Novel molecular data for monogenean parasites of sparid fishes in the Mediterranean and a molecular phylogeny of the Microcotylidae Taschenberg, 1879

During a study of the monogeneans of four sparid fishes (Diplodus vulgaris, Pagellus bogaraveo, Pagrus pagrus and Sparus aurata) from the Western Mediterranean off Algeria, a large collection of parasites was characterised molecularly (28S rRNA and cox1 genes). A total of 46 partial sequences (23 for each gene) were generated from 38 isolates of monogeneans which included four species (Atrispinum acarne, Microcotyle erythrini (sensu stricto), Sparicotyle chrysophrii and Prostatomicrocotylinae gen. sp.) of the family Microcotylidae, two putative species of the family Capsalidae (Encotyllabe spp.), and one species (Choricotyle chrysophryi) of the family Diclidophoridae. Our study provides (i) the first molecular data for the Capsalidae in the Mediterranean; (ii) the first record of a member of the Prostatomicrocotylinae in the Mediterranean and in a sparid fish (D. vulgaris); (iii) the first cox1 sequences for A. acarne; (iv) the second record of M. erythrini (s.s.) from P. pagrus; and (v) the second confirmed by molecular data record of S. chrysophrii in wild populations of S. aurata. The first phylogenetic hypotheses for the family Microcotylidae developed here, revealed the monophyly of the subfamily Prostatomicrocotylinae and the genus Microcotyle but the relationships among the subfamilies were still largely unresolved with the best represented subfamily Microcotylinae being polyphyletic. Our results highlight the importance of molecular methods in the assessment of monogenean diversity and the need for a thorough taxon-sampling approach to increase the accuracy of phylogenetic reconstruction of the relationships of the large and taxonomically complex polyopisthocotylean family Microcotylidae.


Introduction
The Mediterranean is one of the first marine areas where parasite diversity has been explored and is a significant area of species occurrence and endemism of fish hosts and their parasites. Parasite diversity in Mediterranean fishes is generally well surveyed, especially for the most diverse host family (31 spp.), the Sparidae (see P erez-del-Olmo et al., 2016, for a recent review of digenean diversity). In the Mediterranean, sparid fishes are important commercial and recreational angling species and a promising source for the diversification of marine aquaculture (Basurco et al., 2011). However, culturing at high density inevitably leads to parasite-related problems associated with spill-over from and to wild fish populations aggregating around sea-cage facilities in coastal areas (see, e.g. Bouwmeester et al., 2020 for a recent review).
Monogeneans are a large group of pathogens with a direct life-cycle causing significant problems in farmed fish and thus may affect the success of the attempts to diversify sparid sea-cage cultures (e.g. Mladineo & Mar si c-Lu ci c, 2007;S anchez-García et al., 2015). A notable example is the microcotylid Sparicotyle chrysophrii (van Beneden & Hesse, 1863) representing a major parasitological problem in sparid cultures in the Mediterranean. There are several reports of lethal epizootics in Mediterranean cultures of Sparus aurata L. (see S anchez-García et al., 2015 and references therein). Therefore, accurate identification of monogeneans in wild sparid populations in areas prospective for aquaculture is important as it provides the baseline data for local pathogen diversity.
However, monogenean identification requires careful morphological studies and taxonomic expertise but similar to other platyhelminth groups, progressively fewer expert morphologists and taxonomists are being trained (Brooks, 2000). Molecular-based methods may provide a standardised framework for diversity assessment in understudied higher taxa and/or areas that is particularly promising for accelerating identification of previously characterised monogenean species where identification based on morphology is complicated (e.g. Bouguerche et al., 2019a, b;Ono et al., 2020;Víllora-Montero et al., 2020). The success of the latter approach depends on the existence of comprehensive sequence databases at least for the most diverse and widely distributed monogenean higher-level taxa.
