Comparison of species composition between morphology, bulk DNA and eDNA datasets
Morphological identification of the four samples yielded a total of 57 macrobenthos species. In addition, juveniles of Ophiura,Echinoidea, Ophelia , Nephtys , Phoronis ,Pontocrates , Spio, Magelona , Cirratulidae, Glycera ,Lanice as well as specimens belonging to Nemertea, Oligochaeta and Anthozoa were identified to the respective order, genus or phylum. A detailed list of all taxa and their abundance in each location is presented in ESM Table 5. The number of morphological species was highest in Location 120, where 39 species were identified, and decreased over Locations 330, 840 and ZVL (13, 10 and 3, respectively) (ESM Fig 4). For six morphologically identified species (Capitella minima, Caulleriella alata, Mediomastus fragilis, Pseudopolydora pulchra, Spirobranchus lamarckii and Tanaissus liljeborgi ), no COI sequence was present in our reference COI database. For the bulk DNA datasets, all primer sets except primer set D showed the same decreasing trend in species numbers from Location 120, over 330, 840 and ZVL (ESM Fig 4). For the eDNA samples from the ethanol preservative, primer sets B and D identified very low species numbers, and for all primer sets the decreasing trend in species numbers was less clear compared to the bulk DNA dataset. Bulk DNA generally detected a higher number of species than the eDNA samples from the ethanol preservative (52, 37, 37, 25 and 44 species for bulk DNA and 45, 13, 22, 4 and 46 species for eDNA from the ethanol preservative with primer sets A, B, C, D and E, respectively). The biggest discrepancy between primer sets was situated in the number of species detected for the most diverse sample 120B (ESM Fig 4).
Of the 57 species that were identified morphologically, primer set A was able to pick up the highest number of species (Fig 3): 34 and 23 of the morphological species were detected in the bulk DNA and in the eDNA from the ethanol preservative, respectively. All primer sets missed a substantial portion of the morphologically identified species: 21, 34, 33, 40 and 25 species of the 57 morphological species were only found in the morphological dataset, for primer sets A, B, C, D and E respectively (Fig 3). The number of morphological species detected in the bulk DNA samples (34, 23, 24, 17 and 28 for primer sets A, B, C, D and E respectively) was substantially higher than the number of morphological species detected in the eDNA samples from the ethanol preservative (23, 7, 9, 3 and 22 for primer sets A, B, C, D and E, respectively). Primer sets A and E clearly outperformed the detection of the morphological species in the eDNA from the ethanol preservative.