Received: 29 November 2022 | Revised: 24 January 2023 | Accepted: 30 January 2023

DOI: 10.1002/ece3.9829

RESEARCH ARTICLE

Beyond bacteria: Reconstructing microorganism connections

and deciphering the predicted mutualisms in mammalian gut
metagenomes

Qinlong Dai1 | Jingjing Ding2 | Xinyuan Cui3 | Yudong Zhu1 | Hua Chen4 |
Lifeng Zhu5

1
Sichuan Liziping National Natural
Reserve, Shimian, China Abstract
2
Jiangsu Academy of Forest, Nanjing, Numerous gut microbial studies have focused on bacteria. However, archaea, viruses,
China
3
fungi, protists, and nematodes are also regular residents of the gut ecosystem. Little is
College of Life Science, Nanjing Normal
University, Nanjing, China known about the composition and potential interactions among these six kingdoms in
4
Mingke Biotechnology (Hangzhou) Co., the same samples. Here, we unraveled the complex connection among them using ap-
Ltd., Hangzhou, China
5
proximately 123 gut metagenomes from 42 mammalian species (including carnivores,
College of Pharmacy, Nanjing University
of Chinese Medicine, Nanjing, China omnivores, and herbivores). We observed high variation in bacterial and fungal fami-
lies and relatively low variation in archaea, viruses, protists, and nematodes. We found
Correspondence
Lifeng Zhu, School of Medicine and that some fungi in the mammalian intestine might come from environmental sources
Holistic Integrative Medicine, Nanjing (e.g., soil and dietary plants), and some might be native to the intestine (e.g., the oc-
University of Chinese Medicine, Nanjing
210023, China. currence of Neocallimastigomycetes). The Methanobacteriaceae and Plasmodiidae fami-
Email: zhulf2020@126.com lies (archaea and protozoa, respectively) were predominant in these metagenomes,
whereas Onchocercidae and Trichuridae were the two most common nematodes, and
Siphoviridae and Myoviridae the two most common virus families in these mammalian
gut metagenomes. Interestingly, most of the pairwise co-­occurrence patterns were
significantly positive among these six kingdoms, and significantly negative networks
mainly occurred between fungi and prokaryotes (both bacteria and archaea). Our
study revealed some inconvenient characteristics in the mammalian gut microorgan-
ism ecosystem: (1) the community formed by members of the analyzed kingdoms re-
flects the life history of the host and the potential threat posed by pathogenic protists
and nematodes in mammals; and (2) the networks suggest the existence of predicted
mutualism among members of these six kingdoms and of the predicted competition,
mainly among fungi and other kingdoms.

KEYWORDS
mammal metagenomes, microorganisms, putative interaction

Qinlong Dai, Jingjing Ding and Xinyuan Cui contributed equally to this work.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2023 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd.

Ecology and Evolution. 2023;13:e9829.  www.ecolevol.org | 1 of 11

https://doi.org/10.1002/ece3.9829
2 of 11 | DAI et al.
TA X O N O M Y C L A S S I F I C AT I O N
Ecosystem ecology
1 | I NTRO D U C TI O N
The gut microbiome plays an important role in host health and dis-
eases, and many studies have associated it with bacteria (prokary-
otes), which are the main microorganisms in the digestive system
(Huang et al., 2022; Ley et al., 2008a, 2008b; Wei et al., 2019;
Youngblut et al., 2019). Other microorganisms in the gut include ar-
chaea (prokaryotes), viruses, and eukaryotes, such as fungi, protists,
and nematodes (Richard & Sokol, 2019). The balance in the symbiotic
relationships that exist among the members of these kingdoms is
a key factor in maintaining the health of the host. Until now, most
gut microbial research has focused on the interaction between the
host and its gut bacteria. However, the putative networks existing
among the members of the previously mentioned kingdoms have
been poorly explored. The development of metagenomic sequenc-
ing has pushed forward microbial research, and an increasing num-
ber of studies have discussed the function of gut microorganisms,
especially that of bacteria (Qin et al., 2010; Streit & Schmitz, 2004;
Thomas et al., 2012; Zhu et al., 2011). Interestingly, metagenomes
enable researchers to simultaneously study the entire collection
of genomes from a mixed population of microorganisms. This is in-
valuable for understanding animal gut ecosystems. Therefore, we
can simultaneously investigate the co-­occurrence patterns among
archaea, bacteria, viruses, fungi, protists, and nematodes. This is a
neglected characteristic in studies on gut microorganisms.
Network-­based approaches have proven valuable for exploring
complex interactions in systems biology and are also widely used in
gut microbial studies (Layeghifard et al., 2017). We aimed at decipher-
ing the existing connections among gut microorganisms using mam-
malian gut metagenomes. To begin, we chose 123 gut metagenomes
from 42 mammalian species (including carnivores, omnivores, and
herbivores); most of these metagenomes (approximately 104) came
from our laboratory; therefore, they were obtained using the same
sequencing platform to decrease the sequencing bias (Appendix S1).
The details on the origin of the 123 metagenomes are as follows:
52 from giant pandas (9 from the Qinling Mountains [GPQIN, wild]
(Wu et al., 2017), 7 from the Qionglai Mountains [GPQIO, wild]
(Guo et al., 2019), 19 from the Xiaoxiangling Mountains [GPXXL,
wild] (Yao et al., 2021; Zhu, Yang, et al., 2018), 7 from the Chengdu
Breeding Center [GPCD, captive] (Zhang et al., 2018), 10 from the
Yaan Research Base of the Wolong Research Center [GPYA, cap-
tive]) (Guo et al., 2019), 6 from red pandas from the Xiaoxiangling
Mountains (RP, wild; Zhu, Yang, et al., 2018), 19 from meat-­eating
carnivorans (CA) (Zhu, Wu, et al., 2018), 10 from omnivorous carniv-
orans (OC) (Guo et al., 2018; Zhu, Wu, et al., 2018), 12 from herbi-
vores (HE) (Zhu, Wu, et al., 2018), and 24 from Yunnan snub-­nosed
monkeys (YSNM, wild) (Xia et al., 2022). We investigated the gut
microorganism composition regarding the previously mentioned
kingdoms and compared their beta diversities. Finally, we investi-
gated the interactions among such microorganisms.
2 | R E S U LT S A N D D I S CU S S I O N
2.1 | Family-­level identification among six
kingdoms across the different mammal groups
We retained the reads from the previously mentioned kingdoms (ar-
chaea, bacteria, viruses, fungi, protists, and nematodes) in the 123
metagenomes analyzed; we observed high variation in the bacterial
and fungal families and relatively low variation in archaea, viruses,
and protists-­nematodes (Figure 1). In the metagenomes, most of the
reads belonged to the bacterial kingdom (>90%, Figure 1c). As previ-
ously identified, the relative abundance of Pseudomonadaceae is high
in wild giant panda (GPXXL and GPQIO, but not GPQIN) and RP pop-
ulations, and this bacterial family is putatively involved in secondary
plant metabolites (e.g., dietary cyanide compounds) detoxification
(Hu et al., 2021; Wang et al., 2021; Zhu, Yang, et al., 2018). Captive
giant pandas and OC harbor a high proportion of Enterobacteriaceae
(Figure 1c), most probably due to their living environments (Huang
et al., 2021; Zhu, Wu, et al., 2011, 2018). HE and YSNM harbor a
high proportion of Prevotellaceae (Figure 1c), which might be related
to the high proportion of carbohydrates in their diet (Xia et al., 2022;
Zhu, Wu, et al., 2018).
Regarding fungi, the proportion of Mucoraceae was high in four
mammal groups (three captive groups: GPYA, GPCD, and HE, and
one wild group: GPQIN; Figure 1d). Mucoraceae are saprophytic
fungi; many common molds that are destructive to food products
belong to this family (Cavalcanti & Trufem, 2008). We speculated
that dietary plants (e.g., bamboo) might be the source of these
fungi. Wild mammals GPXXL, RP, and YSNM groups had a rela-
tively high proportion of Pyronemataceae compared with the cap-
tive mammal groups GPYA, GPCD, CA, and OC (Figure 1d). The
wild RP and GPXXL, which live in the same region, have a similar
abundance pattern: a relatively high proportion of fungi from the
Mrakiaceae family. Pyronemataceae comprises saprobic, ectomy-
corrhizal, bryosymbiotic, and parasitic species that occur in a broad
range of habitats (in soil, burnt ground, debris, wood, and dung and
inside living bryophytes, plants, and lichens; Hansen et al., 2013).
For example, the main diet of the YSNM is lichen. Thus, a high
proportion of Pyronemataceae in wild herbivorous carnivoran
(giant panda and RP) and monkey feces may be associated with di-
etary plants. OC harbored a high proportion of Sclerotiniaceae and
Ophiostomataceae (Figure 1d). Sclerotiniaceae is a family of fungi
in the order Helotiales, and many species in this family are plant
pathogens (Whetzel, 1945). The Ophiostomataceae family has a
DAI et al. | 3 of 11

