) selec ed o me agenomic sequencing included: i) ecal homogena es om heal hy (n =3) and CD (n =3) dono s, ii) ecal e men a ions o apple pomace (n =9) and pec in (n = 9) de i ed om Pe ico, Regona and Sola ina apple a ie ies collec ed a 8h using heal hy dono s, iii) ecal e men a ions o apple pomace (n =9) and pec in (n =9) de i ed om Pe ico, Regona and Sola ina apple a- ie ies collec ed a 8h using CD dono s. Each expe imen was pe o med in iplica e. Then, mic obial DNA was isola ed by using he Powe Soil P oKi (Qiagen) acco ding o ou p e ious wo k (Cal e e-To e e al., 2022). To al DNA ex ac ed was submi ed o an ex e nal sequencing se ice (www.BaseClea .com) o gene a e pai ed-end sequence eads (2 × 150bp) showing an a e age minimum o 25 million eads pe sample using an Illumina No aSeq sys em. Reads we e demul iplexed o gene a e FASTQ ead sequence iles using bcl2 as q2 ( 2.18) and eads con aining (pa ial) adap e s we e clipped (up o a minimum ead leng h o 50 bp). Addi ional quali y-con ol s eps included he emo al o con aminan low-quali y eads using Kneadda a ( 0.7.4) and T immo- ma ic ( 0.39) so wa e. Fo his pu pose, Bow ie2 ( 4.5) was used o map me agenomic eads agains he e e ence da abases "Homo sapiens hg37 and human con amina ion Bow ie2" ( 0.1), o emo e hos con amina ion. The aw sequences da a we e deposi ed in he Sequence Read A chi e (SRA) o he NCBI (h ps://www.ncbi.nlm.nih.go /s a, accessed on July 18, 2023) unde BioP ojec PRJNA995911. I. Cal e e-To e e al. LWT 187 (2023) 115362 3 2.2. Assembly- ee analysis o me agenomes To cha ac e ise mic obio a composi ion and me abolic unc ions o ecal e men a ion samples, an assembly- ee analysis was i s pe - o med. This assembly- ee analysis allows he de e mina ion o axa abundances in he co e mic obio a composi ion and he iden i ica ion o low-abundance axa ha migh be los du ing he me agenome assembly p ocess. Decon amina ed eads we e analyzed ollowing Me aPhlAn 3.0 ( 3.0.4) and HUMAnN 3.0 ( 3.0.0) pipelines o elucida e axonomic and unc ional p o iles, espec i ely (F anzosa e al., 2018; T uong e al., 2017). Wi h his aim, ChocoPhlAn da abase ( e sion "mpa_ 30_Choco- PhlAn_201,901 ″ ) con aining clade-speci ic ma ke genes and UniRe 90 ( e sion "uni e 90_201,901 ″ ) p o ein da abase, we e used o pe o m axonomic and unc ional analyses. The abundances o gene amilies and me abolic pa hways we e e-no malized and exp essed in uni s o copies pe million. 2.3. Reco e y o me agenome-assembled genomes (MAGs) To be e cha ac e ise pec in-deg ading ac i i ies o mic obial com- muni ies ound in ecal e men a ion samples, me agenome-assembled genomes (MAGs) we e eco e ed ollowing an assembly-based pipe- line desc ibed in a p e ious wo k o ou esea ch g oup (Saba e e al., 2022). B ie ly, he assembly o quali y- il e ed eads was pe o med wi h MEGAHIT ( 1.2.9) conside ing a maximum k-me size o 127 o gene a e he ollowing se ies o k-me s: k-21, k-31, k-41, k-51, k-61, Fig. 1. P incipal coo dina es analysis (PCoA) o axonomic p o iles (A), gene amilies (B) and me abolic pa hways (C) ound in he mic obio a o each g oup o samples: apple pomace (Pom) and pec in (Pec ) de i ed om di e en apple a ie ies (Pe ico, Regona and Sola ina) ha we e subjec ed o ecal e men a ion using ecal homogena es (Homog) om heal hy and C ohn’s disease (CD) dono s. PC: p incipal coo dina e. The pe cen age o a iance explained by each PC is indica ed in he axis. Numbe s iden i y samples o igina ing om di e en ecal dono s. I. Cal e e-To e e al. LWT 187 (2023) 115362 4 k-71, k-81, k-91, k-101, k-111, k-121, k-127. Me agenome eads we e mapped agains he assembly using Bow ie2 ( .4.5) and ou pu bam iles gene a ed we e so ed and indexed. Con igs la ge han 1.5 kilobases we e binned sepa a ely using Me aba 2 ( .2.2.15). MAGs comple eness and con amina ion we e assessed using CheckM ( .1.1.3) lineage-speci ic wo k low o selec MAGs showing comple eness highe han 50% and con amina ion lowe han 5%. Taxonomic iden i ica ion o MAGs was pe o med using GTDB-Tk ( 2.1.1) pipeline and ca bohyd a e-ac i e enzymes (CAZy) p o iles o MAGs we e anno a ed ollowing “ un_dbcan” so wa e de eloped by Zhang e al. (2018). 