Table 2. Pairwise comparisons between replicate systems for RAS and AS separately, based on DGGE and 454 pyrosequencing data sets.

Statistical test	Analytical Method		D00		D07		D14		D28		D42
			AS1-5	Ra-Rb	AS1-5	Ra-Rb	AS1-5	Ra-Rb	AS1-5	Ra-Rb	AS1-5	Ra-Rb
Perm. ANOVA	GUT	DGGE	2/10	P:0.0001	6/10	P:0.0003	6/10	P:0.0001	7/10	P:0.0003	7/10	P:0.0001
		454			P:0.2079	P:0.0025					P:0.0242	P:0.016
ANOSIM		DGGE	R:0.428	R:0.925	R:0.787	R:0.772	R:0.542	R:0.922	R:0.745	R:0.596	R:0.907	R:0.862
			P:0.0005	P:0.0001	P:0.0001	P:0.0001	P:0.0001	P:0.0001	P:0.0001	P:0.0001	P:0.0001	P:0.0001
		454			R:1	R:1					R:1	R:0.846
					P:0.1	P:0.002					P:0.1	P:0.002
Perm. ANOVA	WATER	DGGE		P:0.0002		P:0.0212		P:0.0002		P:0.0001		P:0.0008
ANOSIM				R:0.996		R:0.972		R:1		R:1		R:0.988
				P:0.008		P:0.008		P:0.008		P:0.008		P:0.008

Tests were performed by sample type per sampling day, with P values for each comparison from two different statistical tests (PERMANOVA and ANOSIM). ANOSIM results are complementary to the PERMANOVA to provide information on the degree of separation between groups, suggested by R statistic. The 2/10, 6/10 & 7/10 indicate the number of significant out of the total available comparisons. D00, 07, 14, 28, 42: sampling day 0,7,14,28 & 42 respectively, AS1-5: replicate active suspension system 1 till 5, Ra & Rb: replicate recirculating system a & b. P values are based on 9999 Monte Carlo permutations.