(a) (b) (c)

(d) (e) (f)

F I G U R E 1 Relative abundance of microorganism from six different kingdoms in the gut of different mammals. (a) All six kingdoms;
(b) archaeal families; (c) bacterial families; (d) fungal families; (e) protist and nematode families (combined); (f) virus families. The 123
metagenomes used in these analyses belonged to 52 giant pandas (9 from the Qinling Mountains [GPQIN, wild] (Wu et al., 2017), 7 from
the Qionglai Mountains [GPQIO, wild] (Guo et al., 2019), 19 from the Xiaoxiangling Mountains [GPXXL, wild] (Yao et al., 2021; Zhu, Wu,
et al., 2018), 7 from the Chengdu Breeding Center [GPCD, captive] (Zhang et al., 2018), and 10 from the Yaan research base of the Wolong
Research Center [GPYA, captive]) (Guo et al., 2019); 6 red pandas from the Xiaoxiangling Mountains (RP, wild) (Zhu, Wu, et al., 2018);
19 meat-­eating carnivorans (CA) (Zhu, Yang, et al., 2018); 10 omnivorous carnivorans (OC) (Guo et al., 2018; Zhu, Wu, et al., 2018); 12
herbivores (HE) (Zhu, Wu, et al., 2018); and 24 Yunnan snub-­nosed monkeys (YSNM, wild) (Xia et al., 2022).