2.4. Me abolic modelling Once me abolic unc ions o MAGs we e anno a ed, po en ial me a- bolic in e ac ions be ween MAGs we e calcula ed. Fo his pu pose, MAGs we e anno a ed using P okka ( 1.14.6) o gene a e s anda d Genbank iles (.gbk) iles con aining sequences and anno a ions ha we e used as inpu o me age2me abo ( 1.5.0) so wa e. To simula e mic obial me abolic in e ac ions in he p esence o pec in, a “seeds” ile con aining pec in and di e en nu ien s ha may be ound in human gu acco ding o p e ious s udies was p o ided (Belcou e al., 2020; Saba e e al., 2022). Di e en simula ions we e compu ed o MAGs eco e ed om ecal e men a ions o apple pomace and pec in using heal hy and CD dono s. 2.5. S a is ical analysis o samples S a is ical analysis o assembly- ee analysis ou pu s was pe o med on R ( .4.2.3). B ie ly, alpha and be a di e si y es ima o s we e compu ed using “Phyloseq” and “Mic obiome” R packages (McMu die & Holmes, 2013; Lah i & She y, 2017) and P incipal Coo dina e Analysis (PCoA) and composi ion ba plo s we e gene a ed using “Mic obiome” R package o illus a e majo di e ences in he mic obio a p o iles. S a- is ically signi ican di e ences (p <0.05 and p adj <0.25) in mic o- biome composi ion and unc ion we e calcula ed using se e al mic obio a-speci ic s a is ical me hods (ANCOM, LE Se and DESeq2) implemen ed in “mic obiomeMa ke ” R package (Cao e al., 2022; Lo e e al., 2014; Mandal e al., 2015; Sega a e al., 2011). Me agenomic eads we e no malized p io di e en ial analysis. Taxonomic clades and unc ions classi ied as signi ican ly di e en ial mic obes (mic obiome ma ke s) by any o hese me hods we e selec ed o u he analysis. To in es iga e he associa ions be ween mic obial clades and me aboli e p oduc ion, signi ican co ela ions (p <0.05) be ween mic obial composi ion and sho -chain a y acids (SCFAs) le els epo ed by Cal e e-To e e al. (2022) we e calcula ed and exp essed as Pea son co ela ion coe icien s. CAZy domains in ol ed in pec in me abolism we e compa ed h ough hie a chical clus e ing compu ed using R ( .4.2.3) basic unc- ion “hclus ”. 3. Resul s 3.1. Assembly- ee analysis o me agenomes Mic obial p o iles analysis o ecal e men a ions by assembly- ee me hods was i s ca ied ou allowing a p ope es ima ion o mic obi- al abundances including low-abundance mic obial eads axa ha migh be disca ded du ing me agenome assembly. Fi s , alpha di e si y mea- su es including Chao1, Shannon, Simpson and In e se Simpson indices we e calcula ed o es ima e a iabili y o gene a wi hin samples. Mean Chao1 index conside ing all samples unde s udy was 93.0 ±14.1 ye Chao1 indexes showed no s a is ically signi ican (p >0.05) di e ences among sample subse s including: i) ecal homogena es om heal hy (91.7 ±10.1) and CD (90.0 ±20.4) dono s, ii) apple pomace e men- a ion using samples om heal hy (95.3 ±9.8) and CD (91.9 ±18.5) dono s, iii) apple pec in e men a ion using samples om heal hy (96.0 ±9.7) and CD (89.6 ±18.5) dono s (Supplemen a y Ma e ial Fig. S1). The di e en alpha di e si y coe icien s showed simila ends, indi- ca ing a high in e indi idual a iabili y wi hin g oups: ecal e men a- ions pe o med using samples om heal hy dono 1 showed lowe alpha-di e si y alues han hose ob ained o he es o heal hy do- no s, while ecal e men a ions pe o med using samples om CD dono 4 showed highe alpha-di e si y alues han hose ob ained o he es o CD dono s. Be a di e si y analysis o measu e mic obial di e si y be ween g oups o samples was hen calcula ed. No signi ican (p >0.05) di - e ences in B ay-Cu is dis ances be ween apple and pec in ac ions and be ween heal hy and CD dono s we e ound. This migh be a ibu ed o in ag oup a iabili y, in acco dance o he alpha di e si y analysis p e iously desc ibed. Clus e ing based on B ay-Cu is dis ances g ouped samples mainly acco ding o he heal h s a us o he dono (i.e. samples om heal hy and CD dono s) al hough se e al samples co esponding o he same g oup (i.e. apple pomace o pec in) we e clus e ed in o di e en b anches o he hie a chical model (Supplemen a y Ma e ial Fig. S2A). This model could no comple ely disc imina e be ween sam- ples om di e en g oups due o in e indi idual a iabili y. In ag ee- men wi h his obse a ion, P incipal Coo dina