Table 1 List of identified hub species and keystone species in the gut microbiome networks of pediatric HSCT patients.

From: Levofloxacin prophylaxis and parenteral nutrition have a detrimental effect on intestinal microbial networks in pediatric patients undergoing HSCT

Treatment group

Node property

T0

T1

T2

EN

Hub

Clostridium innocuum (1.60), Faecalibacterium prausnitzii (0.94), Eggerthella lenta (0.71)

Bifidobacterium longum (4.17), Bacteroides fragilis (0.3), Enterococcus faecalis (0.262), Bacteroides uniformis (0.2)

Bifidobacterium breve (6.60), Bifidobacterium longum (4.17), Ruthenibacterium lactatiformans (3.35), Parabacteroides distasonis (0.2)

Keystone

Bifidobacterium breve (−58.89%), Bifidobacterium longum (−57.72%), Faecalibacterium prausnitzii (51.51%), Eggerthella lenta (−52.8%), Ruthenibacterium lactatiformans (−51.6%)

Eggerthella lenta (−53.37%), Bacteroides fragilis (−54.29%), Clostridium innocuum (−53.83%), Erysipelatoclostridium ramosum (−53.7%), Bacteroides vulgatus (−51.2%)

Bifidobacterium longum (−53.99%); then slightly lower than threshold: Erysipelatoclostridium ramosum (−49.75%)

PN LVX (–)

Hub

Akkermansia muciniphila (2.32), Bifidobacterium adolescentis (1.37), Faecalibacterium prausnitzii (1.29), Eggerthella lenta (0.19), Alistipes finegoldii (0.1)

Bifidobacterium longum (4.07), Bifidobacterium adolescentis (1.37), Eggerthella lenta (0.71)

Ruthenibacterium lactatiformans (3.35), Bacteroides vulgatus (1.23), Erysipelatoclostridium ramosum (0.5), Bacteroides uniformis (0.31)

Keystone

Bifidobacterium longum (−57.04%), Bifidobacterium breve (−54.18%), Bifidobacterium adolescentis (−51.85%), Escherichia coli (−50.54%), Bacteroides uniformis (−50.39%)

Akkermansia muciniphila (−55.21%), Bifidobacterium breve (−51.57%)

Bacteroides fragilis (−51.03%), Bacteroides uniformis (−50.03%)

PN LVX (+)

Hub

Bifidobacterium adolescentis (1.37), Flavonifractor plautii (0.95), Fusicatenibacter saccharivorans (0.92), Roseburia inulivorans (0.2)

Clostridium innocuum (1.6), Klebsiella quasipneumoniae (1.02), Bacteroides dorei (0.86), Clostridium symbiosum (0.1)

Ruthenibacterium lactatiformans (3.55), [Ruminococcus] gnavus (2.15), Flavonifractor plautii (0.95), Enterococcus faecium (0.91)

Keystone

No species detected whose absence caused a reduction >50% in TC. The more impactful was: [Ruminococcus] gnavus (−13.87%)

No species detected whose absence caused a reduction >50% in TC. The more impactful were: Akkermansia muciniphila (−25.5%), Klebsiella pneumoniae (−14.99%), Klebsiella quasipneumoniae (−13.49%), Klebsiella variicola (−13.85%)

No species detected whose absence caused a reduction >50% in TC. The more impactful were: Klebsiella pneumoniae (−24.99%), Klebsiella quasipneumoniae (−23.49%), Klebsiella variicola (−18.85%)

  1. Hub species and keystone species are reported for each patient group (PN combined with LVX prophylaxis—PN LVX (+) vs PN alone—PN LVX (–) vs EN alone—EN) before (T0) and at two time points after HSCT (T1 and T2). Hub nodes were detected using three measures of centrality, i.e., node degree, betweenness centrality, and closeness centrality. Only nodes with values of all three parameters greater than the median estimated value of their normal distribution were considered hubs. Relative abundance values are shown in the corresponding brackets. Keystone taxa were defined as such when their absence caused at least a halving of the total cohesion (TC) values in the network. A leave-one-out approach was adopted for one species at a time to check the impact of its absence over the microbial network. Percentages of reduction caused over TC values are shown in the corresponding brackets. Taxa scored as both hub node and keystone species for a given treatment group and timepoint are underlined.
Back to article page