widespread distribution, and its members are pathogens of both
coniferous and deciduous trees (Solheim, 1986). Their presence
in OC mammals might be related to dietary plants as a source of
fungi. CAs have different dominant microorganism families com-
pared with other mammal groups. Therefore, bacterial and fungal
diversities are affected at the family level by captivity (i.e., the
living environment), given that captive and wild pandas harbor dif-
ferent core bacteria and fungi; by dietary plants, given that typical
plant pathogenic fungi may be found in the feces of some of the
mammals under study; and by the host itself, given that CA and OC
harbor different bacterial and fungal families. Anaerobic gut fungi
(e.g., class Neocallimastigomycetes) are important members of the
gut microbiome of herbivores (Swift et al., 2021). We also found that
some reads (mainly from the metagenomes from giant pandas, OC,
HE, and YSNM) matched Neocallimastigomycetes (Appendix S1);
most of them putatively belonged to Neocallimastix californiae and
Anaeromyces robustus, which live in the native rumen environment
and can degrade cellulose (Henske et al., 2018; Swift et al., 2019).
The relative abundance of the putative Neocallimastigomycetes in
HE or YSNM was significantly higher than that of other mammal
groups in this study (Appendix S1).
We do not know whether the fungi identified in this study by
metagenomic methods can live in the mammalian intestine because
we used genomic DNA extracted from feces and not culture meth-
ods. However, indirect evidence suggested that some of the fungi
found in the mammalian intestine might come from environmental
sources (e.g., soil and dietary plants), whereas some might be native
to the intestine (e.g., putative Neocallimastigomycetes).
4 of 11 | DAI et al.
Moreover, we observed that the microbial composition of the
studied mammal groups was relatively conserved at the family level
for archaea, protists, and nematodes compared with that for bac-
teria and fungi, whereas their alpha diversity was highly different.
Archaea Methanobacteriaceae (with a predominance of the genus
Methanobrevibacter) was the dominant family across the mam-
mal groups under study (except for GPCD), and the relative abun-
dance of Methanocorpusculaceae (with a predominance of the genus
Methanocorpusculum) was high in HE (Figure 1b, Appendix S1). The
relative abundance of the putative Neocallimastigomycetes in HE
was nearly significantly higher than that of other mammal groups in
this study (Appendix S1). The species within Methanobrevibacter are
strictly anaerobic archaea that produce methane, mostly through the
reduction of carbon dioxide via hydrogen. Most species live in the
intestines of animals, such as termites (Leadbetter & Breznak, 1996)
and humans (Samuel et al., 2007). Putatively, Methanobrevibacter
smithii, Methanobrevibacter olleyae, and Methanobrevibacter sp.
were identified in the samples, confirming the results of previous
studies (Leadbetter & Breznak, 1996; Samuel et al., 2007). Among
the mammal groups, herbivores (including HE and YSNM) had the
high alpha diversity (Shannon index), especially in Archaea and
Bacteria (Figure 2), which was similar to the previous studies (Ley
et al., 2008a, 2008b; Zhu, Wu, et al., 2018). The complexity of the
dietary diversity might partially result in this pattern.
Regarding the Protista and Nematoda kingdoms, Protozoa
Plasmodiidae (mainly represented by the genus Plasmodium) was the
domain family in our metagenomes (Figure 1e). Plasmodium is a para-
site that infects humans, primates, rodents, birds, and reptiles and is
distributed worldwide (Aravind et al., 2003). Onchocercidae (mainly
represented by the genera Brugia, Onchocerca, and Wuchereri) and
Trichuridae (mainly represented by Trichuris) are the two most com-
mon Nematoda families in our mammalian metagenomes. Among
Trichuris, most reads were putatively mapped to Trichuris trichiura,
which is one of the main parasites in humans (especially children)
(Stephenson et al., 2000). Regarding Brugia, some reads were puta-
tively mapped to Brugia malayi, a human filarial parasite that threat-
ens human health (Ghedin et al., 2007). Although we cannot confirm
that the previously mentioned reads in our metagenomes belong to
Plasmodium, T. trichiura, and B. malayi, we can conclude that these
parasite species are our reads' closest relatives according to the cur-
rent NCBI micro-­NR database. Therefore, regardless of the living
conditions (captive or wild habitat) or the host species, most of the
mammals included in this study face the potential threat of being
infected by parasitic protists and nematodes.
In addition, Siphoviridae (including Jerseyvirus, Kagunavirus,
Dhillonvirus, Lambdavirus, and Seuratvirus) and Myoviridae (in-
cluding Dhakavirus, Asteriusvirus, Gaprivervirus, Kanagawavirus,
Krischvirus, Phapecoctavirus, and Felixounavirus) are the two most
common bacterial virus (phages) families in our mammal gut metag-
enomes (Figure 1f). Siphoviridae and Myoviridae can infect strains
of Enterobacteriaceae, Pseudomonadaceae, and Lactobacillaceae
(Brüssow & Desiere, 2001; Lavigne et al., 2009). Despite the high
variation in the bacterial families present in the mammal groups
included in this study, large broad-­spectrum polyvalent infections
(including both gram-­negative and gram-­positive bacteria) resulted in
a conserved bacteriophage pattern at the family level in our metag-
enomes. However, the virus composition partially shows the specific
features of the mammal group. For example, the relative abundance
of Ackermannviridae and Demerecviridae was high in GPYA, while OC
harbored a high proportion of Guelinviridae, GPXXL harbored a high
proportion of Schitoviridae, and CA harbored the highest proportion
of Podoviridae (Figure 1f).
We speculated that host-­specific patterns of microorganisms
from the six kingdoms included in this study could be determined at
the species or strain level. Results from NMDS analyses using Bray–­
Curtis distance of the putative species-­level abundance support our
hypothesis that mammal groups have significantly different mean
abundance dissimilarities regarding microorganisms from the six
kingdoms under study (Figure 3a, Figure 2f; Adonis test: p = .001).
The highest variation in distances was observed among bacteria,
whereas the lowest was observed among viruses in all mammal
groups (Figure 2; R 2 of bacteria: 43.7%; R 2 of fungi: 30.1%; R 2 of
protists-­nematodes: 29.5%; R 2 of archaea: 27.7%; R 2 of virus: 21.6%).
Beta analyses at the putative species level further confirmed the rel-
atively high variations in bacterial and fungal kingdoms compared
with those in archaea, protists-­nematodes, and viruses.
Overall, at the family level, core microorganisms of the six king-
doms under study reflected the effects of the life history or ecolog-
ical niches of the host. The variation in the abundance of bacteria
and fungi among mammal groups indicates the potential effects of
diet, host phylogeny, and living conditions. The other four kingdoms
(i.e., archaea, viruses, protists, and nematodes) show three kinds of
“broad-­spectrum polyvalent” characteristics. Archaea in the mam-
malian intestine are strictly anaerobic and reduce carbon dioxide
via the hydrogen produced by the fermentation of carbohydrates by
bacteria. Phages in the mammalian gut can infect both gram-­negative
and gram-­positive bacteria. These characteristics mimic the varia-
tion in the abundance of specific bacterial groups among mammals.
Furthermore, although the parasitic protist and nematode species
found in mammals might be different, the close phylogenetic position
at the family level and potential capability for cross-­species infection
among different hosts lead to similar core parasite families among the
mammal groups, regardless of their life history differences. Parasites
are more dependent on the host than symbiotic bacteria. Thus, the
“broad-­spectrum polyvalent” characteristics described here may
have led to “convergence” at the family level in archaea, viruses, pro-
tists, and nematodes. However, the potential interactions among the
six microorganism kingdoms require further exploration.
2.2 | Co-­occurrence networks among the six
microorganism kingdoms reflecting the predicted
competition
We next focused on transkingdom interactions among the six king-
doms in the mammal metagenomes through the lens of specific life
DAI et al. | 5 of 11

Shannon Archaea(p=1e-15) Protist+Nematoda(p=4.24e-4)

Viruses(p=4.19e-14)
Bacteria(p=5.15e-12)
Shannon

RP
N

CD

OC

HE

NM
YA

CA
Fungi(p=2.68e-7) XX
QI

QI

GP

GP

YS
GP

GP

GP
Shannon

RP
N

YA

CD

OC

HE
CA

NM
XX
QI

QI

GP

GP

YS
GP

GP

GP

F I G U R E 2 Shannon index of each microorganism kingdom among these 10 mammal groups.

histories or ecological niches. Interestingly, most of the pairwise co-­
occurrence patterns were significantly positive among the kingdoms,
and significantly negative networks mainly occurred between fungi
and prokaryotes (both bacteria and archaea; Appendix S1). Most of
the interactions between archaea genera and members of the other
kingdoms (e.g., bacteria) were significantly positive among mammal
groups, including captive giant pandas (GPC; combined GPYA and
GPCD), wild giant pandas (GPW; combined GPQIN, GPQIO, and
GPXXL), YSNM, CA, and HE (Figures 4 and 5; Appendix S1). Virus
families displayed a similar pattern (Appendix S1). These findings indi-
cate predicted mutualisms between bacteria and archaea or viruses.
Archaea (e.g., Methanobacteriales, M. smithii, and Methanosphaera
stadtmanae) are naturally occurring components of the animal gut
microbiota, and they participate in functions such as methanogen-
esis, transformation of heavy metals, and trimethylamine metabo-
lism (Brugère et al., 2014; Leadbetter & Breznak, 1996; Samuel
et al., 2007). Thus, the complex network across different mammalian
groups investigates the predicted mutualism between archaea and
bacteria in the mammalian gut ecosystem. Bacteriophages are the
predominant viruses in the human gut microbiome, and they shape
the microbial composition and drive bacterial diversity and nutrient
turnover through continuous cycles of predation and coevolution
6 of 11 | DAI et al.

