Application of Exon Primed Intron Crossing Markers to Cross‐Amplify Oreochromis Species in Eastern Africa

TL;DR

Researchers developed genetic markers to study and conserve fish species in Eastern Africa, finding they can track genetic diversity and help with conservation efforts.

Contribution

The study introduces 50 EPIC markers for Oreochromis species, demonstrating their effectiveness in cross-amplification and biodiversity monitoring.

Findings

45 out of 50 EPIC markers successfully cross-amplified across four Oreochromis species.

Oreochromis niloticus showed the highest genetic diversity compared to other species.

AMOVA results confirmed significant genetic differentiation among species.

Abstract

Oreochromis species are of economic importance in fisheries and aquaculture but threatened by human‐mediated activities. Effective conservation and their sustainable management require genetic monitoring tools that can assess genetic variation across species. Various neutral markers have been used to monitor biodiversity in organisms, but they are limited in cross‐amplification among various taxa. Non‐neutral markers such as Exon‐Primed Intron Crossing (EPIC) not only cross amplify various taxa but also target gene regions that are likely to be involved in adaptive responses to selective pressure. This study therefore developed 50 EPIC markers from
Oreochromis niloticus
reference genome targeting immune related genes to assess their potential to cross‐amplify Oreochromis species. Genetic diversity, population structure, and differentiation was measured among
Oreochromis niloticus, O. jipe, O. esculentus and O. leucostictius species and their respective populations from eleven African water bodies and two aquaculture farms. 45 EPIC markers successfully amplified target regions across the four Oreochromis species.
O. niloticus
exhibited the highest diversity (He = 0.50), followed by
O. jipe
(He = 0.29),
O. leucostictus
(He =0.28) and
O. esculentus
with the lowest (He = 0.25). The highest differentiation values were observed among populations of different species while the lowest was among farmed
O. niloticus
. This was supported by AMOVA results that showed highest genetic differentiation among species. Population structure analyses were concordant with species boundaries while showing detailed intraspecific clustering patterns among the
O. niloticus
populations that reflected geographical origin. The observed genetic diversity, gene flow, and population differentiation confirms the applicability of EPIC markers as molecular tools for monitoring biodiversity and guiding the conservation of Oreochromis species under increasing anthropogenic and environmental pressures.

This study developed 50 exon‐primed intron‐crossing (EPIC) markers in Oreochromis niloticus to assess their cross‐amplification potential in four Oreochromis species. Forty‐five markers successfully amplified across species, revealing significant genetic differentiation among species and populations. These findings demonstrate the utility of EPIC markers for monitoring genetic diversity and informing conservation of Oreochromis species under anthropogenic and environmental pressures.

Figures

Map showing sample collection sites.

Bar plot of Oreochromis species showing observed (Ho) and expected (He) heterozygosity measures.

AMOVA output showing genetic differentiation among Oreochromis species.

PCoA analysis showing population differentiation in: a = Oreochromis species, b =  O. niloticus and O. jipe from Lake Kanyaboli, c = Oreochromis species from L. Kyoga, d =  O. esculentus from Lake Kanyaboli and Bisina.

PCoA analysis showing population differentiation in: a = Oreochromis populations, b =  O. niloticus populations (of Ethiopia, Kenya and Uganda), c =  O. niloticus from Uganda.

PCA loading plot shows the contribution of the 45 EPIC markers to genetic differentiation among Oreochromis species.

Genetic differentiation among Oreochromis populations as shown by Neighbor‐Joining network (a = species, b =  O. niloticus , c = Oreochromis populations), and UPGMA with bootstrap values for Oreochromis populations showing branch support (d).

The genetic structure of Oreochromis populations inferred using STRUCTURE analysis. K* = optimal number of clusters; the population codes: ABK = Albert, KNS = Kyoga, VMS = Victoria, KA = Kajjansi, KAY = Kyanamira, GKH = George, TU = Turkana, ON = Jipe, TA = Tana, H = Hashenge, CHA = Chamo, BSE = Kanyaboli, ESC = Bisina, JIPE = Jipe, NNL = Kyoga.

Tables

TABLE 1. Sample information for Oreochromis populations; water body, sample code, species, population status, number of samples per site, and geographic coordinates.

