3.3 Within-site variation composition and relative abundance (RA)
We next explored similarity between PCR replicates by comparing the RA of taxa detected in them. Qualitatively, for both PITS and FITS, within-site PCR replicates appear similar to each other and each site is readily distinguishable from other sites (Figure 3). Only one replicate appeared to have a different community profile, and this was in PITS results from the YL.1 site (Figure 3). Analysis of ‘local contribution to beta diversity’ (LCBD) provides a quantitative examination of differences in composition and RA. At a read sampling depth of 5,000 and with a five read minimum cutoff, LCBD statistics identified several PCR outliers in the PITS data set (11 PCR replicates from the YL.1 site and 2 PCR replicates from the FO.2 site; Table S5) and one outlier in the FITS data set (1 outlier replicate from the FO.2 data set; Table S5).
To explore how read sampling depth and minimum read cutoff influence LCBD, we repeated these analyses at all three read sampling depths (1,000, 5,000, and 10,000 reads) with minimum read cutoffs of two, five, or ten reads (Table S5), and performed Chi-squared tests for significant differences among groups. The number of LCBD-based outliers increased significantly with higher read sampling depth in the FITS data set (p= 3.861E-15), but not in the PITS data set (p=0.25). We found no significant effect of minimum read cutoff for either the PITS (p=0.71) or FITS (p=0. 79) data set, suggesting that low abundance taxa, which are most likely to be impacted by changing the minimum read cutoff, are not causing outliers. For both the PITS and FITS data set, we found that site itself has a significant effect on the number of observed PCR outliers (both p <2.2e-16).