(a) Five kingdoms NMDS(stress:0.151) (b) Archaea NMDS(stress:0.181) (c) Bacteria NMDS(stress:0.15)
GPQIN GPQIN GPQIN
GPQIO GPQIO GPQIO
GPXXL GPXXL GPXXL
GPYA GPYA GPYA
GPCD GPCD GPCD
RP RP RP
CA CA CA
OC OC OC
HE HE HE
YSNM YSNM YSNM

(d) Fungi NMDS(stress:0.148) (e) Protist+Nematoda NMDS(stress:0.232) (f) Viruses NMDS(stress:0.222)

GPQIN GPQIN GPQIN
GPQIO GPQIO GPQIO
GPXXL GPXXL GPXXL
GPYA GPYA GPYA
GPCD GPCD GPCD
RP RP RP
CA CA CA
OC OC OC
HE HE HE
YSNM YSNM YSNM

nmds1 nmds1 nmds1

F I G U R E 3 NMDS ordination analysis among different mammal groups based on Bray–­Curtis distance matrices using the TPM of six
microorganism kingdoms. NMDS: (a) using all kingdoms at the species level; (b) using archaea at the species level; (c) using bacteria at the
species level; (d) using fungi at the species level; (e) using protists and nematodes at the species level (in combination); (f) using viruses at the
species level.

(Sutton & Hill, 2019). Maintaining a dynamic balance between bac-
teriophages and bacteria is beneficial. Here, we also revealed that
DNA viruses (e.g., phages) show some kind of interdependence with
bacteria across different mammalian groups. Therefore, in this study,
we provide evidence of potential coevolution between bacteria and
archaea or bacteriophages in the mammalian gut ecosystem.
However, we found that some fungi genera (e.g., some
Ascomycota genera) showed the predicted competition mainly with
some bacterial genera (e.g., from Proteobacteria and Firmicutes) in
six mammal groups (including GPC, YSNM, RP, CA, OC, and HE;
Figures 4 and 5; Appendix S1), but the specific fungi genera were
different in each mammalian group (Appendix S1). For example,
in GPC, Beauveria, Bipolaris, Corynespora, and Trematosphaeria
were the primary fungal genera showing a significant negative
relationship with some Proteobacteria and Actinobacteria gen-
era (Figure 4b and Appendix S1). In YSNMs, the fungal genus
Phialophora mainly showed a significant negative relationship with
some Proteobacteria genera (Figure 4c). In the RP group, some
fungal genera showed a significant negative relationship with
some genera of bacteria, archaea, and nematodes (Figure 4d and
Appendix S1). It should be noted that the sample size from the
RP was small (including six samples only), which might lead to bi-
ases in co-­o ccurrence analyses. In CA, Polyporus and Xylona were
the main fungal genera showing a significant negative relationship
with some genera from Firmicutes (Figure 5a and Appendix S1). In
OC, Ascochyta, Monilinia, and Tremella were the main fungal genera
showing a significant negative relationship with some genera from
Proteobacteria and Firmicutes (Figure 5b and Appendix S1). The in-
teraction between fungi and bacteria in the gut is complicated,
and most of them show a friendly relationship (Sam et al., 2017),
as found in our study (significant positive: most co-­o ccurrence
networks). A healthy animal with an intact microbiome can be re-
sistant to pathogenic fungi colonization such as that by Candida al-
bicans (Van der Waaij et al., 1971). Interestingly, Lactobacillus and
C. albicans may antagonistically compete in the murine digest sys-
tem (Savage, 1969). Anaerobic gut fungi are important members of
the gut microbiome of herbivores, existing in small numbers rela-
tive to bacteria, and can degrade cellulose and produce a wealth
of secondary metabolites (e.g., they may show antibiotic activ-
ity) that may regulate the gut microbiome (Henske et al., 2018;
Swift et al., 2019). Therefore, we deduced that the small amount
of the predicted competition between fungi and bacteria in the
DAI et al. | 7 of 11

(a) (b)

GPW GPC

(c) (d)

RP
Correlation types Significantly positive
Significantly negative

YSNM

F I G U R E 4 Complex interactions among the six microorganism kingdoms studied in specific mammal populations. (a) GPW group
(combining GPQIN, GPQIO, and GPXXL; Spearman: r ≥ .8, p ≤ .01); (b) GPC group (combining GPYA and GPCD; Spearman: r ≥ .8, p ≤ .01);
(c) YSNM group (Spearman: r ≥ .8, p ≤ .01); (d) RP (Spearman: r ≥ .9, p ≤ .01). Green lines represent significantly positive correlations. Red lines
represent significantly negative correlations. Each point represents a single genus from the indicated family belonging to one of the six
kingdoms analyzed (except for viruses for which families are represented).

mammalian gut ecosystem might be caused by niche overlap (com-
petition for diet) and antifungal/antibiotic metabolites.
In addition, we found that many protist and nematode genera
showed a significantly positive relationship with many other gen-
era, including bacteria, archaea, and fungi, in GPW, GPC, YSNM,
CA, OC, and HE (Figures 4 and 5; Appendix S1). Nonhuman primate
gut-­associated protists and nematodes are weakly structured by
primate phylogeny, with a minimal signal from the diet, which in-
dicates that gut-­associated eukaryotes offer different information
than gut-­associated bacteria (Mann et al., 2020). We also found
“convergence” at the family level between protists and nematodes
in mammal groups, including carnivores, omnivores, and herbivores,
regardless of their diet type or the host species.
Gut eukaryotes modulate other microbes through predation,
resource and niche competition, and interactions with the host im-
mune system (Lukeš et al., 2015; Stensvold & van der Giezen, 2018).
Bacteria are one of the main components of the diets of protists and
bacteria (Fu et al., 2005; Saleem et al., 2013). Here, except in the
RP group (which may have a potential bias due to the small sample
size), complex networks suggest putative mutualistic relationships
between mammalian gut eukaryotes (fungi, protists, and nematodes)
and prokaryotes (archaea and bacteria). However, parasitic protists
(Plasmodium) and nematodes (Onchocercidae and Trichuridae) may
differ in their interactions with the host and have different roles
in the intestine compared with those of fungi, most of which may
play a role in dietary component digestion. In addition, some fungi
can prey on nematodes (Dijksterhuis et al., 1994). We found that,
in RP, Onchocerca spp. had a significantly negative relationship with
some fungi genera, including Rasamsonia, Emmonsia, Chaetomium,
Hysterangium, Rhodotorula, Puccinia, Phaffia, Naganishia, and
Kwoniella. In GPC, Kytococcus (Actinobacteria) showed a significant
negative correlation with Trichuris. In OC, Trichuris showed a significant
negative correlation with some Proteobacteria (e.g., Mesosutterella
and Rhizobacter). In HE, the protist parasite Hepatocystis showed a
8 of 11 | DAI et al.