Regarding sparid fishes worldwide, the polyopisthocotylean family Microcotylidae Taschenberg, 1879 is one such group characterised by a remarkable diversity (more than 160 species in 50 genera of 7 subfamilies; see Mamaev, 1986;WoRMS, 2021) and cosmopolitan distribution. However, although sequences generated for microcotylid species in the Mediterranean have been instrumental in the pioneer phylogenies of the Polyopisthocotylea (see Mollaret et al., 2000;Jovelin & Justine, 2001), the available genetic data for monogeneans of sparid fishes are scarce both globally and for the Mediterranean, and this may hinder diversity assessments for understudied areas and hosts and pathogen identification in aquaculture.
During a study of the monogeneans of sparid fishes from the western Mediterranean off Algeria, a large collection of parasites, which included four species of the family Microcotylidae, two putative species of the family Capsalidae Baird, 1853, and one species of the family Diclidophoridae Cerfontaine, 1895, was characterised molecularly (28S rRNA gene and cox1 gene). We show that both markers are effective for identification to the species level within the Microcotylidae and highlight a potential problem for the use of cox1 to differentiate Encotyllabe spp. Finally, we used the currently available DNA sequence data to explore the phylogenetic relationships within the most speciose polyopisthocotylean family, the Microcotylidae.

Sample collection
Four species of sparid fishes [Pagrus pagrus (L.) (n ¼ 308); Pagellus bogaraveo (Brünnich) (n ¼ 315); S. aurata L. (n ¼ 420); and Diplodus vulgaris (Geoffroy Saint-Hilaire) (n ¼ 390)] collected at four locations along the Algerian coasts of the Western Mediterranean (off Bouzedjar (35 34 0 24.58 00 N, 1 10 0 2.63 00 W); off Algiers (Bay of Algiers) (36 46 0 17.60 00 N, 3 03 0 57.00 00 E); off Skikda (36 53 0 15.94 00 N, 6 54 0 0.86 00 E); and off Annaba (36 57 0 23.03 00 N, 7 54 0 3.88 00 E)] were examined for parasite infections. Fishes were purchased between 2013 and 2016 from local fishermen during spring and autumn and subsequently examined at the Laboratory of the Environmental Monitoring Unit (LRSE), University of Oran 1, Algeria. Fishes were identified according to Bauchot & Hureau (1986) and examined fresh or at a later stage, after immediate freezing upon collection. Gills and pharyngeal tooth pads were removed, placed in individual Petri dishes, washed in saline solution (9 g NaCl/l) and examined for monogenean parasites under a stereomicroscope. Monogeneans were killed and preserved in molecular biology grade ethanol. Monogenean specimens were subsampled from the total collection and photomicrographs were made for each specimen selected for DNA isolation using an Olympus BX51 microscope (Olympus Corporation, Tokyo, Japan) with the aid of Quick Photo Camera 2.3 image analysis software. A small piece of the body just above the haptor was excised for DNA extraction. The voucher specimens (hologenophores sensu Pleijel et al., 2008) were stained with iron acetocarmine (Georgiev et al., 1986)

Sequence data generation
Genomic DNA (gDNA) was isolated using 5% suspension of deionised water and Chelex®, containing 0.1 mg/ml proteinase K; samples were incubated at 56 C for 3 h, boiled at 90 C for 8 min, and centrifuged at 14,000Â g for 10 min. Partial sequences of the nuclear 28S rRNA gene (domains D1-D3) and the mitochondrial cytochrome c oxidase subunit 1 gene (cox1) were amplified as follows. Partial fragments of the 28S rRNA gene were amplified using primer combinations LSU5  and 1500R (Tkach et al., 1999) (amplification profile: 5 min at 94 C, followed by 40 cycles of 94 C for 30 s, 55 C for 30 s, 72 C for 2 min, and a final extension step at 72 C for 7 min).