Water body	Sample code	Species	Status	Number of samples	Latitude	Longitude
Uganda
L. Albert	ABK	O. niloticus	Native	23	N 1.509900	E 30.93610
L. Kyoga	KNS	O. niloticus	Non‐Native	23	N 1.438730	E 32.86809
L. Victoria	VMS	O. niloticus	Non‐Native	23	N 0.436500	E 33.24081
L. George	GKH	O. niloticus	Native	23	N 0.017390	E 30.08698
L. Kyoga	NNL	O. leucostictus	Native	23	N 1.254444	E 33.34278
L. Bisina	BSE	O. esculentus	Native	24	N 1.600556	E 33.96139
Kajjansi	KA	O. niloticus	Non‐Native	24	N 0.221667	E 32.53444
Kyanamira	KAY	O. niloticus	Non‐Native	23	S 1.264122	E 30.03284
Kenya
L. Turkana	TU	O. niloticus	Native	34	N 3.556170	E 35.91599
L. Jipe	ON	O. niloticus	Non‐Native	12	S 3.511944	E 37.76194
L. Kanyaboli	ESC	O. esculentus	Non‐Native	12	N 0.073060	E 34.17331
L. Jipe	JIPE	O. jipe	Native	11	S 3.511944	E 37.76194
Ethiopia
L. Tana	TA	O. niloticus	Native	20	N 11.966667	E 37.30000
L. Hashenge	H	O. niloticus	Non‐Native	9	N 12.574604	E 39.49667
L. Chamo	CHA	O. niloticus	Non‐Native	12	N 5.821283	E 37.57472

TABLE 2. Heterozygosity (Mean ± SE) for Oreochromis species.

Populations	N	Na	Ne	I	Ho	He	uHe	F
O. niloticus	185.89 ± 4.16	12.07 ± 1.45	3.22 ± 0.40	1.17 ± 0.12	0.38 ± 0.05	0.50 ± 0.04	0.50 ± 0.04	0.26 ± 0.05
O. esculentus	26.33 ± 1.10	2.69 ± 0.29	1.56 ± 0.12	0.46 ± 0.07	0.30 ± 0.05	0.25 ± 0.04	0.26 ± 0.04	−0.11 ± 0.05
O. jipe	9.51 ± 0.41	2.42 ± 0.25	1.77 ± 0.19	0.52 ± 0.07	0.38 ± 0.06	0.29 ± 0.04	0.32 ± 0.04	−0.21 ± 0.06
O. leucostictus	21.02 ± 0.59	2.53 ± 0.27	1.73 ± 0.15	0.50 ± 0.08	0.34 ± 0.05	0.28 ± 0.04	0.29 ± 0.04	−0.17 ± 0.06
Total	60.69 ± 5.53	4.93 ± 0.49	2.07 ± 0.13	0.66 ± 0.05	0.35 ± 0.03	0.33 ± 0.02	0.34 ± 0.02	−0.03 ± 0.03

TABLE 3. Heterozygosity (Mean ± SE) for Oreochromis populations.

Species	Code	N	Na	Ne	I	Ho	He	F
Uganda
O. niloticus	ABK	21.04 ± 0.54	3.29 ± 0.34	1.92 ± 0.17	0.63 ± 0.09	0.38 ± 0.06	0.33 ± 0.04	−0.09 ± 0.05
	GKH	16.07 ± 0.40	3.29 ± 0.36	2.10 ± 0.22	0.67 ± 0.09	0.40 ± 0.06	0.35 ± 0.04	−0.23 ± 0.06
	KA	22.07 ± 0.61	4.07 ± 0.40	2.11 ± 0.22	0.75 ± 0.09	0.43 ± 0.05	0.37 ± 0.04	−0.21 ± 0.06
	KAY	20.58 ± 0.63	3.82 ± 0.36	2.24 ± 0.22	0.77 ± 0.09	0.43 ± 0.05	0.39 ± 0.04	−0.11 ± 0.03
	VMS	21.11 ± 0.55	3.31 ± 0.39	2.12 ± 0.21	0.69 ± 0.09	0.40 ± 0.05	0.36 ± 0.04	−0.19 ± 0.04
	KNS	21 ± 0.41	2.58 ± 0.19	1.80 ± 0.12	0.55 ± 0.07	0.37 ± 0.05	0.31 ± 0.04	−0.15 ± 0.06
O. esculentus	BSE	16.42 ± 0.79	2.16 ± 0.24	1.44 ± 0.12	0.36 ± 0.07	0.26 ± 0.06	0.20 ± 0.04	−0.18 ± 0.06
O. leucostictus	NNL	21.02 ± 0.59	2.53 ± 0.28	1.73 ± 0.15	0.50 ± 0.08	0.34 ± 0.06	0.28 ± 0.04	−0.04 ± 0.03
Kenya
	ON	10.38 ± 0.48	2.47 ± 0.30	1.99 ± 0.23	0.567 ± 0.09	0.37 ± 0.06	0.31 ± 0.05	−0.25 ± 0.05
O. niloticus	TU	15.51 ± 0.89	5.38 ± 0.65	2.73 ± 0.31	0.989 ± 0.11	0.49 ± 0.05	0.47 ± 0.04	−0.10 ± 0.04
O. esculentus	ESC	9.91 ± 0.41	2.13 ± 0.19	1.59 ± 0.12	0.457 ± 0.07	0.32 ± 0.05	0.27 ± 0.04	−0.17 ± 0.06
O. jipe	JIPE	9.511 ± 0.41	2.42 ± 0.25	1.77 ± 0.19	0.522 ± 0.08	0.38 ± 0.06	0.29 ± 0.04	−0.58 ± 0.05
Ethiopia
O. niloticus	TA	19 ± 0.49	1.6 ± 0.14	1.33 ± 0.60	0.24 ± 0.06	0.25 ± 0.06	0.15 ± 0.04	−0.12 ± 0.05
	H	8.51 ± 0.18	1.6 ± 0.12	1.28 ± 0.07	0.24 ± 0.05	0.22 ± 0.05	0.14 ± 0.03	−0.33 ± 0.07
	CHA	10.62 ± 0.38	1.96 ± 0.16	1.50 ± 0.11	0.39 ± 0.06	0.32 ± 0.05	0.24 ± 0.04	−0.04 ± 0.04
Total		14.84 ± 0.23	2.76 ± 0.08	1.82 ± 0.04	0.55 ± 0.02	0.36 ± 0.01	0.30 ± 0.01	0.19 ± 0.037