(a)
(b)

CA HE
Correlation types Significantly positive
Significantly negative

(c)

OC

F I G U R E 5 Complex interactions among the six microorganism kingdoms studied in specific mammal populations. (a) CA group (Spearman
r ≥ .8, p ≤ .01); (b) HE group (Spearman r ≥ .9, p ≤ .01); (c) OC (Spearman r ≥ .9, p ≤ .01). Green lines represent significantly positive correlations.
Red lines represent significantly negative correlations. Each point represents a single genus from the indicated family belonging to one of the
six kingdoms analyzed (except for viruses for which families are represented).

significantly negative relationship with some prokaryotes, such as
Methanoculleus, Duncaniella, Hornefia, Calorimonas, Breznakia, and
Gilliamella. Hepatocystis (within the family Plasmodiidae) parasites
infect monkeys, bats, squirrels, and ungulates in Africa, Asia, and
Australia (Ejotre et al., 2021). One study reported coinfection and
cross-­species transmission of divergent Hepatocystis lineages in a
wild African primate community (Thurber et al., 2013). This protist
lives with other microorganisms and can predate and kill bacteria in
the environment; however, bacteria are also able to resist this ac-
tion, resulting in profound changes to the protist lifestyle (Henriquez
et al., 2021). Thus, we deduced that some prokaryotes might have a
potential antiprotist role in the mammalian gut ecosystem.
Here, the reconstruction of mammalian gut microorganism in-
teractions (the predicted mutualism patterns) provides a basic but
fresh view of mammalian gut ecosystems, but this does not indicate
that parasitic fungi, protists, and nematodes are an important part of
the gut ecosystem. In this study, we cannot answer the question of
whether these potential parasites are key to maintaining the balance
of gut microorganisms.
(including that of organisms from different kingdoms) extracted di-
rectly from environmental samples. Implying that we could explore
putative interactions among different kingdom organisms in the
intestine. Here, we conducted a primary study and revealed some
basic features of the mammalian gut ecosystem. We observed high
variation in the bacterial and fungal families and relatively low vari-
ation (convergence pattern) in those of archaea, viruses, and protist-­
nematodes, which reflects the effects of life history. Although we
know the limitations of metagenomic analyses in species identi-
fication, mainly due to incomplete or short contigs obtained from
complex genetic material, we found a potential threat by parasitic
protists and nematodes in the studied mammals (in both wild and
captive populations). The complex networks among the six microor-
ganism kingdoms studied (archaea, bacteria, viruses, fungi, protists,
and nematodes) have deciphered the predicted mutualism in the gut
ecosystem. However, we still need to understand the potential com-
petition among gut-­associated eukaryotes, prokaryotes, and viruses.
4 | M ATE R I A L S A N D M E TH O DS