Partial fragments of the cox1 gene were amplified using the primers JB3 (¼ COI-ASmit1) and JB4.5 (Bowles et al., 1992) (amplification profile: 5 min at 94 C, followed by 35 cycles of 94 C for 1 min, 50 C for 1 min, 72 C for 1 min, and a final extension step at 72 C for 10 min).
PCR amplifications were performed in a total volume of 25 μl using Ready-To-Go-PCR Beads (GE Healthcare, Chicago, USA) following the manufacturer's instructions. In the case of poor amplification, the PCR reactions were performed with 2Â MyFi™ DNA Polymerase mix (Bioline Inc., Taunton, USA) in a total volume of 20 μl containing 8 pmol of each primer and c.50 ng of gDNA. PCR amplicons were purified using the QIAquick PCR purification kit (Qiagen Ltd., Hilden, Germany) following the manufacturerʼs instructions and sequenced directly for both strands using the PCR primers with ABI BigDye chemistry (ABI PerkinElmer, London, UK), alcohol-precipitated, and run on an Applied Biosystems 3730XL DNA analyser at GATC Biotech, Konstanz, Germany.

Phylogenetic analyses
Contiguous sequences were assembled using MEGA v.7 (Kumar et al., 2016) and subjected to a BLASTn search on the NCBI GenBank database. The newly generated and published 28S rDNA sequences were aligned with MAFFT v.7 (Katoh & Standley, 2013) under default gap parameters on the EMBL-EBL bioinformatics web platform (http://www.ebi.ac. uk/Tools/msa/mafft/). The newly generated and published cox1 sequences were aligned with MUSCLE implemented in MEGA v. 7 (Kumar et al., 2016) with reference to the amino acid translation, using the echinoderm and flatworm mitochondrial code (translation Table 9; Telford et al., 2000); the alignments included no insertions or deletions. Due to the insufficient variability provided by the amino acids alone, these alignments were analysed solely as nucleotides; first, second and third positions within the included codons were included in these analyses.
Molecular identification and assessment of phylogenetic relationships were achieved using neighbour-joining (NJ) analyses performed on Kimura 2-parameter distances with MEGA v.7 (Kumar et al., 2016). Bayesian inference (BI) analysis was carried out for the 28S rDNA dataset for the Microcotylidae with MrBayes version 3.2.6 using Markov chain Monte Carlo (MCMC) searches on 2 simultaneous runs of 4 chains for 10 7 generations, sampling trees every 10 3 generations. The "burn-in" was set for the first 25% of the trees sampled; consensus topology and nodal support were estimated from the remaining 75% of the trees. Prior to the analysis, the best-fitting model of nucleotide substitution was estimated with jModelTest 2.1.4 (Darriba et al., 2012) based on the corrected Akaike (AICc) and the Bayesian information criteria (BIC). This was the general time reversible model with gamma distributed among-site rate variation (GTRþГ). BI analysis was run on XSEDE on the Cipres Science Gateway v. 3.1 (http://www.phylo.org/sub_sections/portal/; Miller et al., 2010). Tree topologies were visualised in FigTree v.1.4.3 (http://t ree.bio.ed.ac.uk/software/figtree/.).
Pairwise genetic distances (uncorrected p-distance model) were calculated in MEGA v.7.

Overview of the novel molecular data and morphological assessment
A total of 46 partial sequences (23 for each gene) were generated from 38 isolates of monogeneans found in the four sparid fish hosts sampled off the Algerian coast of the Western Mediterranean (Table 1). The preliminary morphological assessment of the specimens indicated that these belong to three families: Capsalidae [Encotyllabe spp., a total of 23 isolates ex D. vulgaris (n ¼ 4), S. aurata (n ¼ 5) and P. bogaraveo Comparative sequence and phylogenetic analyses carried out here confirmed the identification based on morphology for the diclidophorid C. chrysophryi and the microcotylids except for Microcotylidae gen. sp. ex D. vulgaris and indicated the presence of two putative species of Encotyllabe Diesing, 1850.
