Genome-wide SNP Data and Integrated Morphology Reveal a New Trout Species from İvriz and Berdan Streams, Türkiye

Abstract

Salmo sengulae sp. nov. is described from the Berdan (a drainage of Mediterranean Sea) and İvriz (a drainage of Konya closed basin) streams. It is distinguished from Salmo species in adjacent water by having 8–10 parr marks on flank; no red spots on flank in specimens larger than about 180 mm SL, if the red spot present in specimens larger than 180 mm SL, they are almost covered with black dots; number of black spots increasing with size, while number of red spots decreasing with size; a longer maxilla in males; a shorter predorsal length; fewer scale rows between dorsal-fin origin and lateral line; more scale rows between anal-fin origin and lateral line; and 16–19 gill rakers on first gill arch. Salmo sengulae sp. nov. was also compared the two closed species (Salmo chilo and Salmo kottelati) by using PCA analysis. The results of the PCA confirmed that Salmo sengulae sp. nov. differs from the other two species (S. chilo and S. kottelati). Salmo sengulae sp. nov. shares the same mtDNA Cyt b (991 bp) haplotype with S. chilo and S. kottelati. Additionally, genome-wide SNP data confirmed the separation of S. sengulae from those of Mediterranean trouts of Türkiye.

BACKGROUND

Türkiye is situated at a latitude of 36–42°N and a longitude of 26–45°E, constitutes one of the richest biodiversity regions, conveniently located at the intersection of major hot spots as (i) Caucasus at the north, (ii) Irano-Anatolia at the east and (iii) Mediterranean at the southern part of the country (Noroozi et al. 2019) out of 36 biodiversity hotspots identified globally (Myers et al. 2000).

For a long time, the brown trout (Salmo trutta, Linnaeus, 1758) has been widely acknowledged as a species spread across Europe, with its range extending southward to the Atlas Mountains (Morocco, Algeria) and eastward to the upper Amu-Darya drainage in Afghanistan. Despite the recognition of numerous nominal subspecies or distinct species of S. trutta over the years, there has always been a prevailing tendency to overlook this diversity (Guinand et al. 2021), and instead asserting that they all fall under a highly variable “species complex” of which traditional taxonomy struggles to adequately classify such variation (Turan et al. 2010). Recent phylogenetic studies have identified five main genetically distinct lineages based on mtDNA control region, namely Adriatic (AD), Atlantic (AT), Danubian (DA), Marmoratus (MA), and Mediterranean (ME) (Bernatchez et al. 1992; Bernatchez and Osinov 1995; Bernatchez 2001) some of which further splits as a result of shorter period of isolation within localities such as: Duero (DU) reported in Spain, Tigris (TI) reported in Türkiye (Sušnik et al. 2005; Bardakçı et al. 2006) and Dades (Snoj et al. 2011) proposed in Morocco. Four of these lineages are distributed in Anatolia. Twenty valid Salmo species inhabit Türkiye. Ten species belong to the Danubian lineage which are S. abanticus Tortonese, 1955 (Lake Abant), S. araxensis Turan, Kottelat & Kaya, 2022 (Aras River), S. ardahanensis Turan, Kottelat & Kaya, 2022 (upper drainages of Kura River), S. brunoi Turan, Bayçelebi, Aksu & Oral, 2024 (Nilüfer Stream, a drainage of Marmara Sea), S. coruhensis Turan, Kottelat & Engin, 2010 (the streams and rivers from Turkish Black Sea coast and Marmara drainages), S. duhani Turan & Aksu, 2021 (Gönen Stream, south western Marmara drainage and Ayazma Stream, a drainage of Aegean Sea), S. euphrataeus Turan, Kottelat & Engin, 2014 (Karasu River, northern Euphrates drainage), S. fahrettini (Karasu River, northern Euphrates drainage), S. murathani Turan, Kottelat & Kaya, 2022(Aras River), S. rizeensis Turan, Kottelat & Engin, 2010 (Turkish Black Sea coast). Eigth species belong to Adriatic lineage which are S. baliki Turan, Aksu, Oral, Kaya & Bayçelebi, 2021(upper drainages of Murat River, Euphrates drainage), S. ekmekciae Küçük, Kalaycı, Güçlü, Oral & Turan, 2024 (Köprüçay River), S. kottelati Turan, Doğan, Kaya & Kanyılmaz, 2014 (Alakır Stream, Mediterranean drainage), S. chilo Turan, Kottelat & Engin, 2012 (Ceyhan River), S. labecula Turan, Kottelat & Engin, 2012 (lower drainages of Seyhan River), S. munzuricus Turan, Kottelat & Kaya, 2017 (Munzur Stream northern Euphrates drainage), S. okumusi Turan, Kottelat & Engin, 2014 (Tohma and Göksu streams,western Euphrates drainage), S. platycephalus Behnke, 1968 (upper drainages of Seyhan River). S. opimus Turan, Kottelat & Engin, 2012 (Alara Stream,Mediterranean drainage) and S. tigridis Turan, Kottelat & Bektaş, 2011(Tigris River) are the single members of the Marmoratus and the Tigris lineages, respectively (Tortonese 1955; Behnke 1968; Bernatchez and Osinov 1995; Sušnik et al. 2005; Bardakçı et al. 2006; Turan et al. 2010 2011 2012 2014a–b 2017 2020 2021 2022 2024; Ninua et al. 2018; Turan and Aksu 2021; Küçük et al. 2024).

