Table 2.
Bacterial taxa whose quantity changed significantly with anti-TNF therapy in CD
| IBD | The change with anti-TNF therapy in patients with IBD | JIA | The change with anti-TNF therapy in patients with JIA | Notes | |||
|---|---|---|---|---|---|---|---|
| Bacterial taxon | Mean taxon abundance in IBD patients [% of total reads] | Mean abundance change in therapy [percentage points] | p-value: raw [corrected for number of tested taxa] | Mean taxon abundance in JIA patients [% of total reads] | Mean abundance change on therapy [percentage points] | Raw p-value [1] | |
| p. Bacteroidota └─ c. Bacteroidia └─ o. Bacteroidales |
29% | −8.1 | 6.2 × 10−4 [0.0061] |
22% | +4.1 | 0.34, NS | [2] |
| └─ f. Rikenellaceae │ |
2.6% | −1.9 | 2.3 × 10−5 [9.2 × 10−4] | 2.4% | +0.68 | 0.38, NS | |
| └─ g. Alistipes | 2.7% | −2.0 | 2.1 × 10−5 [0.0024] | 2.4% | +0.71 | 0.37, NS | |
| p. Firmicutes | 54% | +9.7 | 1.7 × 10−5 [1.7 × 10−4] | 64% | −4.8 | 0.19, NS | |
| └─ c. Clostridia │ |
43% | +11 | 2.0 × 10−6 [2.7 × 10−5] | 58% | −3.6 | 0.33, NS | |
| │─ o. Peptostreptococcales │ |
2.0% | +1.4 | 2.2 × 10−4 [0.0055] | 1.8% | +0.29 | 0.53, NS | [3] |
| │ └─ f. Peptostreptococcaceae │ |
1.7% | +1.1 | 4.3 × 10−4 [0.017] | 1.7% | +0.35 | 0.44, NS | |
| │ └─ g. Intestinibacter │ |
0.74% | +0.75 | 1.6 × 10−5 [0.0019] | 0.56% | +0.13 | 0.47, NS | |
| │─ o. Lachnospirales │ |
26% | +9.4 | 1.5 × 10−7 [3.7 × 10−6] | 30% | −4.3 | 0.12, NS | |
| │ └─ f. Lachnospiraceae │ |
26% | +9.4 | 1.6 × 10−7 [6.4 × 10−6] | 31% | −4.1 | 0.14, NS | |
| │ └─ g. Ruminococcus │ |
2.3% | +2.7 | 1.4 × 10−4 [0.015] | 0.2% | 0.0 | 0.94, NS | [4] |
| └─ ..... g. Flavonifractor | 0.4% | +0.38 | 3.7 × 10−4 [0.042] | 0.2% | −0.05 | 0.32, NS | [5] |
[1] Neither of the tested taxa yielded a corrected p-value of the mean change after anti-TNF in JIA lower than 0.05. There were 13 taxa having a raw uncorrected p-value <0.05 in this test. Neither of them had a mean abundance >1% in either CD or JIA, and all these spurious p-values were caused by one or a few outliers, as confirmed by inspection of the abundance graphs.
[2] The three taxonomic categories completely overlapped in our human faecal dataset. There were no Bacteroidia other than Bacteroidales and no Bacteroidota other than Bacteroidia.
[3] The order is named Peptostreptococcales–Tissierellales in the SILVA database v. 138.
[4] Classified as R. gnavus group by SILVA database v. 138.
[5] Genus Flavonifractor, class Clostridia, order Oscillospirales, family Oscillospiraceae [SILVA database v. 138].
The regression coefficient corresponds to a mean increase or decrease associated with the anti-TNF therapy. Here it is expressed as a change in percentage points, as the bacterial abundance enters the model as a relative count of reads per 10 000.
At each taxonomic level, those taxa were tested whose abundance exceeded 0.1% [10 reads per 10 000] in at least 3% of studied samples. This resulted in a Bonferroni correction for 115 tests at the genus level, 40 tests at the family level, 26 at the order level, 13 at the class level, and 10 at the phylum level.
NS, not significant; p., phylum; c., class; o., order; f., family; g., genus.