es Analysis (PCoA) be e disc imina ed samples om heal hy and CD dono s (Fig. 1A). In addi ion, se e al subg oups o samples co esponding o he same heal hy o CD dono we e gene a ed. Howe e , samples could no be Fig. 2. Mos abundan gene a ound in he mic obio a o each g oup o samples: apple pomace (Pom) and pec in (Pec ) de i ed om di e en apple a ie ies (Pe ico, Regona and Sola ina) ha we e subjec ed o ecal e men a ion using ecal homogena es (Homog) om heal hy and C ohn’s disease (CD) dono s. Da a a e exp essed as abundance pe cen ages (%). I. Cal e e-To e e al. LWT 187 (2023) 115362 5 Table 1 Mic obial axa showing he highes abundances in each g oup o samples: apple apple pomace (Pom) and pec in (Pec ) de i ed om di e en apple a ie ies (Pe ico, Regona and Sola ina) ha we e subjec ed o ecal e men a ion using ecal homogena es (Homog) om heal hy and C ohn’s disease (CD) dono s. Mean abundances and s anda d de ia ions (SD) o axonomic clades a e shown. Highe in: Taxa Taxonomic le el: Heal hy - Homog (Mean) Heal hy - Homog (SD) Heal hy - Pom (Mean) Heal hy - Pom (SD) Heal hy - Pec (Mean) Heal hy - Pec (SD) CD - Homog (Mean) CD - Homog (SD) CD - Pom (Mean) CD - Pom (SD) CD - Pec (Mean) CD - Pec (SD) Heal hy - Homog Ruminococcaceae Family 17.75 11.16 15.48 9.84 12.64 8.76 8.71 6.04 8.54 3.47 7.46 3.78 Heal hy - Homog Lachnospi aceae unclassi ied Genus 11.85 11.19 10.87 10.25 8.02 9.12 1.23 2.13 3.25 4.89 1.58 2.93 Heal hy - Homog Eubac e ium ec ale Species 11.85 11.19 10.87 10.25 8.02 9.12 1.23 2.13 3.25 4.89 1.58 2.93 Heal hy - Pom Faecalibac e ium Genus 7.58 4.57 8.31 3.72 7.06 3.85 2.02 1.89 2.95 2.85 2.16 2.41 Heal hy - Pom Faecalibac e ium p ausni zii Species 7.58 4.57 8.31 3.72 7.06 3.85 2.02 1.89 2.95 2.85 2.16 2.41 Heal hy - Pec Ac inobac e ia Class 19.00 26.24 14.66 15.90 28.05 31.08 15.20 24.06 7.47 10.39 6.49 10.96 Heal hy - Pec Bi idobac e iales O de 18.97 26.25 14.63 15.92 28.00 31.12 15.08 24.06 7.36 10.41 6.37 10.98 Heal hy - Pec Bi idobac e iaceae Family 18.97 26.25 14.63 15.92 28.00 31.12 15.08 24.06 7.36 10.41 6.37 10.98 Heal hy - Pec Bi idobac e ium Genus 18.97 26.26 14.63 15.92 27.99 31.13 15.08 24.06 7.36 10.41 6.37 10.97 CD - Homog Ac inobac e ia Phylum 25.48 34.83 17.01 17.68 30.03 31.98 30.71 27.66 16.13 11.52 16.31 12.05 CD - Homog Co iobac e iia Class 6.48 8.59 2.34 1.96 1.97 1.16 15.51 5.52 8.66 6.08 9.83 6.80 CD - Homog Co iobac e iales O de 6.28 8.54 2.19 2.02 1.82 1.27 14.54 5.35 8.33 5.89 9.48 6.62 CD - Homog Co iobac e iaceae Family 6.23 8.45 2.19 2.02 1.81 1.27 14.54 5.35 8.33 5.89 9.48 6.62 CD - Homog Collinsella Genus 6.21 8.43 2.18 2.01 1.81 1.26 14.53 5.35 8.33 5.89 9.48 6.62 CD - Pec Ve ucomic obia Phylum 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 CD - Pec P o eobac e ia Phylum 0.19 0.17 0.49 0.38 0.70 0.53 0.74 0.47 1.47 1.37 2.00 1.62 CD - Pec Ve ucomic obiae Class 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 CD - Pec Ve ucomic obiales O de 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 CD - Pec Akke mansiaceae Family 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 CD - Pec Akke mansia Genus 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 CD - Pec Akke mansia muciniphila Species 0.15 0.25 0.27 0.40 0.26 0.40 8.97 12.91 9.72 12.55 11.18 13.55 I. Cal e e-To e e al. LWT 187 (2023) 115362 6 disc imina ed acco ding o he apple a ie y (Pe ico, Regona and Sola ina), hence sugges ing ha in e indi idual a iabili y is a s onge sou ce o a ia ion han he speci ic apple a ie y. Taxonomic composi ion analysis e ealed ha , among bac e ial gene a wi h majo abundances, Bi idobac e ium, Bac e oides, Faecali- bac e ium and Ruminococcus exhibi highe abundances in samples om heal hy dono s han hose o CD dono s (Fig. 2). In con as , P e o ella, Collinsella and an unclassi ied P e o ellaceae gene a showed highe abundances in CD samples han heal hy samples. S a is ically signi ican (p <0.05 and p adj <0.25) di e ences in he abundances o indi idual axa be ween g oups we e also de e mined (Table 1). Acco ding o he esul s ob ained, Ruminococcaceae, an unclassi ied Lachnospi aceae genus and Eubac e ium ec ale showed he highes abundances in ecal homogena es o heal hy dono s whe