3 | CO N C LU S I O N 4.1 | Data used

Metagenomic technology has made considerable progress in gut
microbial research. However, most researchers have subconsciously
focused on the function of bacteria when using metagenomes.
Therefore, we neglected the core advantages of this method: poten-
tially recovering and completing the sequencing of genetic material
We used raw datasets of 123 gut metagenomes from 42 mammalian
species (including carnivores, omnivores, and herbivores); most of
these metagenomes (approximately 104 metagenomes) came from
our laboratory using the same sequencing platform to decrease
sequencing bias (Appendix S1). From these 123 metagenomes, 52
DAI et al.
belonged to giant pandas (9 GPQIN (Wu et al., 2017), 7 GPQIO (Guo
et al., 2019), 19 GPXXL (Yao et al., 2021; Zhu, Yang, et al., 2018),
7 GPCD (Zhang et al., 2018), and 10 GPYA) (Guo et al., 2019), 6 to
RP, 19 to CA (Zhu, Wu, et al., 2018), 10 to OC (Guo et al., 2018; Zhu,
Wu, et al., 2018), 12 to HE (Zhu, Wu, et al., 2018), and 24 to YSNM
(Xia et al., 2022).
4.2 | Metagenomic analyses
The raw reads of the 123 metagenomes were trimmed using
Trimmomatic (Bolger et al., 2014) to remove all reads <50 bp in length
and reads with degenerate bases (N's). All duplicates were defined
as sequences in which the initial 20 nucleotides were identical and
shared an overall identity of more than 97% throughout the length of
the shortest read. Megahit (Li et al., 2015) was used to assemble the
clean reads, and Salmon was used for quality control of the contigs
and to remove contigs with coverage below 60% (Patro et al., 2017).
We used BWA (Li, 2013) to clean the host genome and delete po-
tential host contaminants. All coding regions (CDS) of metagenomic
contigs were predicted by Prodigal (Hyatt et al., 2012) and clustered
using CD-­HIT (identity: 95%; overlap: 90%) (Fu et al., 2012), generat-
ing unigene pools. Salmon (Patro et al., 2017) was used to map the
clean reads to each unigene pool and calculate the transcripts per
million (TPM) to determine unigene abundance. Diamond (Buchfink
et al., 2015) was used to conduct the alignment of unigenes against
the NCBI micro-­NR database (including bacteria, fungi, archaea, vi-
ruses, protists, and nematodes) and obtained the TPM of each taxon
group using our own software.
4.3 | Alpha and beta diversity analysis
The TPM of each taxon group was used to calculate Shannon's index.
We applied NMDS (nonmetric multidimensional scaling) ordination
and the Adonis test in the VEGAN package (Dixon, 2003) based on
Bray–­Curtis dissimilarity matrices (Beals, 1984) using the TPMs of
the taxon groups.
4.4 | Reconstructing the interactions among
microorganism kingdoms
We used our own Perl commands (Appendix S1) to calculate the
Spearman correlation (Myers & Sirois, 2004) based on the TPM of
the genera of the microorganism kingdoms, and the results were
input to Cytoscape (Shannon et al., 2003) for visualization. Here,
we used the family level for this calculation to decrease bias in the
taxonomic assignment. Overall, we set up the following strict crite-
ria for interaction visualization analysis: (1) all the genera (family in
viruses) should be identified in six or more samples (the RP group
has six samples only); (2) Spearman correlation r ≥ .8 (0.9 for HE, OC,
RP, and YSNM due to excessive relationships); (3) p ≤ .01; and (4) the
| 9 of 11
interactions among kingdoms were kept, whereas the relationships
within each kingdom were excluded.
AU T H O R C O N T R I B U T I O N S
Qinlong Dai: Formal analysis (equal); visualization (equal); writ-
ing –­ original draft (equal). Jingjing Ding: Visualization (equal); writ-
ing –­ original draft (equal). Xinyuan Cui: Formal analysis (equal);
visualization (equal); writing –­ original draft (equal). Yudong Zhu:
Formal analysis (equal); visualization (equal); writing –­ original draft
(equal). Hua Chen: Visualization (equal). Lifeng Zhu: Conceptualization
(lead); data curation (lead); formal analysis (lead); methodology (lead);
resources (equal); visualization (lead); writing –­ original draft (lead).
AC K N OW L E D G M E N T S
This study is supported by the project of the ecological research of
the translocated Giant in Liziping National Nature Reserve, Forestry
Science and Technology Innovation and Promotion Project of
Jiangsu Province (Jiangsu Province Forest and Wetland Positioning
Monitoring Long-­term Scientific Research Base (LYKJ [2020] 21)),
and the basic research project of Sichuan Academy of Giant panda.
C O N FL I C T O F I N T E R E S T S TAT E M E N T
The authors declared no conflicts of interest relevant to this
manuscript.
DATA AVA I L A B I L I T Y S TAT E M E N T
This is not applicable for this study. The data used in this study come
from the published data from several research groups. The personal
script used in this study has been included in the Appendix S1.
ORCID
Lifeng Zhu https://orcid.org/0000-0002-9231-2786
REFERENCES
Aravind, L., Iyer, L. M., Wellems, T. E., & Miller, L. H. (2003). Plasmodium
biology: Genomic gleanings. Cell, 115, 771–­785.
Beals, E. W. (1984). Bray-­Curtis ordination: An effective strategy for
analysis of multivariate ecological data. In Advances in ecological re-
search (pp. 1–­55). Elsevier.
Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: A flexible
trimmer for Illumina sequence data. Bioinformatics, 30, 2114–­2120.
Brugère, J. F., Borrel, G., Gaci, N., Tottey, W., O'toole, P. W., & Malpuech-­
Brugère, C. (2014). Archaebiotics: Proposed therapeutic use of ar-
chaea to prevent trimethylaminuria and cardiovascular disease. Gut
Microbes, 5(1), 5–­10.
Brüssow, H., & Desiere, F. (2001). Comparative phage genomics and the
evolution of Siphoviridae: Insights from dairy phages. Molecular
Microbiology, 39, 213–­223.
Buchfink, B., Xie, C., & Huson, D. H. (2015). Fast and sensitive protein
alignment using DIAMOND. Nature Methods, 12, 59–­60.
Cavalcanti, M., & Trufem, S. F. B. (2008). Pilobolus (Mucoraceae) from
herbivore dung in Recife, Pernambuco, Brazil. Mycotaxon, 104,
111–­122.
Dijksterhuis, J., Veenhuis, M., Harder, W., & Nordbring-­Hertz, B. (1994).
Nematophagous fungi: Physiological aspects and structure–­
function relationships. Advances in Microbial Physiology, 36,