The morphology of the specimens of the three species previously recorded in sparids (C. chrysophryi, A. acarne and S. chrysophrii) agreed with the redescription of Llewellyn (1941), the original description of Maillard & Noisy (1979) and the redescription of Euzet & Noisy (1979), respectively, as well as with the data provided by Radujkovi c & Euzet (1989) (C. chrysophryi and S. chrysophrii). Morphological identification of species of Microcotyle is not straightforward because of the morphological homogeneity of the species within the M. erythrini species complex (see Bouguerche et al., 2019a, b;Víllora-Montero et al., 2020). Therefore, identification of the present specimens of Microcotyle was based on the morphological and molecular characterisation of M. erythrini (s.s.) of Víllora-Montero et al. (2020).
Unfortunately, the condition of the single specimen ex D. vulgaris provisionally identified as Microcotylidae gen. sp. did not allow examination of its morphology in detail. However, this isolate fell within the clade representing the subfamily Prostatomicrocotylinae with strong support in the analysis of the cox1 sequences. Regarding the specimens of  Encotyllabe sequenced here, it is worth noting that these were not assigned a priory to one of the two forms; this separation was based on the consistent nucleotide differences and clustering of the specimens based on the more conservative 28S rRNA gene. The two forms delineated by the 28S analysis, exhibit morphological differences, the most apparent being the shape and size of the large hamuli. However, detailed morphological description of the present material of Encotyllabe spp. requires a revision of the genus and will be published elsewhere. Taschenberg, 1879 3.2.1. Analysis 1: cox1 dataset The alignment of the cox1 dataset for the Microcotylidae (304 nt) comprised 11 newly generated sequences (one for the outgroup, C. chrysophryi; see Table 1) and 39 sequences retrieved from GenBank (see Supplementary Table S1 for details). The neighbour-joining tree ( Fig. 1) revealed the monophyly of the genus Microcotyle van Beneden & Hesse, 1863 and the subfamily Prostatomicrocotylinae Yamaguti, 1963. However, a species of the only other genus representative of the subfamily Microcotylinae Taschenberg, 1879 in this analysis, Bivagina pagrosomi (Murray, 1931) Dillon & Hargis, 1965, clustered with A. acarne within the clade representing the subfamily Atriasterinae Maillard & Noisy, 1979(Atrispinum Euzet & Maillard, 1974þ Bychowskicotyla Unnithan, 1971þ Sparicotyle Mamaev, 1984.

Family Microcotylidae
The cox1 analysis provided molecular confirmation for the identification of M. erythrini (sensu stricto) as defined by Víllora-Montero et al.
The isolates of S. chrysophrii from off Algeria represented 5 unique novel haplotypes. Additionally, one isolate ex D. vulgaris provisionally identified as Microcotylidae gen. sp. fell within a strongly supported clade representing the subfamily Prostatomicrocotylinae and comprising the sequences for Polylabris halichoeres Wang & Zhang, 1998 and Polylabroides guangdongensis Zhang & Yang, 2000. There were no cox1 sequences for A. acarne on GenBank but two of the three isolates sequenced for cox1 were also sequenced for 28S (see below); thus, our study provides the first cox1 sequences for A. acarne.
Mitochondrial cox1 haplotype networks were constructed for the important fish pathogen in sparid aquaculture S. chrysophrii based on the newly generated and published sequences from sparid fishes in the Central and Western Mediterranean (Jovelin & Justine, 2001;Mladineo et al., 2009) (see Supplementary Table S2 for details). The aligned cox1 dataset for S. chrysophrii comprised 304 nt positions. A total of 12 unique haplotypes was identified in the 36 individual sequences included in this alignment: six from cage-reared S. aurata and wild Boops boops associated with these cages in the Central Mediterranean (Adriatic Sea, see Mladineo et al., 2009); and six (all singletons) from S. aurata sampled in the Western Mediterranean off S ete (Jovelin & Justine, 2001) and Algeria (present study) (Supplementary Table S2). As illustrated in Fig. 2, there is a significant differentiation between the haplotypes of S. chrysophrii from S. aurata in the Central and Western Mediterranean (genetic divergence of 2.3-3.9% compared with 0.3-0.7% for the haplotypes from the Central Mediterranean alone) and that the haplotype from B. boops may represent a cryptic species (genetic divergence of 3.9-7.9% compared with 0.3-3.9% for the haplotypes from S. aurata).