eas Collinsella was highe in ecal homogena es om CD pa ien s. Faecalibac e ium p ausni zii and Bi ido- bac e ium showed he highes abundances a e ecal e men a ion o apple pomace and pec in, espec i ely, using samples om heal hy do- no s. Simila ly, Akke mansia muciniphila showed he highes abundances a e ecal e men a ion o pec in using samples om CD pa ien s (Table 1). In e es ingly, no s a is ically signi ican di e ences (p <0.05 and p adj <0.25) we e obse ed in he axonomic p o iles o samples om di e en apple a ie ies in ag eemen wi h PCoA whe e samples om di e en a ie ies could no be disc imina ed (Fig. 1). Co ela ion analysis e ealed posi i e associa ions be ween se e al SCFAs p oduce s like Bi idobac e ium and ibe -deg ading species like A. muciniphila and Cop ococcus eu ac us, and isobu y ic, ale ic, iso ale ic and cap oic acids le els (Fig. 3). Wi h ega d o he unc ional analysis o me agenomes, he gene coun (i.e., o al numbe o di e en mic obial genes) showed no s a- is ically signi ican (p >0.05) di e ences be ween g oups: i) ecal ho- mogena es om heal hy (76054.7 ±21812.7) and CD (61188.0 ± 40925.8) dono s, ii) apple pomace e men a ion using samples om heal hy (78391.1 ±23745.9) and CD (64005.3 ±38706.4) dono s, iii) apple pec in e men a ion using samples om heal hy (77197.1 ± 29959.8) and CD (60732.3 ±39541.0) dono s. No signi ican (p >0.05) di e ences in be a di e si y es ima o s be ween apple and pomace ac ions and be ween heal hy and CD dono s we e ound. Clus e analysis o unc ional p o iles led o simila clus e s o hose ob ained in he axonomic analysis, whe e samples we e classi ied mainly acco ding o hei heal h s a us (i.e. samples om heal hy and CD dono s) and simila g oups o samples (i.e. apple pomace o pec in ac ions) we e clus e ed in he same b anches (Supplemen a y Ma e ial Figs. S2B and C). PCoA o unc ional p o iles disc imina ed samples om heal hy and CD dono s and samples om he same heal hy o CD dono we e g ouped oge he (Fig. 1B and C). This beha iou was obse ed o bo h mic obial gene amilies and me abolic pa hways. Mic obial gene amilies and me abolic pa hways showing he highes (p <0.05 and p adj <0.25) abundances in ecal homogena e and e - men a ions o apple pomace samples om heal hy dono s co esponded o F. p ausni zii and E. ec ale while hose genes and pa hways inc eased in e men a ion o pec in samples co esponded o B. adolescen is and F. p ausni zii (Table 2). These pa hways we e in ol ed in amino acid, nucleo ide, i amin and ca bohyd a e me abolism (Supplemen a y Ma e ial Table S1). In con as , CD ecal homogena es we e en iched in C. ae o aciens genes. In e es ingly, pec in samples om CD dono s was en iched in a wide ange o A. muciniphila gene amilies and me abolic pa hways (Table 2), in ol ed in nucleo ide and amino acid me abolism (Supplemen a y Ma e ial Table S1). These esul s ag ee wi h he axo- nomic analysis o samples (Fig. 1, Table 1). 3.2. Reco e y o me agenome-assembled genomes (MAGs) Assembly-based analyses allowed o eco e a o al 1080 MAGs om me agenome samples including: i) ecal homogena es om heal hy (n = 237) and CD (n =314) dono s, ii) apple pomace e men a ion using samples om heal hy (n =117) and CD (n =134) dono s, iii) apple pec in e men a ion using samples om heal hy (n =142) and CD (n = 136) dono s (Supplemen a y Ma e ial Table S2). Mos ep esen ed axa among MAGs eco e ed om samples ob ained using heal hy dono s included Aga hobac e ec alis, C. eu ac us, Pa abac e oides dis asonis, Ruminococcus species and no el Ch is ensenellales UBA11524 sp000437595 species, whe eas A. muciniphila, P. dis asonis and Rumi- nococcus species a e he mos ep esen ed among MAGs eco e ed om samples ob ained om CD pa ien s (Supplemen a y Ma e ial Table S2). Func ional anno a ion o CAZy domains o MAGs e ealed simila - i ies in enzyme domains ound in all e men a ion g oups (Fig. 4). I should be no ed ha pec in is mainly composed o linea chains o α -1,4- D- GalA chains and se e al ami ied domains ha comp ise al e na e sequences o GalA and α -(1, 2) linked α -L- hamnosyl esidues which may be subs i u ed a O-4 wi h linea o b anched oligosaccha ides (Saba e , Fig. 3. Co elog ams showing he associa ions be