111–­143.
10 of 11 | DAI et al.
Dixon, P. M. (2003). VEGAN, a package of R functions for community
ecology. Journal of Vegetation Science, 14, 927–­930.
Ejotre, I., Reeder, D. M., Matuschewski, K., & Schaer, J. (2021).
Hepatocystis. Trends in Parasitology, 37, 456–­457.
Fu, L., Niu, B., Zhu, Z., Wu, S., & Li, W. (2012). CD-­HIT: Accelerated for
clustering the next-­generation sequencing data. Bioinformatics, 28,
3150–­3152.
Fu, S., Ferris, H., Brown, D., & Plant, R. (2005). Does the positive feed-
back effect of nematodes on the biomass and activity of their bac-
teria prey vary with nematode species and population size? Soil
Biology and Biochemistry, 37, 1979–­1987.
Ghedin, E., Wang, S., Spiro, D., Caler, E., Zhao, Q., Crabtree, J., Allen, J.
E., Delcher, A. L., Guiliano, D. B., Miranda-­Saavedra, D., Angiuoli,
S. V., Creasy, T., Amedeo, P., Haas, B., el-­Sayed, N. M., Wortman, J.
R., Feldblyum, T., Tallon, L., Schatz, M., … Scott, A. L. (2007). Draft
genome of the filarial nematode parasite Brugia malayi. Science, 317,
1756–­1760.
Guo, W., Mishra, S., Wang, C., Zhang, H., Ning, R., Kong, F., Zeng, B.,
Zhao, J., & Li, Y. (2019). Comparative study of gut microbiota in wild
and captive giant pandas (Ailuropoda melanoleuca). Genes, 10, 827.
Guo, W., Mishra, S., Zhao, J., Tang, J., Zeng, B., Kong, F., Ning, R., Li, M.,
Zhang, H., Zeng, Y., Tian, Y., Zhong, Y., Luo, H., Liu, Y., Yang, J., Yang,
M., Zhang, M., Li, Y., Ni, Q., … Li, Y. (2018). Metagenomic study sug-
gests that the gut microbiota of the giant panda (Ailuropoda melan-
oleuca) may not be specialized for fiber fermentation. Frontiers in
Microbiology, 9, 229.
Hansen, K., Perry, B. A., Dranginis, A. W., & Pfister, D. H. (2013). A phy-
logeny of the highly diverse cup-­fungus family Pyronemataceae
(Pezizomycetes, Ascomycota) clarifies relationships and evolution
of selected life history traits. Molecular Phylogenetics and Evolution,
67, 311–­335.
Henriquez, F. L., Mooney, R., Bandel, T., Giammarini, E., Zeroual, M., Fiori,
P. L., Margarita, V., Rappelli, P., & Dessì, D. (2021). Paradigms of
protist/bacteria symbioses affecting human health: Acanthamoeba
species and trichomonas vaginalis. Frontiers in Microbiology, 11,
3380.
Henske, J. K., Gilmore, S. P., Haitjema, C. H., Solomon, K. V., & O'Malley,
M. A. (2018). Biomass-­degrading enzymes are catabolite repressed
in anaerobic gut fungi. AICHE Journal, 64, 4263–­4270.
Hu, T., Dai, Q., Chen, H., Zhang, Z., Dai, Q., Gu, X., Yang, X., Yang, Z., &
Zhu, L. (2021). Geographic pattern of antibiotic resistance genes in
the metagenomes of the giant panda. Microbial Biotechnology, 14,
186–­197.
Huang, G., Wang, L., Li, J., Hou, R., Wang, M., Wang, Z., Qu, Q., Zhou, W.,
Nie, Y., Hu, Y., Ma, Y., Yan, L., Wei, H., & Wei, F. (2022). Seasonal
shift of the gut microbiome synchronizes host peripheral circadian
rhythm for physiological adaptation to a low-­fat diet in the giant
panda. Cell Reports, 38, 110203.
Huang, G., Wang, X., Hu, Y., Wu, Q., Nie, Y., Dong, J., Ding, Y., Yan, L., &
Wei, F. (2021). Diet drives convergent evolution of gut microbiomes
in bamboo-­eating species. Science China Life Sciences, 64, 88–­95.
Hyatt, D., LoCascio, P. F., Hauser, L. J., & Uberbacher, E. C. (2012). Gene
and translation initiation site prediction in metagenomic sequences.
Bioinformatics, 28, 2223–­2230.
Lavigne, R., Darius, P., Summer, E. J., Seto, D., Mahadevan, P., Nilsson, A.
S., Ackermann, H. W., & Kropinski, A. M. (2009). Classification of
Myoviridae bacteriophages using protein sequence similarity. BMC
Microbiology, 9, 1–­16.
Layeghifard, M., Hwang, D. M., & Guttman, D. S. (2017). Disentangling
interactions in the microbiome: A network perspective. Trends in
Microbiology, 25, 217–­228.
Leadbetter, J. R., & Breznak, J. A. (1996). Physiological ecology of
Methanobrevibacter cuticularis sp. nov. and Methanobrevibacter
curvatus sp. nov., isolated from the hindgut of the termite
Reticulitermes flavipes. Applied and Environmental Microbiology, 62,
3620–­3631.
Ley, R. E., Hamady, M., Lozupone, C., Turnbaugh, P. J., Ramey, R. R.,
Bircher, J. S., Schlegel, M. L., Tucker, T. A., Schrenzel, M. D., Knight,
R., & Gordon, J. I. (2008). Evolution of mammals and their gut mi-
crobes. Science, 320, 1647–­1651.
Ley, R. E., Lozupone, C. A., Hamady, M., Knight, R., & Gordon, J. I. (2008).
Worlds within worlds: Evolution of the vertebrate gut microbiota.
Nature Reviews Microbiology, 6, 776–­788.
Li, D., Liu, C. M., Luo, R., Sadakane, K., & Lam, T. W. (2015). MEGAHIT:
An ultra-­fast single-­node solution for large and complex metage-
nomics assembly via succinct de Bruijn graph. Bioinformatics, 31,
1674–­1676.
Li, H. (2013). Aligning sequence reads, clone sequences and assembly
contigs with BWA-­MEM. arXiv Preprint arXiv: 13033997.
Lukeš, J., Stensvold, C. R., Jirků-­Pomajbíková, K., & Wegener Parfrey, L.
(2015). Are human intestinal eukaryotes beneficial or commensals?
PLoS Pathogens, 11, e1005039.
Mann, A. E., Mazel, F., Lemay, M. A., Morien, E., Billy, V., Kowalewski,
M., di Fiore, A., Link, A., Goldberg, T. L., Tecot, S., Baden, A. L.,
Gomez, A., Sauther, M. L., Cuozzo, F. P., Rice, G. A. O., Dominy, N.
J., Stumpf, R., Lewis, R. J., Swedell, L., … Wegener Parfrey, L. (2020).
Biodiversity of protists and nematodes in the wild nonhuman pri-
mate gut. The ISME Journal, 14, 609–­622.
Myers, L., & Sirois, M. J. (2004). Spearman correlation coefficients, dif-
ferences between. Encyclopedia of Statistical Sciences, 12, 1–­3.
Patro, R., Duggal, G., Love, M. I., Irizarry, R. A., & Kingsford, C. (2017).
Salmon provides fast and bias-­aware quantification of transcript
expression. Nature Methods, 14, 417–­419.
Qin, J., Li, R., Raes, J., Arumugam, M., Burgdorf, K. S., Manichanh, C.,
Nielsen, T., Pons, N., Levenez, F., Yamada, T., Mende, D. R., Li, J.,
Xu, J., Li, S., Li, D., Cao, J., Wang, B., Liang, H., Zheng, H., … Wang,
J. (2010). A human gut microbial gene catalogue established by
metagenomic sequencing. Nature, 464, 59–­65.
Richard, M. L., & Sokol, H. (2019). The gut mycobiota: Insights into
analysis, environmental interactions and role in gastrointesti-
nal diseases. Nature Reviews Gastroenterology & Hepatology, 16,
331–­3 45.
Saleem, M., Fetzer, I., Harms, H., & Chatzinotas, A. (2013). Diversity of