The present study aims to describe Salmo populations distributed in Berdan (drainage of the Mediterranean Sea) and İvriz streams (Konya closed basin). Given the ongoing debate of Salmo species, integrated taxonomic investigation scheme has been applied: a large number of specimens was investigated morphologically, these were combined with the widely accepted phylogenetic approach of mtDNA and additionally, supported with the next generation sequencing technologies providing reduced representation of genome of interest.

MATERIALS AND METHODS

Fish sampling

The care of experimental animals followed the animal welfare laws and guidelines declared by the Republic of Türkiye, thus following the policies approved by the RTE University Local Ethics Committee for experimentations (Permit reference number 2014/72). First, 80 mg/L MS222 was performed for anaesthesia. Secondly, fish were collected for faunal surveys, preserved in 5% formaldehyde or 96% ethanol, and stored in 70% ethanol. Surgical procedures were only performed for excision of fin clips. Thus, the experimental conditions did not cause severe stress on specimens under investigation.

Morphological analyses

All measurements were performed point to point (never by projections) as specified in Turan et al. (2010) with a dial caliper calibrated to 1 mm. The number of lateral line scale count, standard length, and caudal peduncle length were measured according to Turan et al. (2010). The last two branched rays articulating on a single pterygiophore in the anal and dorsal fins are counted as “1½”. Thirty-five measurements of new species (n = 23), S. chilo (n = 42) and S. kottelati (n = 18) were analyzed with a principal component analysis (PCA) using the software PAST version 1.8 (Hammer et al. 2001). After proportioning to SL, all metric characters (raw data) underwent log10-transformation prior being subjected to PCA.

Comparison material

For S. abanticus, S. araxensis, S. ardahanensis,

S. brunoi, S. coruhensis, S. duhani, S. euphrataeus, S. fahrettini, S. murathani, S. baliki, S. ekmekciae, S. kottelati, S. labecula, S. munzuricus, S. okumusi, S. platycephalus, S. opimus, S. rizeensis and S. tigridis see Turan et al. (2024), Küçük et al. (2024).

DNA extraction

The Hibrigen Genomic DNA isolation kit and a robot (King-Fisher Flex DNA extraction device, Thermo Fisher Scientific) were used to extract total DNA from fin clips. DNA quality metrics were checked on agarose gel electrophoresis and Nanodrop spectrophotometry (Thermo Fisher Scientific).

PCR and Sequencing for mtDNA analysis

Mitochondrial cytochrome b gene (Cyt b) (991 bp) was amplified using SsaL14437 (Warheit and Bowman 2008) and StrCBR (Turan et al. 2010) primer pairs. PCRs were carried out in a 50 μL reaction volume in the T100 thermal cycler (Bio-Rad, Hercules, CA, USA), including 100 ng of template DNA, 10X PCR buffer, 3 mM MgCl2, 5 μL of 0.5 mM dNTPs mix, 1U Taq DNA polymerase (Thermo Scientific Inc.) and 0.5 mM of each primer. PCR amplifications were conducted under the following conditions: initial denaturation for 2 min at 95°C, denaturation for 30 s at 95°C, annealing for 30 s at 56°C, extension 70 s at 72°C through 35 cycles and a final extension of 7 min at 72°C. The PCR products were run at 1% agarose gel and visualized under the UV Quantum–Capt ST4 system (Vilber Lourmat, France). Purification and sequencing of PCR products were performed by Macrogen Europa Inc. (Amsterdam, Netherlands).

Molecular Data Analysis of mtDNA

We used the newly generated eleven Cyt b sequences from the present study and 64 confirmed sequences from published studies deposited to NCBI GenBank for all phylogenetic analysis. (Crête et al. 2012; Tougard et al. 2018; Turan et al. 2020 2022 2024). We used the Clustal W algorithm (Thompson et al. 1994) in Bioedit v7.2.5 (Hall 1999) to align Cyt b sequences. Sequences submitted to NCBI GenBank with accession numbers PP884104–PP884114. The nucleotide substitution model TN93 model (Tamura and Nei 1993) was chosen as the best nucleotide substitution model according to the Bayesian information criterion (BIC) in Smart Model Selection (SMS) (Lefort et al. 2017) in PhyML software (Guindon et al. 2010). Phylogenetic relationships among species carried out maximum likelihood (ML) using PhyML 3.0 with 1000 bootstrap and Bayesian inference (BI) analysis using MrBayes 3.2 software (Ronquist et al. 2012). BI analysis ran using a Metropolis-coupled Markov chain Monte Carlo (MCMC) algorithm for one million generations, and the initial 25% of the saved trees sampled in each MCMC run were discarded as burnin. Salmo ohridanus (JX960763) was selected as an outgroup taxon for all phylogenetic analyses. Pairwise genetic distance estimation among the species was calculated by MEGA 11 (Tamura et al. 2021) software using the p-distance substitution model.