ween mic obial composi ion and sho -chain a y acids (SCFAs) le els. Blue and ed do s indica e posi i e and nega i e co ela ions exp essed as Pea son co ela ion coe icien s. Colou in ensi y and do size a e in p opo ion o magni ude. These co ela ions we e s a is ically signi ican (p <0.05). (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) I. Cal e e-To e e al. LWT 187 (2023) 115362 7 Cal e e-To e, e al., 2021; Tan & Nie, 2020). The e o e, CAZy domains ha may be in ol ed in pec in me abolism comp ise a wide ange o ac i i ies including a abino u anosidases, xylosidases, xylanases, poly- galac u onases, hamnosidases, pec in es e ases and hamnogalac u - onan lyases, among o he s (Fig. 4). Wi h ega d o pec in es e ases, pec in me hyles e ase domains we e anno a ed in Selenomonas and no el Ch is ensenellales UBA11524 MAGs eco e ed om pomace and pec in samples e men ed using aecal samples om heal hy dono s (Fig. 4A). Simila ly, pec in ace yles e ase domains we e anno a ed in Gemmige and Ch is ensenellales UBA11524 eco e ed om hese e men a ions. These axa and o he gene a (Akke mansia, Allop e o ella, Fimenecus, Pa abac e oides, Ruminococcus and Selenomonas) eco e ed om ecal e men a ions o bo h subs a es (apple pomace and pec in) showed he highes numbe o pec in- deg ading ac i i ies including a abinases, galac u onases, galac u - onidases and enzymes ac ing on complex hamnogalac u onan-II ami- ied domains like ace ic acid and KDO (2-ke o-3-deoxy-D-manno- oc ulosonic acid) hyd olases (Fig. 4A). In e es ingly, hamnogalac u - onan endolyase was cha ac e is ic o Ruminococcus species (Fig. 4A). In e men a ions wi h aecal samples om CD pa ien s, pec in es- e ases we e anno a ed in En e oclos e , Gemigge , Lachnospi a and P e- o ella MAGs eco e ed om pomace and pec in samples (Fig. 4B). These axa and o he gene a (Akke mansia, C yp obac e oides, Eisenbe gella and no el species Acu alibac e aceae CAG-177 and UBA1417) showed he highes numbe o pec in-deg ading domains in MAGs eco e ed om apple pomace and pec in samples (Fig. 4B). Simila CAZy p o iles we e obse ed o MAGs eco e ed om bo h pomace and pec in samples e men ed using CD dono s (Fig. 4B). 3.3. Me abolic modelling Once pec in-deg ading ac i i ies o MAGs we e anno a ed, mic obial in e ac ions we e simula ed compu a ionally, wi h he aim o de ining gu mic obial conso ia capable o me abolise pec in in he con ex o CD, as a a ional p ebio ic o synbio ic design s a egy. Essen ial sym- bion s, needed o sa is y speci ic me abolic unc ions on each minimal communi y o in e ac ing mic obes, and al e na i e symbion s, occu only in some o hese minimal communi ies, we e de e mined acco ding o Belcou e al. (2020). Di e ences in he po en ial me abolic in- e ac ions o MAGs we e obse ed acco ding o mul iple c i e ia: heal h s a us (heal hy and CD dono s), subs a e (apple pomace and pec in) and apple a ie y (Pe ico, Regona and Sola ina) (Figs. 5 and 6; Supple- men a y Ma e ial Fig. S3). Key symbion s de e mined o each apple a ie y comp ise Akke - mansia, Bi idobac e ium and Ruminococcus species among o he s (Figs. 5 and 6). In gene al, Bi idobac e ium and Ruminococcus MAGs comp ise essen ial symbion s in all e men a ion expe imen s while A. muciniphila MAGs comp ise al e na i e symbion s in e men a ion expe imen s ca ied ou using CD samples. No ably, essen ial and al e na i e sym- bion s p edic ed a e di e en acco ding o he heal h s a us and sub- s a e used o he e men a ions. Fo ins ance, in apple pomace om Pe ico a ie y e men ed using heal hy dono s, essen ial Ruminococcus symbion s (Ruminococcus sp000433635 and R. bici culans) es ablish mul iple in e ac ions wi h wo di e en communi ies o al e na i e symbion s comp ising o he Ruminococcus and P e o ella species: i) Ruminococcus sp000980705, ii) R. b omii and P e o ella sp002251385 (Supplemen a y Ma e ial Fig. S3A). Howe e , pec in samples om his a ie y e men ed using heal hy dono s showed a di e en in e ac ion p o ile: essen ial symbion s P. dis asonis and Anae o ib io sp900548165 in e ac wi h al e na i e symbion s P e o ella sp000434975 and Fi mi- cu es bac e ium CAG-41 sp900066215, while essen ial symbion Fi mi- cu es bac e ium CAG-83 