. 2012 Jun 13;7(6):e26284. doi: 10.1371/journal.pone.0026284

Table 3. Permutation-based repeated measures ANCOVA results for each of the six phyla/subphyla bacterial categories (Clade).

Category/Clade	ACTINOBACTERIA	R²	P value	FDR
Main effects	Disease phenotype	0.126	0.001	0.01
	Steroids	0.021	0.029	0.08
Interactions	Age of Surgery * ATG16L1	0.041	0.014	0.07
Measurements		0.001	0.328	0.47

Category/Clade	BACTEROIDETES	R²	P value	FDR
Main effects	Disease phenotype	0.117	0.001	0.01
	Smoking	0.018	0.034	0.09
	5-ASA	0.021	0.025	0.08
	Steroids	0.023	0.013	0.07
Interactions	*Steroids Immunomodulators**	0.040	0.006	0.04
	Disease phenotype * 5-ASA	0.019	0.023	0.08
	5-ASA * Age	0.016	0.050	0.12
	5-ASA * C. difficile	0.019	0.023	0.08
Measurements		0.002	0.047	0.12
Category/Clade	FIRM.CLOSTRIDIUM GROUPIV	R²	P value	FDR
Main effects	Disease phenotype	0.161	0.001	0.01
	Gender	0.038	0.003	0.02
	Smoking	0.025	0.015	0.07
	NOD2	0.021	0.027	0.08
Interactions	*Disease phenotype Age**	0.040	0.003	0.02
	BMI * 5-ASA	0.025	0.010	0.06
Measurements		0.008	0.001	0.01
Category/Clade	FIRM.CLOSTRIDIUM GROUPXIVa	R²	P value	FDR
Main effects	Gender	0.023	0.019	0.08
Interactions	*Disease phenotype Age**	0.049	0.004	0.03
	Steroids * BMI	0.030	0.011	0.06
Measurements		0.006	0.006	0.04
Category/Clade	FIRM.BACILLUS	R²	P value	FDR
Main effects	Disease phenotype	0.175	0.001	0.01
	NOD2	0.016	0.046	0.12
Interactions	Disease phenotype * Steroids	0.040	0.002	0.02
	Disease phenotype * Age	0.027	0.032	0.09
	5-ASA * ATG16L1	0.025	0.025	0.08
	Steroids * Immunomodulators	0.019	0.027	0.08
Measurements		0.008	0.001	0.01
Category/Clade	PROTEOBACTERIA	R²	P value	FDR
Main effects	NOD2	0.027	0.016	0.07
Interactions	*Steroids NOD2**	0.040	0.002	0.02
	*5-ASA Race**	0.032	0.007	0.04
	Steroids * Immunomodulators	0.021	0.030	0.08
	NOD2 * ATG16L1	0.056	0.019	0.08
Measurements		0.040	0.001	0.01

Sequencing results for all three platforms were available for 157 samples (44 ileal CD, 53 colitis and 60 control non-IBD). The variables and first order interactions with P≤0.05 are listed above. To address multiple comparison issues, the Benjamini-Hochberg method was applied to adjust P-values to the false discovery rate (FDR). The variables and first order interactions with FDR ≤0.05 are bolded.