Metabarcoding Reveals High Diversity of Benthic Foraminifera Driven by Atlantication of Coastal Svalbard

Arctic marine biodiversity is undergoing rapid changes due to global warming and 13 modifications of oceanic water masses circulation. These changes have been demonstrated in 14 the case of mega- and macrofauna, but much less is known about their impact on the 15 biodiversity of smaller size organisms, such as foraminifera that represents a main component 16 of meiofauna in the Arctic. Several studies analysed the distribution and diversity of Arctic 17 foraminifera. However, all these studies are based exclusively on the morphological 18 identification of specimens sorted from sediment samples. Here, we present the first assessment 19 of Arctic foraminifera diversity based on metabarcoding of sediment DNA samples collected 20 in fjords and open sea areas in Svalbard Archipelago. We obtained a total of 5,968,786 reads 21 that represented 1,384 ASVs. More than half of the ASVs (51.7%) could not be assigned to 22 any group in the reference database suggesting a high genetic novelty of Svalbard foraminifera. The sieved and unsieved samples resolved comparable communities, sharing 1023 ASVs, comprising over 97% of reads. Our analyses show that the foraminiferal assemblage differs 25 between the localities, with communities distinctly separated between fjord and open sea 26 stations. Each locality was characterized by a specific assemblage, with only a small overlap 27 in the case of open sea areas. Our study demonstrates a clear pattern of the influence of water 28 masses on the structure of foraminiferal communities. The stations situated on the western 29 coast of Svalbard that is strongly influenced by warm and salty Atlantic Water (AW) are 30 characterized by much higher diversity than stations in the northern and eastern part, where the 31 impact of AW is less pronounced. This high diversity and specificity of Svalbard foraminifera 32 associated with water mass distribution indicate that the foraminiferal metabarcoding data can 33 be a very useful tool for inferring present and past environmental conditions in the Arctic.

The sieved and unsieved samples resolved comparable communities, sharing 1023 ASVs, 24 comprising over 97% of reads. Our analyses show that the foraminiferal assemblage differs

38
The Arctic Ocean is strongly impacted by the increased influence of warm and saline 39 88 2. Study area 89 The oceanography of Svalbard region is shaped mainly by the interplay between warm 90 and saline Atlantic Water (AW) and cold Arctic Water (ArW), as well as locally formed water 91 masses 16,49 . AW is transported northward along the Spitsbergen shelf edge as the West 92 Spitsbergen Current (WSC, Fig. 1) 50,51 . WSC is one of the major heat contributors to the Arctic   involved the removal of unique (occurring in only one sample) and rare ASVs (having less 167 than 10 reads).

168
The remaining ASVs were compared to the curated database of foraminiferal 18S 169 rDNA sequences 67,68 and the PR2 database v4.11.1 69 using VSEARCH, implemented in 170 SLIM, and BLASTN 70 based on minimum similarity (-perc_identity 80%) and minimum 171 coverage (-qcov_hsp 80%) for the taxonomic assignment to six taxonomic levels (phylum; 172 class; order; family; genus; species). The representative sequences of ASVs that remained 173 unclassified with the foraminiferal database, were aligned in a stand-alone BLAST using 174 BLAST (v2.7.1) search against the NCBI's non-redundant nucleotide database. The sequences 175 diverging by less than 1% were considered as belonging to the same species/genus. ASVs 176 below 99% identity were classified at the family, order, or class or as unassigned foraminifera.
Finally, taxonomic compositions in terms of cluster abundance were compared among 178 processing methods only using clusters reliably assigned at the species/genus level.      Table S4.    (Table S5).
to environmental sequences, that possibly represent non-described taxa. More than half of the 287 ASVs (51.73%) were assigned with low similarity (< 0.9).  Table S6). In Fig. 5a, the Venn diagram showed that 1023 ASVs 296 (corresponding to 97.23% of the reads) were shared among sieved and unsieved samples. The The taxonomic composition of benthic foraminifera also changed between the 320 locations. At the class level (Fig. 4), the monothalamous taxa were the dominant group which  identity percentage with the reference database of more than 99% and no less than 10 reads 337 were picked and those attributed to the same taxa were merged (Fig. 6).   The four alpha-diversity indices (Observed ASVs, Chao1, Simpson and Shannon) were 358 measured separately for sieved and unsieved datasets (Fig. 7) and showed clear variation

397
In the results of the sPLS regression, we detected several foraminiferal ASVs lineages 398 for which relative sequence abundance correlated with environmental parameters (Fig. 9 and 399 Table S8). The sPLS regression and subsequent hierarchical clustering suggested that the data 400 was separated into three clusters (Fig. 9). These include lineages identified as potential  (Table S8).

444
As shown by our study, the taxonomic composition of different size fractions is not the 445 same. For example, the order Rotaliida was the most abundant in 500-100 µm and 100-63 µm 446 size fractions. Also, another hard-shelled order Textulariida, which is microscopically studied 447 in the 500-100 µm fraction, in metabarcoding data is present mainly in fractions 500-100 µm and 100-63 µm (Fig. 5c) This is congruent with the rotaliids and textulariids dominating 449 microscopic assemblage found in >63 µm sieved fraction. On the other hand, the smallest 450 fraction (<63 µm) was dominated by monothalamiids and undetermined Foraminifera (Fig.   451 5C), which may suggest the presence of some genetically unknown, tiny monothalamous 452 species. 453 We also observed some differences between sieved and unsieved samples regarding the showed sieved and unsieved samples clustered together (Fig. 8, Table S7), indicating that there 467 is no significant difference in community composition inferred by the two methods.  (Fig. 6), or the increase of alpha diversity from glacier proximal/inner to glacier-487 distant/outer stations (Fig. 7), which are in agreement with the previous morphology-based 488 studies 81-83 , the composition of foraminiferal communities is generally specific to each 489 location. Each fjord forms a separate cluster and only some stations at open-water areas overlap 490 with each other (Fig. 8a).

491
The high-Arctic settings are usually considered as a cold system influenced at different 492 levels by ArW during summer to late autumn 84 , and covered by sea ice in winter 9,59,85 .

493
However, the increased influence of AW and winter sea-ice loss is observed in recent years    representing genera Hipocrepinella and Psammosphaera (Fig. 6). In particular, station EDG1 where the impact of warmer and more saline Atlantic Water (AW) was much pronounced, 550 confirmed by our CTD profile (Fig. 2). Also, sub-clusters that formed within cluster 2 reflected 551 different impacts of AW. The first sub-cluster comprises stations located in the glacial-distant