Full text

Introduction

1

Fisheries and aquaculture sectors provide a vital source of animal protein and support livelihoods globally, forming an integral part of food security, nutrition, and income generation (FAO 2024). In Eastern Africa, fish consumption remains below global averages due to persistent supply deficits, making aquaculture development and improved fisheries management essential for meeting future nutritional needs (Obiero et al. 2019). Among the diverse fish groups exploited in these sectors, tilapiines are central to inland capture fisheries and aquaculture because of their ecological versatility and economic value and play a key role in tropical ecosystems across Africa (Trewavas 1982; Pouyaud and Agnèse 1995; López‐Olmeda et al. 2021). Tilapiines belong to three main genera: Sarotherodon, Coptodon and Oreochromis (Trewavas 1983). Among these, Oreochromis species are highly adaptable to diverse environments, reproduce efficiently, and have robust feeding strategies, which have supported their global introductions (Cnaani and Hulata 2008; Prabu et al. 2019; El‐Sayed and Fitzsimmons 2023). Wild tilapia populations displayed high levels of genetic and ecological diversity (Ukenye and Megbowon 2023), which support their resilience in fluctuating and often challenging environments (O'Brien and Evermann 1988; Lively 2010; Reid et al. 2017; Ukenye and Megbowon 2023). Understanding the genetic structure and diversity of these populations is essential for fisheries management, conservation, and sustainable use of genetic resources. In Eastern Africa, wild fish populations are increasingly influenced by anthropogenic activities, including translocations, aquaculture introductions, and habitat modifications (Sanon et al. 2020; Yongo et al. 2021; Sadler et al. 2023). These processes can alter patterns of genetic diversity and connectivity among populations, potentially leading to erosion of locally adapted gene pools (Tibihika et al. 2019; Kwikiriza et al. 2024). These changes necessitate molecular tools that monitor population structure and genetic diversity and support sustainable aquaculture management and conservation of wild fish populations.

Neutral molecular markers such as microsatellites are widely applied to study population structure and genetic diversity in tilapiines (Angienda et al. 2011; Kwikiriza et al. 2024) due to their polymorphic nature (Mojekwu and Anumudu 2013). However, their application is often constrained by high cost for developing species specific markers (Olubunmi 2019). In addition, there is limited success in cross‐species amplification due to primer mismatches and sequence divergence (Moran 2002; DeWoody et al. 2006; Yue et al. 2010; Mojekwu and Anumudu 2013). These limitations restrict the applicability of traditional microsatellites for comparative studies across multiple taxa or populations lacking reference genomes.

Exon‐Primed Intron Crossing (EPIC) markers offer a robust alternative for population genetic studies. EPIC markers, unlike neutral microsatellites, are designed to amplify conserved exon regions flanking variable introns, allowing reliable cross‐species amplification while capturing sequence variability suitable for population‐level analyses (Li et al. 2010; Silva et al. 2017). EPIC markers enable evaluation of population structure, genetic differentiation, and admixture among populations and species through hybridization. Admixture has been reported among Oreochromis species in East African lakes (Angienda et al. 2011; Kwikiriza et al. 2024). Such admixture can challenge the maintenance of genetic diversity and affect population structure by introducing alleles from non‐native species, potentially compromising locally adapted gene pools (Lawlor and Hutchings 2004; Goldberg et al. 2005; McKinna et al. 2010; Bourret et al. 2011; Popovic and Bernatchez 2021).