sp000431575 in e ac s wi h al e na i e symbion P e o ella sp003447235 (Supplemen a y Ma e ial Fig. S3B). On he o he hand, apple pomace ac ions om his a ie y e men ed using CD dono s e ealed in e ac ions be ween essen ial symbion s A. muciniphila and P. cop i and al e na i e symbio ic communi ies including Odo ibac e splanchnicus, and P. dis asonis among o he s (Supplemen a y Ma e ial Fig. S3C). In con as , in apple pec in om he same a ie y e men ed using CD dono s, Akke mansia muciniphila s ains comp ise al e na i e symbion s ha show po en ial in e ac ions wi h all essen ial symbion s (Supplemen a y Ma e ial Fig. S3D). These esul s highligh he ole o Ruminococcus in symbio ic in e ac ions o all subs a es ega dless heal h s a us, as well as A. muciniphila in CD mic obio a. Di e ences be ween apple a ie ies we e also in es iga ed (Figs. 5 and 6). Po en ial in e ac ions be ween R. callidus and Ruminococcus CAG-177 wi h A. muciniphila we e iden i ied in apple pomace om Regona a ie y e men ed using CD dono s (Fig. 5B). In e ac ions be- ween A. muciniphila s ains and all essen ial symbion s we e also ound in ecal e men a ions o apple pomace om Sola ina a ie y (Fig. 5C). Simila ly, A. muciniphila and R. bici culans showed po en ial c oss- eeding mechanisms wi h all essen ial symbion s in e men a ion ex- pe imen s o pec in om hese a ie ies (Fig. 6B and C). 4. Discussion This wo k p o ides a comp ehensi e cha ac e iza ion o he ca bo- hyd a e me abolism o MAGs eco e ed om ecal e men a ion wi h apple pomace and pec in and de ines no el mic obial conso ia capable Table 2 Numbe o mic obial gene amilies and me abolic pa hways showing he highes abundances in each g oup o samples: apple apple pomace (Pom) and pec in (Pec ) de i ed om di e en apple a ie ies (Pe ico, Regona and Sola ina) ha we e subjec ed o ecal e men a ion using ecal homogena es (Homog) om heal hy and C ohn’s disease (CD) dono s. These gene amilies and pa hways a e summa ised by bac e ial species. Mic obial gene amilies Highe in: Taxa n genes/pa hways Heal hy - Homog Faecalibac e ium p ausni zii 2104 Heal hy - Homog Eubac e ium ec ale 1534 Heal hy - Homog Collinsella ae o aciens 91 Heal hy - Homog Bi idobac e ium adolescen is 4 Heal hy - Homog Bi idobac e ium longum 3 Heal hy - Pom Faecalibac e ium p ausni zii 899 Heal hy - Pom Eubac e ium ec ale 546 Heal hy - Pec Bi idobac e ium adolescen is 1267 Heal hy - Pec Faecalibac e ium p ausni zii 542 Heal hy - Pec Eubac e ium ec ale 34 Heal hy - Pec Bi idobac e ium longum 11 CD - Homog Collinsella ae o aciens 932 CD - Homog Akke mansia muciniphila 14 CD - Homog Bi idobac e ium longum 10 CD - Homog Faecalibac e ium p ausni zii 3 CD - Homog Bi idobac e ium adolescen is 2 CD - Homog Eubac e ium ec ale 2 CD - Pom Akke mansia muciniphila 8 CD - Pom Faecalibac e ium p ausni zii 6 CD - Pom Bi idobac e ium longum 1 CD - Pom Eubac e ium ec ale 1 CD - Pec Akke mansia muciniphila 1433 CD - Pec Faecalibac e ium p ausni zii 8 CD - Pec Bi idobac e ium longum 2 CD - Pec Collinsella s e co is 1 Mic obial me abolic pa hways Highe in: Taxa n genes/pa hways Heal hy - Homog Faecalibac e ium p ausni zii 50 Heal hy - Homog Eubac e ium ec ale 41 Heal hy - Homog Collinsella ae o aciens 1 Heal hy - Pom Faecalibac e ium p ausni zii 12 Heal hy - Pom Eubac e ium ec ale 11 Heal hy - Pec Bi idobac e ium adolescen is 49 Heal hy - Pec Eubac e ium ec ale 1 Heal hy - Pec Faecalibac e ium p ausni zii 1 CD - Homog Collinsella ae o aciens 47 CD - Pec Akke mansia muciniphila 33 I. Cal e e-To e e al. LWT 187 (2023) 115362 8 o me abolise pec in in he con ex o CD. In his ega d, A. muciniphila and se e al Ruminococcus clades including no el species showed a wide ange o pec in-deg ading enzymes and po en ial syne gis ic in- e ac ions wi h o he gu mic obes ound in CD mic obio a. I should be no ed ha A. muciniphila and Ruminococcus species a e eme ging p o- bio ics. P e ious esea ch highligh s he in e es o de elop nex gen- e a ion p obio ics o p omo e hese axa and o a ge speci ic disease condi ions such as CD (Lin e al., 2019; Lo dan e al., 2020). Resul s he e p esen ed ag ee wi h p e ious bioin o ma ic app oaches whe e a wide ange o polysaccha ide-deg ading enzymes we e