Mitochondrial COI: a promising molecular marker for species identification in Foraminifera

Traditional morphological methods for species identification are highly time consuming, especially for small organisms, such as Foraminifera, a group of shell-building microbial eukaryotes. To analyze large amounts of samples more efficiently, species identification methods have extended to molecular tools in the last few decades. Although a wide range of phyla have good markers available, for Foraminifera only one hypervariable marker from the ribosomal region (18S) is widely used. Recently a new mitochondrial marker (COI) has been sequenced. Here we investigate whether this marker has a higher potential for species identification compared to the ribosomal marker. We explore the genetic variability of both the 18S and COI markers in 22 benthic foraminiferal morphospecies (orders Miliolida and Rotaliida). Using single-cell DNA, the genetic variability within specimens (intra) and between specimens (inter) of each species was assessed using next-generation sequencing. Amplification success rate was twice as high for COI (151/200 specimens) than for 18S (73/200 specimens). The COI marker showed greatly decreased intra- and inter-specimen variabillity compared to 18S in six out of seven selected species. The 18S phylogenetic reconstruction fails to adequately cluster multiple species together in contrast to COI. Additionally, the COI marker helped recognize misclassified specimens difficult to morphologically identify to the species level. Integrative taxonomy, combining morphological and molecular characteristics, provides a robust picture of the foraminiferal species diversity. Finally, we suggest the use of a set of sequences (two or more) to describe species showing intra-genomic variability additionally to using multiple markers. Our findings highlight the potential of the newly discovered mitochondrial marker for molecular species identification and metabarcoding purposes. Since this is a new marker, interpretation of metabarcoding data is still limited by the small number of available references.