At the generic level, seven species of Microcotyle formed a strongly supported clade, sister to a clade representing Omanicotyle Yoon, Al-Jufaili, Freeman, Bron, Paladini &Shinn, 2013 andBivagina Yamaguti, 1963. Two otherwise unpublished sequences by Chou (Microcotyle spp. 1 and 2, no host and locality data available) and a sequence (GenBank: EF653385 Microcotylidae gen. sp. M10) for an unidentified microcotylid ex Sebastes sp. off UK published by Aiken et al. (2007) clustered within the clade of Microcotyle spp. and thus clearly represent sequences for three distinct species of Microcotyle. However, the sequence KT191025 reported by Jun (2015) as Microcotyle sp. ex Dentex dentex (L.) from the Eastern Mediterranean (off Crete) was strongly associated with B. pagrosomi and Omanicotyle heterospina Yoon, Al-Jufaili, Freeman, Bron, Paladini & Shinn, 2013 and thus does not belong to Microcotyle (Fig. 3).
The newly generated sequences from isolates parasitic in the Mediterranean sparids off Algeria clustered with a strong support with: (i) a sequence for A. acarne ex Pagellus acarne (GenBank: AF311702) from off France by Jovelin & Justine (2001) (Fig. 3).

Family Capsalidae Baird, 1853
The 15 newly generated 28S rDNA sequences (771 nt) from specimens of Encotyllabe spp. formed two strongly supported monophyletic clades tentatively labelled Encotyllabe sp. 1 ex P. bogaraveo and Encotyllabe sp. 2 ex D. vulgaris and S. aurata (Fig. 4). The sequences of these two forms differed at 2 nt positions and collectively differed from the available sequences for the remaining congeneric species at 4-12 nt positions. Additionally, the sequences for E. cf. spari ACAC-2017 (Gen-Bank: KY553149; Camargo et al., 2017), Encotyllabe sp. h FAS-2020 isolate F44 (GenBank: MT968927; unpublished) and Encotyllabe  Table 1  By contrast, in the phylogeny based on the cox1 dataset (269 nt) for Encotyllabe spp. comprising 13 newly generated sequences and 12 sequences retrieved from GenBank (outgroup: Benedenia seriolae (Yamaguti, 1934) Meserve, 1938), all novel sequences formed a single strongly supported monophyletic clade (Fig. 5). However, due to sequencing problems only a single isolate of Encotyllabe sp. 2 ex S. aurata was sequenced for both genes (indicated with a red star in Fig. 5). Of the eight isolates for which we were not able to generate 28S rDNA sequences, three (OL675218-OL675220) were identical with a cox1 sequence for Encotyllabe sp. 1 (OL675216), one (OL675221) was identical with another sequence for Encotyllabe sp. 1 (OL675215) and four could not be assigned to one of the two forms of Encotyllabe based on the phylogenetic results. The overall range of genetic divergence for the cox1 gene in the newly collected material of Encotyllabe was low (0-1.9%; 0-5 nt) but the upper limit was slightly higher than the upper limits of the intraspecific ranges of 0.7% (2 nt) and 0.4% (1 nt) for E. antofagastensis and Encotyllabe cheilodactyli Sepúlveda, Gonz alez & Oliva, 2014, respectively. Encotyllabe sp. 1 and Encotyllabe sp. 2 differed at 3-5 nt positions (1.1-1.9%), values above the known intraspecific divergence for cox1 in Encotyllabe.

