Table 2

Pairwise association between alpha diversity indexes and patient and sample parameters

			Smoke	Gender	B27	Age	Disease duration	Family	Psoriasis	Uveitis	Radio	NSAIDs	AntiTNF	DMARDs	Library	CRP	ASDAS
M0	Chao	Effect size	9.4	1.0	7.1	9.3	−4.5	7.8	0.8	2.0	0.1	0.6	1.0	1.3	0.7	5.1	−5.1
	Chao	P value	0.2262	0.7817	0.2262	0.7817	0.7817	0.2262	0.7817	0.7817	0.7817	0.7817	0.7817	0.7817	0.7817	0.7817	0.7817
	invSimpson	Effect size	0.8	0.9	11.6	8.8	−10.8	0.7	0.9	4.3	1.0	1.4	6.7	2.5	2.3	−6.8	−18.9
	invSimpson	P value	0.6288	0.6288	0.1409	0.6288	0.6288	0.6288	0.6288	0.6288	0.6288	0.6288	0.3834	0.6288	0.6288	0.6365	0.6288
M3	Chao	Effect size	9.5	2.0	6.0	7.9	−5.1	14.3	0.9	1.0	0.0	0.5	0.0	0.9		−3.8	−16.8
	Chao	P value	0.1370	0.7717	0.3111	0.9390	0.9390	0.0543	0.9390	0.9390	0.9638	0.9390	0.9760	0.9390		0.9390	0.6698
	invSimpson	Effect size	0.4	1.1	0.5	−1.7	−2.1	2.3	0.4	0.0	1.0	1.3	0.6	4.0		2.5	−24.3
	invSimpson	P value	0.8243	0.8128	0.8243	0.9503	0.9503	0.8019	0.8243	1.0000	0.8128	0.8128	0.8243	0.8128		0.9503	0.8019

Chao index was used as a measure of diversity and invSimpson index as a measure of evenness. For comparisons between a numerical and a categorical variable, Kruskal-Wallis test was used with Eta for effect size; for comparisons between two numerical variables, Spearman correlation was used. P values were adjusted for multiple testing using the false discovery rate.
ASDAS, Ankylosing Spondylitis Disease Activity Score; CRP, C reactive protein; DMARDs, disease-modifying anti-rheumatic drugs; NSAIDs, non-steroidal anti-inflammatory drugs.