Collection codes

IFC-ESUF, Inland Fishes Collection, Faculty of Eğirdir Fisheries, Isparta University of Applied Sciences, Isparta; and FFR, Zoology Museum, Faculty of Fisheries, Recep Tayyip Erdoğan University, Rize.

Samples Information

In total, over 80 specimens were investigated however, morphological analyses (metric and meristic data collection) were performed on 21 formalin-fixed samples (details in paratypes and Table 2), while genetic analysis for mtDNA was carried out in 75 samples (details in molecular data analysis of mtDNA) and 25 ethanol fixed fin samples for whole genome analysis, respectively. In total, 11 samples were collected from the new species, Salmo sengulae sp. nov., inhabiting İvriz stream, Konya (closed basin) and used for the mtDNA analysis and two samples (of which one inhabits İvriz, Ereğli, and the second specimen inhabits Mersin Çamlıyayla) were analysed via whole genome DNA analysis. The rest of the samples included in the genome-wide analysis involved existing Salmo species inhabiting the Mediterranean basin of Anatolia (3 samples of Salmo chilo from Ceyhan River, Sivas, Akdere, three samples of S. kottelati originates from Alakır Stream, Antalya, Kumluca, three samples of S. labecula originates from Seyhan River Niğde, Çamardı, three samples of S. platycephalus from Seyhan River, Kayseri Pınarbaşı). In total, three samples of Salmo tigridis originates from Van Çatak (upper Botan) Euphrates-Tigris basin were used as an outgroup of Anatolian trout. Additionally, globally recognised reference trout species were included in the analysis followed as: three samples from Adriatic lineage (unidentified species level) originate from first two specimens from Kalamos in Greece and one specimen from Ohrid-Drin-Skadar in Albania, respectively, two samples of Danubian lineage, S. ischchan, originate from Armenia, two samples of S. obtusirostris originate from Bosnia and Herzegovina and one specimen was used as representative of Atlantic lineage originates from a hatchery at south France, origin of common commercial domestic Atlantic lineage of Salmo trutta (pers.comm. P. Berrebi).

Genome-wide DNA analysis

Accurate identification of double stranded DNA available for the next generation sequencing library construction was quantified on Qubit spectrophotometry (Thermo Fisher Scientific). The library generation protocol was followed by Turan et al. (2024) and the references therein. Briefly, standardized dsDNA samples were doubled-digested using EcoRI and MspI enzymes, individually adapter barcoded, size selected with a fragment window of 300–700 bp and amplified through PCR. Final library was assessed precisely on Qubit and sequenced on a short read platform sequencing by synthesis.

SNP calling

The first round of quality control was carried out from the raw reads produced and operated on FastQC (Andrews 2010). Low quality reads (Phred score < 30 with an average of 39 quality score), missing both restriction sites and/or mismatching barcodes were removed from the dataset. Process radtags module incorporated in Stacks v.2.55 (Catchen et al. 2013) was used for genotyping of the retained reads (-c -r -q --renz_1 eco-RI --renz_2 mspI) following they had been sorted into locus. Reference based mapping pipeline (ref_map.pl) was used to align reads to Salmo trutta genome assembly (acc. no: GCA_901001165.2; Hansen et al. 2021) using BWAmem (Li and Durbin 2010) (-k 19 -c 500 -O 0,0 -E 2,2 -T 0 -R) and samtools v1.11 (Li 2011) (-Sb -q 1 -F 4 -F 256 -F 2048). Afterwards, gstacks (--max-clipped 0.01) module was run with the following parameters: a minimum of two populations (-p 2), a minimum of 20% of individuals in a population (-r 0.2), a maximum observed heterozygosity of 60% (--max-obs-het 0.6), a minimum allele frequency of 1% (--min-maf 0.01), and a single representative of each overlapping site (--ordered -export). Following the previously outlined procedures, genotypes were gathered and then filtered using vcftools v0.1.16 (Danecek et al. 2011). The specific filters were applied for SNP calling: minimum stack depth of at least 30; maximum number of mismatches allowed at a locus in an individual; minimum allele frequency of %1 and maximum heterozygosity was kept under %60 to eliminate fixed heterozygote paralog sites (--minDP 4 --minGQ 30 --max-missing 0.4 --min-alleles 2 -max-alleles 2 --maf 0.01). As a final step, we further eliminated the SNP markers that were in high linkage disequilibrium using 11_extract_unlinked_snps_ genome.py (diff_threshold = 0.5 and max_distance = 50) from stacks_workflow v2.62 (https://github.com/ enormandeau/stacks_workflow). The bioinformatic analysis were performed under the guidelines from LDgenX (www.ldgenx.com). The dataset used in the present study is a part of a detailed investigation aims to evaluate Anatolian trout diversity which has not been published. Therefore, no availability is provided as of yet thus only a portion of the data was employed for the purpose of the present study.