Family Diclidophoridae Cerfontaine, 1895
Our study provides the first 28S rDNA and cox1 sequences for C. chrysophryi ex P. bogaraveo; these were used as the outgroups in the phylogenetic analyses. The newly generated 28S sequence (876 nt) differed from the only 28S sequence for C. chrysophryi available on GenBank (AF311705; ex P. acarne off France, see Jovelin & Justine, 2001) at 2 nucleotide positions (overlap 505 nt).
A comparison of the newly generated cox1 sequence for C. chrysophryi (alignment length of 304 nt) with the sequence for C. cf. chrysophryi ex P. acarne off France by Jovelin & Justine (2001) and the two sequences of C. cf. chrysophryi (as C. chrysophryi on GenBank) ex P. acarne off Algeria published recently by Bouguerche et al. (2021) revealed a wide range of genetic divergence: (i) very low (0.7%, 2 nt; our sequence vs MZ127222 by Bouguerche et al. (2021)); (ii) intermediate (6.3-7.0%, 19-21 nt; our sequence and MZ127222 vs AY009165 by Jovelin & Justine (2001)); and (iii) rather high (9.5-15.0%, 28-44 nt, the three sequences listed above vs the second sequence (MZ127216) by Bouguerche et al. (2021)). It is worth noting that the sequences AY009165 by Jovelin & Justine (2001) and MZ127216 by Bouguerche et al. (2021) have some ambiguously coded nucleotide positions (3 and 8, respectively) thus increasing to a degree the divergence levels. Nevertheless, the genetic distances between the divergent sequence MZ127216 and the remaining sequences for C. chrysophryi seem to fall above the levels of intraspecific divergence.  Supplementary Table S1 for details).
There are three sequences available on GenBank that provide data for congeneric comparisons, i.e. C. australiensis ex Chrysophrys auratus (GenBank: MT783687) off Australia recently re-described by Hossen et al. (2020). Comparative sequence analysis revealed levels of interspecific divergence of 18.8-22.3% between the Mediterranean and Australian species of Choricotyle.

Discussion
To the best of our knowledge this is the first molecular assessment of polyopisthocotyleans in sparids which is based on a focused systematic sampling of a range of hosts and in diverse coastal environments of the Western Mediterranean. Overall, our study expands the 28S rDNA and cox1 sequence databases with 46 sequences for seven species of three monogenean families (Capsalidae, Microcotylidae and Diclidophoridae). Comparative sequence analyses and phylogenetic reconstructions confirmed the morphological identification of four monogenean species (A. acarne, C. chrysophryi, M. erythrini (s.s.) and S. chrysophrii), helped assign the microcotylid from D. vulgaris provisionally identified as Microcotylidae gen. sp. to the subfamily Prostatomicrocotylinae and revealed that the capsalid specimens sequenced may belong to two putative species. Our study provides (i) the first molecular data for the Capsalidae in the Mediterranean; (ii) the first record of a member of the Prostatomicrocotylinae in the Mediterranean and in a sparid fish (D. vulgaris); (iii) the first cox1 sequences for A. acarne; (iv) the second record of M. erythrini (s.s.) from P. pagrus; and (v) the second confirmed by molecular data record of S. chrysophrii in wild populations of S. aurata.
The recent accumulation of 28S rDNA sequences for the Microcotylidae offers the opportunity for an updated phylogenetic hypothesis for this monogenean group. The first attempt to assess the relationships of the Microcotylidae based on 28S rDNA data was included within a wider reconstruction of the phylogenies of the Monopisthocotylea and Polyopisthocotylea (see Mollaret et al., 2000). These authors included sequences for representatives of five microcotylid genera, i.e. A. acarne  Mollaret et al., 2000). Jovelin & Justine (2001) inferred the relationships within the Polyopisthocotylea, and especially, between the terminal groups (Gastrocotylinea, Discocotylinea and Microcotylinea), based on partial (domain D2) 28S rDNA sequences. These authors provided new 28S sequences for representatives of four microcotylid genera, i.e. A. acarne (GenBank: AF311702), C. cf. chrysophryi (GenBank: AF311705), M. erythrini (Gen-Bank: AF311712) and S. chrysophrii (GenBank: AF311719); they also provided novel cox1 sequences for the latter three species (GenBank: AY009165, AY009159 and AY009161, respectively). Jovelin Table 1 and Supplementary Table S1 for details).
