DNA sequencing
Surface sediment samples from the Baltic Sea are sequenced through Sanger with haptophyte-specific primers targeting 18S rRNA as described in 22. In short, DNA was extracted and purified from 0.5 g of dried sediments. One round of PCR was performed with primer pair Prym-429F and Prym-887R 44. The PCR products were run on gel electrophoresis, and gel bands with targeted length were purified and cloned. The clone products were sequenced on an ABI 3730XL capillary sequencer (Applied Biosystems, Foster City, CA). Operational Taxonomic Units (OTUs) were clustered by UCLUST. Representative sequences of each OTU were deposited in GenBank under accession numbers MH472563-MH472566.
The surface sediment samples from the Chesapeake Bay and Kangerlussuaq Fjord were sequenced through Next Generation Sequencing (NGS) with haptophyte-specific primers targeting 18S rRNA following methods described in Wang et al. (2021), and stored under accession number PRJNA835469 in GenBank. In short, DNA was extracted from the samples, followed by a two-step PCR. The first round of PCR used an adapted version of the primer pair 528Flong and PRYM01+7 (Egge et al., 2013) followed by a second round of PCR which gives each sample a unique index sequence. The cleaned and normalized amplicons were then sequenced on an Illumina Miseq with V2 600-Cycle Kit (San Diego, CA). Sequences were verified using FastQC v0.11.8 and clustered into Amplicon Sequence Variants (ASVs). Taxonomy of ASVs were assigned using the SILVA reference database.
Water and ice sample collection and sequencing in the Gulf of Finland is described in Enberg et al21. Relative read abundance (RRA) to total haptophyte of Eukaryotes for the i th OTU is defined as:\(\text{RRA}_{i}=\frac{1}{S}\sum_{k=1}^{S}{\frac{n_{i,k}}{\sum_{i=1}^{T}n_{i,k}}\times 100\%}\), where S is the number of samples, T is the total number of haptophyte or Eukaryotes OTU, \(n_{i,k}\) is the number of sequences of OTU i in sample k 46. The presence/absence of Isochrysidales sequences was defined as if any ASV or OTU were identified as Isochrysidales in the sample.
The downcore sediment samples from Upernavik Fjord were sequenced on an Illumina MiSeq 500-cycle paired-end run with haptophyte-specific primers (HAP_LSU_F and LHapto20R_bis) targeting D1-D2 region of 28S rRNA following methods described in Richter et al. (2019). Isochrysidales sequences were selected through BLAST and aligned to the 28S rRNA reference sequences in Richter et al. (2019) by PyNast. Maximum-likelihood phylogenetic trees for 18S and 28S rRNA were built by RAxML under GTRCAT approximation with 1000 bootstrap on CIPRES to determine the sub-groups of the Isochrysidales sequences47,48.
We also reanalyzed environmental sequences recovered in seawater samples from Southern Ocean near Antarctica Peninsula (Supplementary Data), following methods described in Wang et al.3. Sequences classified as Isochrysidales are included in phylogeny tree shown in Fig. S1.