Delimitation and identification of microeukaryotic OTUs
The three different OTU generation methods infer somewhat different community compositions from sequenced long read rDNA environmental DNA amplicons. For instance, the methods resulted in very different estimates of total non-singleton OTU richness, ranging from 1,336 OTU_A detected based on inference of ASVs, compared to 2,046 OTU_S and 2,488 OTU_C for sequence similarity-based clustering using single-linkage or centroid-based clusters respectively (Table 1). The OTU_A dataset represents only 28% of the raw reads while the two other methods were comparable, capturing 81–83% of the reads into OTUs (Table 1). After pooling all reads, OTU accumulation curves for the three methods indicate that sequencing depth was sufficient to reach comparable asymptotic OTU richness estimates in both mesic-dry and wet soil conditions (Fig. S4a). For individual samples, however, increased sequencing depth would be expected to increase OTU detection for all three methods (Fig. S5). Further, the estimated asymptotic OTU richness increased for all methods when analyzing the pooled reads based on number of samples (Fig. S4b), indicating that taking additional samples would be expected to increase the number of detected taxa for all methods. Across the three methods, 38–42% of the detected OTUs were taxonomically assigned to kingdom Fungi. Based on read abundance, the estimated proportion of fungi ranged from 34% for OTU_As compared to just over 40% for the two other methods (Table 1). Protists thus dominated the sequenced microeukaryotic soil community at this site.