anno a ed in MAGs o A. muciniphila, leading o po en ial coope a ion mechanisms be ween hese axa and o he gu mic obes in he p esence o pec in (Saba e e al., 2022). In ac , i has been p oposed ha pec in adminis a ion may amelio a e gu dysbiosis in CD by p omo ing me abolic in e ac ions be ween key species such as A. muciniphila (Saba e e al., 2022). Expe imen al s udies epo c oss- eeding mechanisms be ween A. muciniphila and o he gu commensals like E. hallii (F i s e al., 2021). Simila ly, me abolic coope a ion mechanisms be ween di e en Ruminococcus species in he con ex o ibe me abolism ha e been e- po ed (C os e al., 2018), in ag eemen wi h ou esul s. P e ious in- es iga ions om ou esea ch g oup epo an inc emen in Akke mansia, no el Lachnospi aceae and Ruminococcaceae gene a a e ecal e men a ion o apple pomace and pec in (Cal e e-To e e al., 2022), in ag eemen wi h esul s ob ained he e h ough a sho gun me agenomics app oach. In addi ion, o he au ho s epo ed ha adminis a ion o a abinoxylan oligosaccha ides and high me hoxyla ed pec in p omo e A. muciniphila (Zhai e al., 2019), while Lachnospi a and Ruminococcaceae we e inc eased a e ecal e men a ion o some ci us and suga bee pec in ac ions (La sen e al., 2019). Fecal e men a ion o o he unc ional polysaccha ides such as so ghum a abinoxylans led o an inc emen in Bac e oides and Bi idobac e ium and Lachnospi aceae (Yao e al., 2023). The majo i y o pec in e men a ion s udies epo 16S RNA gene sequence da a al hough sho gun me agenomics analysis desc ibed in his wo k allows axonomic iden i ica ion a species and s ain le el as well as complex me abolic s udies o design no el con- so ia o in e ac ing mic obes (Bikel e al., 2015; V´ azquez-Cas ellanos Fig. 4. Hea map showing he p esence o di e en glycosidases (indica ed as black cells) in me agenome-assembled genomes (MAGs) eco e ed om each g oup o samples: apple pomace (Pom) and pec in (Pec ) de i ed om di e en apple a ie ies ha we e subjec ed o ecal e men a ion using ecal homogena es om heal hy (A) and C ohn’s disease (CD) (B) dono s. The pe cen age (%) o MAGs con aining each unc ional domain is shown. Codes co esponding o he Ca bohyd a e-Ac i e enZYmes Da abase (CAZy) amily o each enzyme ha e been assigned. I. Cal e e-To e e al. LWT 187 (2023) 115362 9 e al., 2019). Mic obial abundances o Bi idobac e ium, Bac e oides, Ruminococcus and Faecalibac e ium we e highe in me agenomes om heal hy dono s han in hose om CD pa ien s in ag eemen wi h p e ious s udies (Clooney e al., 2021; Saba e e al., 2022). I has been sugges ed ha ch onic in lamma o y diseases lead o a dec ease o bac e ia wi h an i-in lamma o y capaci ies like F. p ausni zii (Nishida e al., 2018; Sokol e al., 2008). Simila ly, Ruminococcus abundances migh be dec eased in pa ien s wi h CD, including child en, adolescen s and adul s (Nishida e al., 2018; Vacca e a., 2020), in ag eemen wi h ou esul s, al hough some speci ic clades wi hin his g oup, ha e been associa ed wi h p o-in lamma o y p ope ies (Henke e al., 2019). Associa ions be ween mic obial composi ion and SCFAs le els we e also epo ed in he li e a u e. In his ega d, posi i e co ela ions be- ween Akke mansia and majo SCFAs ace ic, p opanoic and bu y ic acids ha e been epo ed (Cal e e-To e e al., 2022), in ag eemen wi h ou esul s. Pec in u iliza ion by he gu mic obio a may lead o he p o- duc ion o bu y ic acid, mainly h ough u iliza ion o galac ose and galac u onic acid o pec in chains (Fu e al., 2018). Speci ically, A. muciniphila and F. p ausni zii comp ise specialis gu anae obic bac- e ia in ol ed in he b eakdown o die a y ibe (Duncan e al., 2016), leading o SCFAs p oduc ion and nume ous heal h bene i s (Lin e al., 2019). In conclusion, ou esul s sugges ha he s uc u e o he coope - a ing mic obial conso ia a y depending on he pa icula a ie y o apple pomace/pec in. This obse a ion ag ees wi h p io wo ks ha epo ed ha ine polysaccha ide s uc u e de e mines he selec ion o dis inc e men ing conso ia (Yao e al., 2023). Mos s udies up o da e epo changes in he mic obio a axonomic p o iles induced by pec in adminis a ion. Howe e , ew s udies o da e ha e explo ed he po en- ial in e ac ions be ween mic obial species