RESULTS

Phylogenetic placement of Salmo specimens from Berdan (Mediterranean basin) and İvriz streams (Konya closed basin)

The resulting phylogeny indicates that the studied Salmo species are divided into six main clades: Adriatic, Danubian, Tigris, Atlantic, Mediterranean, and Marmoratus. Salmo sengulae sp. nov. is involved in Adriatic lineage with S. kottelati, S. chilo, S. ekmekciae, S. labecula, S. munzuricus, S. okumusi, S. baliki, and S. platycephalus. Salmo sengulae sp. nov. shares the same haplotype with S. chilo and Salmo kottelati (Fig. 1). The Bayesian and Maximum Likelihood analyses of Cyt b gene resulted in coherent trees supported by high bootstrap values. p-distance between species ranged from 0.00% (S. sengulae, S. kottelati, and S. chilo; S. fahrettini and S. abanticus; S. opimus and S. marmoratus; S. duhani and S. brunoi; S. euphrataeus and S. murathani) to 1.61% (S. tigridis and S. araxensis; S. tigridis and S. ardahanensis (Table S1).

Fig. 1.

Maximum likelihood (ML) tree based on mitochondrial cytochrome b gene sequences of Salmo species. Bayesian inference and ML analyses resulted in congruent trees. Bootstrap and posterior probability values are shown above nodes on tree if 50% or higher.

PCA data analysis

Salmo sengulae sp. nov. was compared using PCA analysis using 35 metric characters with the species S. kottelati and S. chilo. In total, 83 individuals [Salmo sengulae (n = 23), S. chilo (n = 42) and S. kottelati (n = 18)] were log standardized after the thirty-five metric characters obtained were proportioned to standard length. The results of the PCA confirmed that Salmo sengulae sp. nov. differs from the other two species (S chilo and S. kottelati). The plot indicated that the Salmo sengulae sp. nov. is unambiguously separated from S. chilo and S. kottelati (Table 1; Fig. 2). The first three extracted components explained 58.0% of the total variation among the examined samples, the first three components were considered due to higher eigenvalues, and these explained 27.85%, 16.05% and 14.14% of the total variance, respectively. The loadings on the first principal component (PC1) include eight metric characters (body width at anal-fin origin, body width at adipose-fin origin, length of dorsal-fin base, height of dorsal-fin, adipose-fin depth and length of adipose-fin base) (see Table 1, highlighted in bold font).

Table 1.

Character loadings on principal components I and II (PCI and PC II) for 35 measurements taken on 83 specimens of S. sengulae sp. nov., S. chilo and S. kottelati

Fig. 2.

A scatter plot of the scores of the first two principal components (PC I and PC II) for 83 specimens of 3 species Salmo sengulae sp. nov. (■:purple), S. chilo (+: red) and S. kottelati ( : blue) based on 32 morphometric characters.

SNP calling and Admixture analysis

Overall, an average of thirteen million raw reads were obtained per sample and mean sequence depth was kept over thirty. Alignment of Salmo trutta reference genome was 95.17% indicating high genotype calling accuracy. Once above indicated filtering applied 215,420 SNP markers obtained. As the accuracy of genetic predictions decrease with the proportion of linked markers (Ling et al. 2021), we further selected a distance of 50 bp unlinked markers resulted in 187,385 SNPs for the population structure analysis. The admixture results revealed 10 individual clusters presenting each species (Fig. 3, k = 10). The crossvalidation graph of Admixture analysis of the best K value is given in figure 4. All reference specimens and Anatolian outgroup (S. tigridis) clustered separately as expected corresponding to the species level difference and geographic origin. Salmo sengulae sp. nov. (Fig. 3 represented with the dark blue bars, Q = 0.99989), sampled respectively form İvriz, Ereğli and Mersin Çamlıyayla shared the same cluster indicating the distribution range for the new species, while being separated from rest of the Mediterranean trouts of Anatolia.

Fig. 3.

Admixture bar plots of ancestral origin retrieved from SNP data. Each vertical line represents a sample, with colours highlights the shared and/or separated ancestry of k = 10 groups. Salmo sengulae sp. nov., indicate a separate cluster thus presented as new species as well as mtDNA (Cyt b) and morphological analysis.

Fig. 4.

Cross validation graph of Admixture analysis indicating the best K value for the SNP data.

Salmo sengulae sp. nov.

(Figs. 5–7)

urn:lsid:zoobank.org:act:64C73A02-F774-4AD8-987E-22CF83EB3CC5

Holotype: FFR 3246, male, 282 mm SL; Türkiye: Mersin Prov.: Berdan Stream, Cehennemderesi Valley, 37°14'52.82"N 34°37'38.16"E.