In this study, EPIC markers were designed and developed from the O. niloticus reference genome. The aims of the study were to (1) develop and validate EPIC markers for cross‐species amplification, (2) characterize patterns of genetic variation and population structure, and (3) provide molecular tools to support future population genetic studies for conservation of Oreochromis species. By using EPIC markers, this approach overcomes limitations of species‐specific microsatellites and provides tools for comparative analyses across Oreochromis species, supporting conservation and resource management in Eastern Africa.

Materials and Methods

2

Description of Sample Collection Sites

2.1

This study utilized DNA of samples obtained from previous collections done by (Tibihika et al. 2019; Kariuki et al. 2021; Kwikiriza et al. 2023; Nyauchi et al. 2025), covering several lakes and fish farms across Eastern Africa (Figure 1). A total of 296 DNA samples consisting of 226 O. niloticus individuals, 36 O. esculentus , 11 O. jipe , and 23 O. leucostictus from 15 populations were used in this study (Table 1).

Map showing sample collection sites.

In Uganda, O. niloticus samples were collected from four lakes: Victoria, Kyoga, Albert, George while Singida tilapia ( Oreochromis esculentus ) henceforth O. esculentus and blue‐spotted tilapia ( Oreochromis leucostictus ), hence O. leucostictus were collected from Lake Bisina and Lake Kyoga respectively. Water and the associated flora and fauna in Lakes Victoria, Kyoga and Albert interact through River Nile (Green 2009). Lake George forms a connected system with Lake Edward within the western rift valley and together drain into Lake Albert through the Semliki River (Acere and Mwene‐Beyanga 1990). Meanwhile, Bisina is a satellite lake that is connected to Lake Kyoga through papyrus fringes (Ogutu‐Ohwayo et al. 1999).

In Ethiopia, O. niloticus was collected from three lakes: Tana, Chamo, and Hashenge. Lake Tana, situated in the northwestern highlands, primarily pours into the Blue Nile River and is the largest lake in Ethiopia (Vijverberg et al. 2009). Lake Chamo is part of a chain of shallow, high fish productivity Rift Valley lakes, located in the Ethiopian Southern Rift Valley (Golubtsov and Habteselassie 2010). Lake Hashenge is a small, shallow, endorheic (Haileselasie et al. 2012), dominated by O. niloticus and Tilapia zilli (Abreha Tesfaye et al. 2017).

In Kenya, O. niloticus was collected from lakes Turkana and Jipe, while O. esculentus was collected from Lake Kanyaboli and Jipe tilapia ( Oreochromis jipe ); hence, O. jipe was collected in Lake Jipe. Lake Kanyaboli is part of Lake Victoria basin, located in the Yala swamp complex in western Kenya, adjacent to the northeastern shore and serving as key habitats for threatened tilapiine species (Abila et al. 2008). Lake Jipe is a shallow transboundary lake situated on the border of Kenya and Tanzania, lying at the foot of the Pare Mountains and within the upper Pangani river drainage, fed by inflows from river Lumi and local springs (Ngugi et al. 2015; Orina et al. 2023, 2024). O. jipe is a critically endangered Oreochromis species endemic to Lake Jipe (Orina et al. 2023).

In addition, fish farms including Kajjansi and Kyanamira represented domesticated fish populations. The Aquaculture Research and Development Centre (ARDC), Kajjansi, is a satellite of National Fisheries Resources Research that is solely responsible for conducting aquaculture adaptive research in Uganda. The centre maintains genetically managed stocks of O. niloticus for research on selective breeding, nutrition, disease management, and production performance. Kyanamira Fish Facility is a zonal aquaculture research centre that carries out basic research under Kachwekano Zonal Agricultural Research and Development Institute. These sampling sites were selected based on ecological importance, presence of target species, and representation of both native and introduced populations.

EPIC Marker Design, Singleplex PCR, and Multiplex Amplicon Genotyping

2.2

Epic Marker Design

2.2.1

To design EPIC markers used in this study, O. niloticus genome was searched in the National Center for Biotechnology Information (NCBI) for immune genes. For each identified gene, the corresponding mRNA sequence was compared to the genome using NCBI‐BLASTn (Altschul et al. 1990) and implemented in Geneious version 2024.7 (Kearse et al. 2012) to identify exon positions and the intervening intron regions. Primers were designed using the default settings of the Primer3 program (Untergasser et al. 2012) with the following adjustments: optimal primer length between 18 and 22 bp, GC content 40%–60%, expected PCR amplicon sizes ranging from 370 to 450 bp and an optimal melting temperature (Tm) of 55°C.