modula ed by pec in in he con ex o in lamma o y diseases. In his ega d, esul s he e p esen ed sugges ha A. muciniphila and Ruminococcus species a e key mic obes p esen in CD mic obio a ha migh be in ol ed in mul iple c oss- eeding mechanisms o me abolise pec in. Bioin o ma ic pipelines employed in his wo k could be o special in e es o assis he a ional design o p ebio ic o symbio ic in e en ions o modula e he gu mic obio a o speci ic popula ion g oups. In addi ion, hese me hods allow he selec ion o key species and mic obial conso ia o me abolise a gi en subs a e including biologically ac i e polysaccha ides (Shakh- ma o e al., 2019). In i o demons a ion o he coope a i e me a- bolism o complex pec in by he mic obial conso ia he ein de ined, will undoub edly aid owa ds he de elopmen o mic obiome- a ge ed p ebio ic and/o symbio ic s a egies o speci ic popula ion g oups. CRediT au ho ship con ibu ion s a emen Ines Cal e e-To e: In es iga ion, Fo mal analysis, W i ing – e iew & edi ing. Ca los Saba e : Me hodology, Fo mal analysis, W i ing – o iginal d a . Abela do Ma golles: Concep ualiza ion, W i ing – e- iew & edi ing. Lo ena Ruiz: Concep ualiza ion, Resou ces, W i ing – e iew & edi ing. Decla a ion o compe ing in e es Au ho s ha e no con lic o in e es o decla e. Fig. 5. Me abolic ne wo k illus a ing po en ial c oss- eeding mechanisms be ween me agenome-assembled genomes (MAGs) eco e ed om apple pomace om Pe ico (A), Regona (B) and Sola ina (C) a ie ies e men ed using C ohn’s disease pa ien s samples. Ne wo k nodes (i. e. ci cles con aining di e en mic obial communi ies showing equi alen me abolic unc ions) a e connec ed by black lines indica ing syne gis ic ela ionships be ween communi ies and complemen a y me abolic. MAGs inside he same node play he same ole and could be eplaced by o he membe s om he same communi y wi h he excep ion o essen ial symbion s included in he g een node. Blue and ed nodes co espond o Akke mansia and Ruminococcus species, espec i ely. MAGs om ne wo k A) include: 1: Clos idium AP scindens, 2: Bi idobac e ium longum, 3: E epia gaba o ous, 4: Campylobac e B sp905373215, 5: CAJFUH01, 6: Cop omonas sp900066535, 7: Eisenbe giella po ci, 8: BX12 sp014333425, 9: Ruminococcus D bici culans, 10: Ruminococcus C callidus, 11: Mailhella exc emen igallina um, 12: CAG-170 sp000432135, 13: Phasco- la c obac e ium A succina u ens, 14: En e oclos e sp001517625, 15: UBA1394 sp900538575, 16: Duodenibacillus in es ina ium, 17: UBA1417 sp003531055, 18: Me hanob e ibac e A smi hii, 19: Akke mansia muciniphila, 20: Alis ipes A ihumii, 21: Fimi icinus sp900544375, 22: Angelakisella sp900547385, 23: Odo ibac e splanchnicus, 24: C yp obac e oides sp000433355, 25: P e o ella cop i, 26: Pa abac e oides dis asonis, 27: Lachnospi a eligens A, 28: Akke mansia muciniphila. MAGs om ne wo k B) include: 1: Ace a i ac o sp003447295, 2: Cop omonas sp900066535, 3: Campylobac e B sp905373215, 4: BX12 sp014333425, 5: En e oclos e clos- idio o mis, 6: Clos idium AP scindens, 7: Gemmige qucibialis, 8: Ruminococcus D bici culans, 9: Ru henibac e ium lac a i o mans, 10: Pa abac e oides dis asonis, 11: ER4 sp000765235, 12: Bilophila sp902373525, 13: Phascola c obac e ium aecium, 14: CAG-217 sp000436335, 15: Duodenibacillus in es ina ium, 16: Phascola c obac e ium A succina u ens, 17: Lachnospi a eligens A, 18: UBA1417 sp003531055, 19: CAG-170 sp000432135, 20: Akke mansia muciniphila, 21: Ruminococcus C callidus, 22: Akke mansia muciniphila, 23: Akke mansia muciniphila, 24: CAG-177 sp003514385, 25: UBA1394 sp900538575, 26: Angelakisella sp900547385. MAGs om ne wo k C) include: 1: Cop omonas sp900066535, 2: Bi idobac e ium longum, 3: Clos idium AP scindens, 4: CAJFUH01, 5: Eisenbe giella po ci, 6: BX12 sp014333425, 7: Ruminococcus C callidus, 8: CAG-217 sp000436335, 9: Phascola c obac e ium aecium, 10: Ru henibac e ium lac a i o mans, 11: Aga hobac e ec alis, 12: Lachnospi a eligens A, 13: Phascola c obac e ium A succina u ens, 14: Me hanob e ibac e A smi hii, 15: Akke mansia muciniphila, 16: Campylobac e B sp905373215, 17: Akke mansia muciniphila, 18: Collinsella. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) I. Cal e e-To e e al.