Paratypes: FFR 1221, 5, 97–204 mm SL; same data as holotype. —IFC-ESUF 02-0026, 9, 131–202 mm SL; same data as holotype.; —IFC-ESUF 02-0025, 4, 125–203 mm SL; Türkiye: Mersin Prov.: Berdan Stream, Kadıncık Valley, 37°08'56.90"N 34°30'11.74"E. —IFC-ESUF 02-0023, 9, 96–163 mm SL; Türkiye: Konya Prov.: İvriz Stream, Delimahmutlu Village, 37°23'12.12"N 34°23'19.14"E. —FFR 03209, 5, 90–121 mm SL; Türkiye: Konya Prov.: İvriz Stream at Osmanköseli Village, 37°24'29.12"N 34°19'59.93"E.

Diagnosis: Salmo sengulae sp. nov. is distinguished from all the species of Salmo in Türkiye and adjacent areas by combination of following characters: Four dark bands on flank, one to five black spots (usually than pupil) in postorbital and suborbital areas, and one to eleven black spots (smaller than pupil) on opercle. Black spots on body large (always greater than pupil), few to numerous (9–113), scattered on back and upper part of flank, and usually whole flank in most specimens more than about 180 mm SL, furthermore their number increasing with increasing size. Red spots few (6–27), ocellated, organized in one or two irregular longitudinal rows on median part of the flank, their number decreasing with increasing size. No red spots in most specimens larger than about 180 mm SL, if present, red spots almost covered black dots. The red and black surrounded with a very large (greater than pupil) white ring. Eight to ten parr marks along lateral line. Maxilla long, length 11–12% SL in males and 8–10 in females. Lateral line with 108–117 scales; 20–24 scale rows between dorsal-fin origin and lateral line; 19–22 scale rows between anal-fin origin and lateral line; 13–15 scale rows between origin of the adipose fin and lateral line. The 16 to19 gill rakers on outer side of first gill arch.

Description: The general appearance is shown in figures 5–7, morphometric data are in table 2. Body moderately deep, compressed laterally, its depth markedly smaller than head length. The dorsal body profile is arched, ventral profile less arched than dorsal profile. The head long, length 1.1–1.6 times body depth at dorsal-fin origin. Upper profile of head slightly convex on the interorbital area and slightly concave at the nostril in males, and convex on both interorbital areas and on the snout in females. Mouth somewhat large, length of mouth gape 13–15% SL in males, 12–14% SL in females, slightly subterminal in males and females. Tip of lower jaw slightly curved upwards, pointed, with a very slightly developed process at symphysis in adult males. Maxilla somewhat long, length of 10–12% SL in males and 8–10% SL in females, reaching beyond the posterior margin of the eye in males larger than 130 mm SL and only reaching the posterior margin of the eye in females. Snout somewhat short, length 7–8% SL in males and females with a slightly pointed tip in males and females. Adipose fin small, height about 4–8% SL in males and in females. Known maximum size 282 mm SL.

Fig. 5.

Salmo sengulae sp. nov.: FFR 3246, holotype, male, 282 mm SL; Türkiye: stream Berdan.

Fig. 6.

Salmo sengulae sp. nov.: IFC-ESUF 02-0026 paratypes, males, a, 195 mm SL; b, 193 mm SL; c, 204 mm SL; Türkiye: stream Berdan.

Fig. 7.

Salmo sengulae sp. nov.: FFR 1221, paratypes, females, a, 204 mm SL; b, 146 mm SL; 97 mm SL; Türkiye: stream Berdan.

Table 2.

Morphometry of Salmosengulae sp. nov. (holotype, FFR 3246; paratypes FFR 3058, n = 5; IFC-ESUF 02-0023, n = 3; FFR 03209, n = 5; IFC-ESUF 02-0023, n = 8). The calculations include the holotype

Dorsal fin with 3–4 unbranched and 9–11 branched rays, its distal margin straight or slightly convex. Pectoral fin with 1 unbranched and 12–14 branched rays, its external margin straight or slightly convex. Pelvic fin with 1 unbranched and 8 branched rays, its external margin convex. Anal fin with 3 unbranched and 8 branched rays, its distal margin convex anteriorly and concave posteriorly. The caudal fin slightly emarginated, lobes slightly pointed. Lateral line with 108–117 scales; 20–24 scale rows between dorsal fin origin and lateral line; 19–22 scale rows between anal fin origin and lateral line; 13–15 scale rows between origin of the adipose fin and lateral line. Sixteen to 19 gill rakers on outer side of first gill arch.

Coloration: In formalin: General body color brownish or light brownish. Back and flank dark brown and belly light brownish. Four dark bands on flank. One to five black spots (usually smaller than pupil) in postorbital and suborbital areas, and one to eleven black spots (smaller than pupil) on opercle. Red spots few (6–27), not conspicuously, ocellated, organized in one or two irregular longitudinal rows on median part of the flank, their number decreasing with increasing size. No red spots in most specimens larger than about 180 mm SL, if present, red spots almost covered black dots. Number of red spots decreasing with increasing size. A conspicuously black spots in postorbital and suborbital areas. Black spots on body large (always greater than pupil), few to numerous (9–113), scattered on back and upper part of flank, and sometime whole flank in most specimens more than about 180 mm SL, their number increasing with increasing size. The red and black surrounded with a very large (greater than half of eye diameter) white ring. Pectoral, pelvic and anal fins yellowish, dorsal and anal fins yellowish or light brownish. Adipose fin without red spots or reddish margin. Eight to ten parr marks along lateral line.