protists and bacteria determines predation performance and stabil-
ity. The ISME Journal, 7, 1912–­1921.
Sam, Q. H., Chang, M. W., & Chai, L. Y. A. (2017). The fungal mycobi-
ome and its interaction with gut bacteria in the host. International
Journal of Molecular Sciences, 18, 330.
Samuel, B. S., Hansen, E. E., Manchester, J. K., Coutinho, P. M., Henrissat,
B., Fulton, R., Latreille, P., Kim, K., Wilson, R. K., & Gordon, J. I.
(2007). Genomic and metabolic adaptations of Methanobrevibacter
smithii to the human gut. Proceedings of the National Academy of
Sciences, 104, 10643–­10648.
Savage, D. C. (1969). Microbial interference between indigenous yeast
and lactobacilli in the rodent stomach. Journal of Bacteriology, 98,
1278–­1283.
Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D.,
Amin, N., Schwikowski, B., & Ideker, T. (2003). Cytoscape: A soft-
ware environment for integrated models of biomolecular interac-
tion networks. Genome Research, 13, 2498–­2504.
Solheim, H. (1986). Species of Ophiostomataceae isolated from Picea
abies infested by the bark beetle Ips typographic. Nordic Journal of
Botany, 6, 199–­207.
Stensvold, C. R., & van der Giezen, M. (2018). Associations between
gut microbiota and common luminal intestinal parasites. Trends in
Parasitology, 34, 369–­377.
Stephenson, L., Holland, C., & Cooper, E. (2000). The public health signif-
icance of Trichuris trichiura. Parasitology, 121, S73–­S95.
Streit, W. R., & Schmitz, R. A. (2004). Metagenomics–­the key to the un-
cultured microbes. Current Opinion in Microbiology, 7, 492–­498.
Sutton, T. D., & Hill, C. (2019). Gut bacteriophage: Current understanding
and challenges. Frontiers in Endocrinology, 10, 784.
DAI et al.
Swift, C. L., Brown, J. L., Seppälä, S., & O'Malley, M. A. (2019). Co-­
cultivation of the anaerobic fungus Anaeromyces robustus with
Methanobacterium bryantii enhances transcription of carbo-
hydrate active enzymes. Journal of Industrial Microbiology and
Biotechnology, 46, 1427–­1433.
Swift, C. L., Louie, K. B., Bowen, B. P., Olson, H. M., Purvine, S. O.,
Salamov, A., Mondo, S. J., Solomon, K. V., Wright, A. T., Northen, T.
R., Grigoriev, I. V., Keller, N. P., & O’Malley, M. A. (2021). Anaerobic
Gut Fungi Are an Untapped Reservoir of Natural Products.
Proceedings of the National Academy of Sciences, 118, e2019855118.
Thomas, T., Gilbert, J., & Meyer, F. (2012). Metagenomics-­a guide from
sampling to data analysis. Microbial Informatics and Experimentation,
2, 1–­12.
Thurber, M. I., Ghai, R. R., Hyeroba, D., Weny, G., Tumukunde, A.,
Chapman, C. A., Wiseman, R. W., Dinis, J., Steeil, J., Greiner, E.
C., Friedrich, T. C., O'Connor, D. H., & Goldberg, T. L. (2013). Co-­
infection and cross-­species transmission of divergent Hepatocystis
lineages in a wild African primate community. International Journal
for Parasitology, 43, 613–­619.
Van der Waaij, D., Berghuis-­de Vries, J., & Lekkerkerk-­Van der Wees, J.
(1971). Colonization resistance of the digestive tract in conventional
and antibiotic-­treated mice. Epidemiology & Infection, 69, 405–­411.
Wang, L., Huang, G., Hou, R., Qi, D., Wu, Q., Nie, Y., Zuo, Z., Ma, R., Zhou,
W., Ma, Y., Hu, Y., Yang, Z., Yan, L., & Wei, F. (2021). Multi-­omics
reveals the positive leverage of plant secondary metabolites on the
gut microbiota in a non-­model mammal. Microbiome, 9, 1–­15.
Wei, F., Wu, Q., Hu, Y., Huang, G., Nie, Y., & Yan, L. (2019). Conservation
metagenomics: A new branch of conservation biology. Science
China Life Sciences, 62, 168–­178.
Whetzel, H. H. (1945). A synopsis of the genera and species of the
Sclerotiniaceae, a family of stromatic inoperculate discomycetes.
Mycologia, 37, 648–­714.
Wu, Q., Wang, X., Ding, Y., Hu, Y., Nie, Y., Wei, W., Ma, S., Yan, L., Zhu, L.,
& Wei, F. (2017). Seasonal variation in nutrient utilization shapes gut
microbiome structure and function in wild giant pandas. Proceedings
of the Royal Society B: Biological Sciences, 284, 20170955.
Xia, W., Liu, G., Wang, D., Chen, H., Zhu, L., & Li, D. (2022). Functional
convergence of Yunnan snub-­nosed monkey and bamboo-­eating
panda gut microbiomes revealing the driving by dietary flexi-
bility on mammal gut microbiome. Computational and structural
Biotechnology Journal, 20, 685–­699.
| 11 of 11
Yao, R., Dai, Q., Wu, T., Yang, Z., Chen, H., Liu, G., Zhu, Y., Qi, D., Yang,
X., Luo, W., Gu, X., Yang, X., & Zhu, L. (2021). Fly-­over phylogeny
across invertebrate to vertebrate: The giant panda and insects
share a highly similar gut microbiota. Computational and Structural
Biotechnology Journal, 19, 4676–­4683.
Youngblut, N. D., Reischer, G. H., Walters, W., Schuster, N., Walzer, C.,
Stalder, G., Ley, R. E., & Farnleitner, A. H. (2019). Host diet and evo-
lutionary history explain different aspects of gut microbiome di-
versity among vertebrate clades. Nature Communications, 10, 1–­15.
Zhang, W., Liu, W., Hou, R., Zhang, L., Schmitz-­Esser, S., Sun, H., Xie,
J., Zhang, Y., Wang, C., Li, L., Yue, B., Huang, H., Wang, H., Shen,
F., & Zhang, Z. (2018). Age-­associated microbiome shows the giant
panda lives on hemicelluloses, not on cellulose. The ISME Journal,
12, 1319–­1328.
Zhu, L., Wu, Q., Dai, J., Zhang, S., & Wei, F. (2011). Evidence of cellulose
metabolism by the giant panda gut microbiome. Proceedings of the
National Academy of Sciences, 108, 17714–­17719.
Zhu, L., Wu, Q., Deng, C., Zhang, M., Zhang, C., Chen, H., Lu, G., &
Wei, F. (2018). Adaptive evolution to a high purine and fat diet
of carnivorans revealed by gut microbiomes and host genomes.
Environmental Microbiology, 20, 1711–­1722.
Zhu, L., Yang, Z., Yao, R., Xu, L., Chen, H., Gu, X., Wu, T., & Yang, X. (2018).
Potential mechanism of detoxification of cyanide compounds by
gut microbiomes of bamboo-­eating pandas. mSphere, 3, e00229-­18.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information can be found online in the
Supporting Information section at the end of this article.
How to cite this article: Dai, Q., Ding, J., Cui, X., Zhu, Y.,
Chen, H., & Zhu, L. (2023). Beyond bacteria: Reconstructing
microorganism connections and deciphering the predicted
mutualisms in mammalian gut metagenomes. Ecology and
Evolution, 13, e9829. https://doi.org/10.1002/ece3.9829