(2001) found the Microcotylidae to be monophyletic but the relationships within the family were poorly supported; a single clade (B. pagrosomi þ M. erythrini) was well supported. The present molecular analyses of the relationships within the Microcotylidae are based on a markedly expanded taxa representation at all levels, i.e. subfamily, genus, species. The phylogeny based on the 28S rRNA gene includes five of the seven subfamilies of the Microcotylidae, i.e. Anchoromicrocotylinae, Atriasterinae Maillard &Noisy, 1979, Metamicrocotylinae Yamaguti, 1963, Microcotylinae andProstatomicrocotylinae, which are represented by one to eight genera (19 genera in total). The largest microcotylid subfamily, the Microcotylinae, is the most densely sampled (represented by species of 10 out of 30 genera). The phylogeny based on the cox1 gene includes fewer taxa, representative for three subfamilies: Microcotylinae (2 genera, 12 species); Atriasterinae (3 genera, 3 species) and Prostatomicrocotylinae (2 genera, 3 species). Overall, 11 of the species included in the phylogenetic analyses represented isolates originating from the Western Mediterranean (Table 1 and  Supplementary Table S1). These are representative for six genera thus comprising nearly 50% of the known microcotylid diversity in the Mediterranean (29 species of 14 genera, see Gibson et al., 2005;Ayadi et al., 2017;Bouguerche et al., 2019a, b;Víllora-Montero et al., 2020;Azizi et al., 2021). The fact that our large-scale intensive sampling off Algeria did not increase the number of species indicates that a substantial proportion of microcotylid diversity in the Western Mediterranean has already been discovered and is represented in our phylogenetic hypotheses.
Both 28S rDNA and cox1 phylogenies provided strong support for the Prostatomicrocotylinae represented by different species of Polylabris and Polylabroides (Figs. 1 and 3). However, the two datasets yielded contradictory results for the Atriasterinae. This family represented by three out of six genera (Atrispinum, Sparicotyle and Bychowskicotyla), was also recovered as monophyletic albeit with poor support in the 28S phylogeny but the only available sequence for Bivagina pagrosomi (GenBank: Z83003) clustered with sequences for A. acarne within the clade of the Atriasterinae in the cox1 analysis. It is difficult to interpret this result as the cox1 sequence for this isolate has only been used as the outgroup in previous analyses (Littlewood et al., 1997;Bouguerche et al., 2019a, b). The identification of the host (as "S. aurata") for this isolate is also ambiguous as the occurrence of this sparid off Australia is considered questionable (Froese & Pauly, 2021). A plausible explanation would seem to be that Bivagina pagrosomi (GenBank: Z83003) represents a misidentification of the only species of Atriaster Lebedev & Parukhin, 1969 reported from a sparid host off  (unpublished data). Stars indicate the isolates of Encotyllabe for which 28S rDNA sequences were generated (see Table 1 and Supplementary Table S1 for details).