In life: The general coloration of freshly preserved specimens greyish on the back and upper part of the flank, light greyish on the lower part of the flank and the belly (Fig. 8). Other color-related characters are the same as in the specimens fixed in formaldehyde.

Fig. 8.

Salmo sengulae sp. nov., not preserved: a, female, ~160 mm SL, (above); b, male, ~260 mm SL (below); Türkiye: stream Berdan.

Distribution and habitat: Salmo sengulae sp. nov. inhabits clear and swift-flowing water, with a substrate (Mediterranean Sea basin) and İvriz (Konya closed basin) (Fig. 9).

Fig. 9.

Type localities of native Turkish trouts and the distribution of Salmo sengulae.

Etymology: The species is named after Şengül Küçük, the beloved wife of Fahrettin Küçük, third author of this study.

Conservation status: According to our observations, Salmo sengulae sp. nov. is under the influence of overfishing, human activities for economic and recreational purposes. Given the highly restricted distribution of S. sengulae n. sp. to a very limited stating that “fishing is prohibited or should not be hunted because the fish are in danger of extinction”, and even to make conspicuous warning signs stating that the area is being monitored using CCTV cameras. Thus, there is also a need for the species to be conserved under international legislation.

Morphological differences and comparisons

Salmo sengulae sp. nov. is distinguished from the other species of trout recorded from the streams and rivers in Turkish Mediterranean coast (S. chilo, S. kottelati, S. labecula, S. ekmekciae and S. opimus and S. platycephalus) by the following characters: Salmo sengulae differs from S. chilo, shared the same mt DNA Cyt b haplotype, by having no red spots on flank in specimens larger than about 180 mm SL, if the red spot present in specimens larger than about 180 mm SL, they almost covered with black dots (vs. presence in all size, and red spots not covered with black dots), most of the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the ring around the red and black spots markedly greater than pupil (vs. mostly maller than pupil) the red and black spots roundish (vs. irregularly shaped), the ring around black and red spots markedly greater than pupil (vs. smaller than pupil), fewer parr marks along lateral line in juveniles (8–10, vs. 11–13), a longer maxilla in males (length 10–12% SL, vs. 8–10), a shorter predorsal length in females (46–47% SL, vs. 47–50), fewer scale rows between dorsal-fin origin and lateral line (20–24, vs. 24–26) and more scale rows between anal-fin origin and lateral line (19–22, vs. 15–18).

Salmo sengulae sp. nov. differs from S. kottelati,shared with same mt DNA Cyt b haplotype, by having no red spots on flank in specimens larger than about 180 mm SL, if the red spot present in specimens larger than about 180 mm SL, they almost covered with black dots (vs. presence in all size, and red spots not covered with black dots), the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the red and black spots roundish (vs. irregularly shaped), the ring around black and red spots markedly greater than pupil (vs. smaller than pupil), fewer scale rows between dorsal-fin origin and lateral line (20–24, vs. 24–29) and more scale rows between anal-fin origin and lateral line (19–22, vs. 17–19), a shorter predorsal length (44–48% SL in males, 46–47 in females; vs, 49–52 in males, 47–50 in females) and a shorter adipose fin in male (4–8% SL, vs. 8–10).

Salmo sengulae sp. nov. differs from S. labecula by having the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the red and black spots roundish (vs. irregularly shaped), the ring around black and red spots markedly greater than pupil (vs. smaller than half of eye diameter), fewer black spot behind eye and suborbital area (1, vs. 2–10), fewer parr marks along lateral line (8–10, vs. 11–12), a deeper caudal peduncle (11–12% SL, vs. 10–11), a longer maxilla in males (length 10–12% SL, vs. 8–10) and more scale rows between anal-fin origin and lateral line (19–22, vs. 16–17) and fewer gill rakers on first gill arch (16–19, vs. 21–23).

Salmo sengulae sp. nov. differs from S. ekmekciae by having the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the ring around black and red spots markedly greater than pupil (vs. smaller than pupil), a deeper caudal peduncle (11–12% SL, vs. 9–10), a longer maxilla in males (length 10–12% SL, vs. 8–10), a greater mouth gape in males (13–15, vs. 12–13) and more scale rows between anal-fin origin and lateral line (19–22, vs. 16–18) and fewer gill rakers on first gill arch (16–19, vs. 22–24) and more scale rows between the adipose fin origin and the lateral line (13– 15, vs. 11–13).

Salmo sengulae sp. nov. differs from S. opimus by having no red spots on flank in specimens larger than about 180 mm SL, if the red spot present in specimens larger than about 180 mm SL, they almost covered with black dots (vs. presence in all size, and red spots not covered with black dots), the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the red and black spots roundish (vs. irregularly shaped), the ring around black and red spots markedly greater than pupil (vs. markedly smaller than half of eye diameter), a longer the head in males (27–30% SL, vs. 25–28), a shorter predorsal length in females (46–47% SL, vs. 47–48), a slender adipose fin (height, 4–8% SL, vs. 8–9), a longer maxilla in males (length 10–12% SL, vs. 9–10) and a greater mouth gape in males (13–15% SL, vs. 11–13), fewer scale rows between dorsal-fin origin and lateral line (20–24, vs. 24–26).

