2014 Env Microbiol, Supplementary Tables (S1-S9)

Moreover, trends between methods (Pearson correlation of diversity estimates), absolute shifts (as binary log-ratio) and shift significance (as p-values in one-sided Mann-Whitney-U tests) are also provided for every index. Finally, an overview of the HSM dataset (sequence counts and internal sequence similarities per habitat), as well as pairwise partition similarities between methods at 97% clustering are also provided. # Table_S4.xlsx Adjusted Mutual Information (AMI) between methods across thresholds when clustering the HSM dataset.

Values as shown in Figure S1. # Table_S5.xlsx Normalized Mutual Information (NMI) between methods across thresholds when clustering the HSM dataset. Values as shown in Figure S2. # Table_S6.xlsx Adjusted Rand Index (ARI) between methods across thresholds when clustering the HSM dataset. Values as shown in Figure S3. # Table_S7.xlsx Adjusted Mutual Information (AMI) between methods across thresholds when clustering the global dataset of 887,870 16S sequences.

Values as shown in Figure 4 in the main text. Missing values for AL at thresholds <92% and UCLUST / UPARSE >99% as clustering to these thresholds was prohibited by memory requirements (see Text S1). # Table_S8.xlsx Normalized Mutual Information (NMI) between methods across thresholds when clustering the global dataset of 887,870 16S sequences. Values as shown in Figure S4.

Missing values for AL at thresholds <92% and UCLUST / UPARSE >99% as clustering to these thresholds was prohibited by memory requirements (see Text S1). # Table_S9.xlsx Adjusted Rand Index (ARI) between methods across thresholds when clustering the global dataset of 887,870 16S sequences. Values as shown in Figure S5. Missing values for AL at thresholds <92% and UCLUST / UPARSE >99% as clustering to these thresholds was prohibited by memory requirements (see Text S1).