Although the important fish pathogen S. chrysophrii has been recorded in wild populations of S. aurata (e.g. Euzet & Audouin, 1959;Radujkovic & Euzet, 1989) and Diplodus puntazzo (Walbaum) (see S anchez-García et al., 2015), most of the genetic data come from cage-reared S. aurata in the Adriatic (Mladineo et al., 2009) (partial cox1 gene, 5 haplotypes, see Supplementary Table S2). Our study revealed a similar genetic diversity (6 haplotypes) based on a very limited sampling of wild populations of S. aurata in the Western Mediterranean. This, coupled with the significant differentiation between the haplotypes of the Adriatic and Western Mediterranean populations of S. chrysophrii (Fig. 2), indicates a much greater genetic diversity of this species in the wild populations of its sparid hosts (S. aurata and D. puntazzo) in the Western Mediterranean; wider sampling in this region is needed to test this hypothesis that might have serious repercussions for sea-cage sparid aquaculture in this region. Finally, the very divergent haplotype from B. boops may represent a cryptic species of Sparicotyle pending additional morphological and molecular study.
Our study provides the first 28S rDNA and cox1 sequences for species of Encotyllabe in the Mediterranean and adds to the rather scarce molecular data for this genus worldwide. The genus currently contains 24 species (WoRMS, 2021). Of these, only three have been recorded in sparid and labrid fishes in the Mediterranean: Encotyllabe nordmanni Diesing, 1850 ex Sparus brama Bonnaterre and Chromis chromis (L.); Encotyllabe paronae Monticelli, 1907 ex Symphodus tinca (L.); and Encotyllabe vallei Monticelli, 1907 ex D. puntazzo and S. aurata (see Euzet et al., 1993). However, the original descriptions of these species are too brief and insufficient for species differentiation/identification and there is just one study (Radujkovich & Euzet, 1989) providing some morphological data for a single specimen of E. vallei. Thus, elucidating the actual diversity of Encotyllabe in the Mediterranean would require detailed morphological studies, preferably in association with molecular data.
Comparative sequence analyses of the Microcotylidae showed that both partial 28S and cox1 genes can be successfully used for species identification. However, the number of sequences available for representatives of the family for both loci is still very low in spite of the recent surge in the use of cox1 DNA data for monogenean species discovery and identification in the Western Mediterranean (e.g. Chaabane et al., 2016a, b;Kheddam et al., 2016;Ayadi et al., 2017;Bouguerche et al., 2019a, b, Víllora-Montero et al., 2020Bouguerche et al., 2021).
Although the use of the short sequences of the barcoding cox1 gene may substantially facilitate and accelerate identification of previously characterised species especially in large-scale inventories and aquaculture settings, amplification success is considered to represent the major limitation to a wider use of cox1 barcoding in monogeneans (Vanhove et al., 2013). Whilst we have been successful in amplifying the cox1 fragment for the microcotylids, this was not the case with the specimens of Encotyllabe; the amplified fragment was short, and amplification failed for most isolates of Encotyllabe sp. 2. The position of the single successfully sequenced isolate in the cox1 phylogeny did not support the distinct status of this putative species. This scenario may occur in other monogenean groups so that for species distinction using cox1 it is necessary to generate more sequences per taxon and rely on the detection of reciprocally monophyletic lineages in phylogenies. Further, the available data are insufficient to assess the intra-and interspecific genetic variation for most polyopisthocotylean genera. Clearly vast additional taxon sampling and preferably sequencing of both 28S rRNA and cox1 genes would help advance species discovery and identification of the monogenean groups studied.

Conclusions
Our focused systematic sampling of monogeneans in four sparid hosts from the Western Mediterranean off Algeria expanded the 28S rDNA and cox1 sequence databases with 46 sequences for seven species of the families Microcotylidae, Capsalidae and Diclidophoridae. Comparative sequence and phylogenetic analyses confirmed species identification based on morphology and helped clarify the status of some species of Microcotyle and Choricotyle. The first phylogenetic hypotheses for the family Microcotylidae provided here, revealed the monophyly of the subfamily Prostatomicrocotylinae and the genus Microcotyle but the relationships among the subfamilies were still largely unresolved with the largest and best represented subfamily Microcotylinae being polyphyletic. We highlight that a thorough taxonsampling approach is required to increase the accuracy of phylogenetic reconstruction of the relationships of the diverse polyopisthocotylean family Microcotylidae.