Salmo sengulae sp. nov. differs from S. platycephalus by having fewer parr marks along lateral line (8–10, vs. 12–13), the presence of black spots in all size (vs. absent in specimens larger than about 200 mm SL), the black spots on flank equal or greater than eye pupil (vs. smaller than pupil), the head not flattened dorsoventrally (vs. flattened dorso-ventrally), and fewer gill rakers on first gill arch (16–19, vs. 23–25) and more scale rows between anal-fin origin and lateral line (19– 22, vs. 17–18).

Salmo sengulae sp. nov. is distinguished from S. baliki, S. okumusi and S. munzuricus by having fewer scale rows in lateral line and dorsal-fin origin (20–24, vs. 26–30), a longer head in males (27–30% SL, vs. 24– 27), a longer maxilla (length 10–12% SL, vs. 8–10), a smaller adipose fin in males (height 4–8% SL, vs. 8–12). It further differs from S. okumusi and S. munzuricus by having fewer parr marks on flank (8–10, vs. 10–14).

Salmo sengulae sp. nov. is also distinguished from other species (S. abanticus, S. araxensis S. ardahanensis, S. brunoi, S. coruhensis, S. duhani, S. euphrataeus, S. fahrettini, S. munzuricus, S. murathani, S. rizeensis, S. tigridis) by having the presence of four broad dark bands on flank (vs. absent), fewer parr marks on flank in juvenile (8–10, vs. 10–14), fewer scale rows lateral line and dorsal-fin origin (20–24, vs. 24–35).

Key to native Salmo species distributed in Türkiye

1a. Red spots larger than eye pupil ................... S. baliki

1b. Red spots smaller than eye pupil .................... 2

2a. There are four dark bands on flank ....................... 3

2b. There are no dark bands on body ....................... 12

3a. There are no red spots in adult specimens ..................... 4

3b. There are red spots in both juvenile and adult specimens ....................... 5

4a. The head flattened dorsoventrally; there are no black spots in specimens larger than 175 mm SL; there are numerous black dots on body ................................ S. platycephalus

4b. The head not flattened dorsoventrally; there are black spots in both juveniles and adults; there is no black dots on body ................................. S. labecula

5a. Length of maxilla more than 10% SL .................... 6

5b. Length of maxilla less than 10% SL ..................... 7

6a. Predorsal length 49–52% SL in males and 47–51 in females .................... S. kottelati

6b. Predorsal length 46–48% SL in males and 44–47 in females ................... S. sengulae

7a. There are 9–10 parr marks on flank in juvenile ..................... S. ekmekciae

7b. There 10–13 parr marks on flank in juvenile .................... 8

8a. 22–23 gill rakers on first gill arch; a large (lager than eye pupil) white ring around red spots ..................... S. opimus

8b. 18–21 gill rakers on first gill arch; a narrow (equal or smaller than eye pupil) white ring around red spots ................... 9

9a. There are 24–26 scales row between dorsal-fin origin and lateral line; 15–18 scale rows between anal-fin origin and lateral line .................... S. chilo

9b. There are 26–35 scales row between dorsal-fin origin and lateral line; 18–26 scale rows between anal-fin origin and lateral line ................... 10

10a. There are 32–35 scales row dorsal-fin origin and lateral line .................... S. tigridis

10b. There are 26–30 scales row dorsal-fin origin and lateral line ................... 11

11a. Lateral line with 103–112 scales .................... S. okumusi

11b. Lateral line with 116–123 scales ................... S. munzuricus

12a. There are no red spots on the body in specimens larger than 200 mm SL; shape of black spots on body polygonal ................... S. abanticus

12b. There are red spots on body of both juvenile and adult specimens; shape of black spots on body circular ...................... 13

13a. There are few black and red spots on body; number of black and red spots not increasing with size and age; black spots scattered on back and upper part of flank; one black spot behind eye ................... 14

13b. There are numerous black and red spots on body; number of black and red spots increasing with size and age; black spots scattered on back, upper part and middle part of flank; more than one spots behind eye in specimens larger than 230 mm SL .................... 17

14a. General body colour silvery in life; length of maxilla 9–10% SL in males .................... S. araxensis

14b. General body colour brownish or greenish in life; length of maxilla 10-12% SL in males ................... 15

15a. Distance between adipose fin and caudal fin bases in females 12–14% SL ..................... S. brunoi

15b. Distance between adipose fin and caudal fin bases 14–17% SL .................................. 16

16a. The head slightly compressed in adult male; anal-fin and adipose fins reaching to caudal-fin base ................... S. euphrataeus

16b. The head not compressed in adult male, anal-fin and adipose fins not reaching to caudal-fin base in adult males ................... S. rizeensis

17a. There are 19–23 gill rakers on first gill arch; adipose-fin reaching to caudal-fin base; head slightly flattened ................... 18

17b. There are 16–19 gill rakers on first gill arch; adipose-fin not reaching to caudal-fin base; head not flattened ..................... 19

18a. Pores on top of head with small black spots (smaller than pupil) ..................... S. ardahanensis

18b. There are no black spots on pores on top of head ................... S. murathani

19a. Dorsal-fin with 8–9 unbranched rays; upper profile of head markedly convex in males ................... S. fahrettini

19b. Dorsal-fin with 9–11 unbranched rays; upper profile of head straight or slightly convex ................... 20

20a. The number of spots conspicuously increasing with sizes in males ...................... S. coruhensis

20b. The number of spots not increasing with sizes in males ................... S. duhani

DISCUSSION

The Mediterranean basin has provided a refuge as being climatically stable for several species during glacial period thus regarded as biodiversity hot spot for the speciation (Lattanzi et al. 2024; Bianchi et al. 2012). As a result, species with different geographic origin are common in the region. Until recently, 6 species of trouts have been reported (Turan et al. 2012 2014b 2024; Küçük et al. 2024) in freshwaters draining into Mediterranean basin.

In this study molecular data revealed that Salmo sengulae sp. nov. belongs to the Adriatic lineage similar to existing Mediterranean trout. The only exception was observed in S. opimus grouped in Marbled lineage (S. marmoratus; Fig. 1). Given the Mediterranean basin’s refuge potential, it is a common phenomenon to find the existence of different geographic origin species in such regions (Bianchi et al. 2012). According to mtDNA (Cyt b) analysis Salmo sengulae sp. nov., S. kottelati and S. chilo grouped together. Yet, the lack of mtDNA resolution power was eliminated (grouping S. sengulae, S. kottelati and S. chilo together) with the use of large number of unlinked SNP markers genome-wide distributed.

At the time of genome-wide SNP analysis we did not have the samples of Salmo opimus (type locality of Alara Stream) and S. ekmekciae (type locality of Köprüçay Stream, reported recently by Küçük et al. 2024) both distributes in Mediterranean basin of Anatolia. As Salmo opimus belongs to Marbled lineage (Salmo marmoratus) different that those of rest Mediterranean trouts of Anatolia, this was treated as negligible for the SNP analysis.

According to our genome wide SNP markers, S. sengulae sp. nov., S. kottelati and S. chilo (Q = 0.99989, 0.999858, 0.99987) respectively separated from each other with high probability thus represented by the separate colour bars indicating unshared ancestral origin. The strategy applied within the Admixture analysis to involve S. tigridis as outgroup of Anatolian trout and globally known reference specimen provided separation among the existing Mediterranean trouts as well as new species, S. sengulae (Q = 0.99989). Admixture analysis results confirming the separation of S. sengulae from those of Mediterranean trouts of Anatolia. Furthermore, mtDNA (Cyt b) separation power (grouping S. sengulae, S. kottelati and S. chilo together) was empowered with the use of large number of unlinked SNP markers genome-wide distributed. Taken all together, integrated taxonomic investigation scheme has empowered our hypothesis, thus here we provide evidence that S. sengulae sp. nov is a new trout species of Anatolia.

CONCLUSIONS

This study combines morphological diagnostic criteria, mitochondrial DNA (Cyt b) analyses, and genome-wide SNP data to demonstrate that trout populations inhabiting the Berdan Stream (Mediterranean basin) and the İvriz Stream (Konya closed basin) constitute a distinct evolutionary lineage, thus here defined as Salmo sengulae sp. nov. Although S. sengulae shares the same Cyt b haplotype as S. chilo and S. kottelati based on mtDNA, comprehensive morphometric assessments and genome-wide SNP data demonstrated a clear and strong genetic separation from all Mediterranean trout species in Anatolia. The specieslevel distinctness of S. sengulae was validated by the high-resolution power 187,385 unlinked genome-wide SNP markers, which offered conclusive resolution in cases of which the discriminatory power of mtDNA alone was insufficient.

Morphologically, S. sengulae is distinguished from all other species by its unique combination of coloration pattern, meristic features, scale numbers, and maxilla and predorsal proportions, supporting its recognition as a distinct taxon. The species has an extremely narrow distribution, occurring in only two isolated river systems, making it particularly vulnerable to human activities such as overfishing, habitat destruction, and recreational pressures.

The discovery of S. sengulae sp. nov. highlights the still-understudied biodiversity of Anatolian inland waters regardless of all efforts and emphasises the need for integrated taxonomic approaches in regions with complex geological and climatic histories. The results highlight the urgency of conservation efforts such as habitat protection, enforcement of sportive fishing restrictions locally, and long-term population monitoring. In addition to protecting a recently discovered species, conservation of S. sengulae will also help preserve the ecological and evolutionary integrity of freshwater habitats in the Mediterranean and Anatolia.

Supplementary materials

Table S1.

Pairwise genetic distance of Salmo species distributed in Türkiye based on cytochrome b gene sequences.