Singleplex PCR

2.2.2

The suitability of primers for downstream applications was assessed; each primer pair was first evaluated in singleplex polymerase chain reactions. Forward and reverse primers (100 μM) were diluted 1:100 (2 μL primer pair +98 μL sterile distilled water), vortexed for 20 s, and centrifuged at 2000 rpm for 2 min. Template DNA from three Oreochromis niloticus samples obtained from Kwikiriza et al. (2023) was diluted 1:10. PCR reactions were prepared by mixing 5 μL of Qiagen Multiplex PCR Master Mix (2×) with 1 μL of DNA, followed by the addition of 4 μL of primer mix to reach a final volume of 10 μL. Reactions were vortexed briefly, centrifuged at 2000 rpm for 2 min, and run in a thermal cycler under the following conditions: initial denaturation at 95°C for 15 min; 35 cycles of 95°C for 30 s, annealing at 55°C for 1 min, and extension at 72°C for 1 min; followed by a final extension at 72°C for 10 min, then cooled to 10°C. PCR products were visualized on a 1.8% agarose gel prepared in 1× TAE buffer with 1.5 μL HDgree+ dye. Gel samples were prepared by mixing 1 μL PCR product with 4 μL loading buffer (1:4). Electrophoresis was performed at 80 V for 15 min using an Enduro Power Supply (Labnet International), and gels were visualized using an Intas GDS imaging system (Japan). Primer pairs that produced clear and specific bands were selected for downstream analysis, resulting in 50 out of 65 primers meeting the target amplicon length of 450 bp.

The best multiplex combinations of primer pairs that successfully amplified were determined using primer pooler (https://ssb22.user.srcf.net/pooler/) with default settings. The following tags were added to primers: forward: TCTTTCCCTACACGACGCTCTTCCGATCT and reverse: CTGGAGTTCAGACGTGTGCTCTTCCGATCT. The primers were pooled into four (4) groups with a minimum of 10 pairs per combination.

Multiplex Amplicon Genotyping

2.2.3

We prepared libraries for amplicon sequencing according to the method of Kwikiriza et al. (2024) and Tibihika et al. (2019). The main change was the fact that reaction mixtures for the first PCR were prepared in a robotic liquid handling station (BRAND, Wertheim‐ Germany) in a 384 PCR plate and the volume of the PCR reaction was 5.0 μL containing 3 μL DNA (diluted 1:4) and 2 μL of mixture comprised of primer mixes (60 + 300 μL). PCR amplifications were carried out using the Qiagen Multiplex PCR Kit (Qiagen, Netherlands). The resulting amplicons were combined and purified with NucleoMag magnetic beads (Macherey‐Nagel) by mixing 2.9 μL of beads with 4 μL of pooled PCR product. The mixture was incubated for 5 min at room temperature, after which DNA–bead complexes were isolated using a VP 407‐AM‐N inverted magnetic bead extractor (V&P Scientific Inc., San Diego, USA). Beads were washed twice with 100 μL of 80% ethanol for 45 s, air‐dried for 5 min at room temperature, and DNA was eluted in 30 μL of elution buffer (10 mM Tris–HCl, pH 8.3, preheated to 65°C). Library indexing followed the procedure described by Tibihika et al. (2019). Indexed libraries were then pooled and sequenced on an Illumina MiSeq platform (paired‐end, 300 bp reads) at the Genomics Service Unit, Ludwig Maximilian University, Munich, Germany.

Population Genetic Analysis

2.3

Sequence quality was first evaluated by trimming low quality reads in Trimmomatic software version 0.39 (Bolger et al. 2014) and USEARCH version 11.0 (Edgar 2010) for merging forward and reverse sequences. The cleaned reads were processed using the SSR‐GBS pipeline (https://github.com/mcurto/SSR‐GBS‐pipeline) by Curto et al. (2019). This workflow was originally designed for SSR markers (Curto et al. 2019; Tibihika et al. 2019) but was subsequently adapted to accommodate EPIC markers. The pipeline provides a genotype matrix used in subsequent analysis.

We inspected the codominant genotype matrix in Microsoft Excel and removed both loci and samples that contained excessive missing data. Specifically, samples with more than 50% missing genotypes and markers that failed in more than 50% of the samples were excluded from subsequent analyses. This threshold was chosen because it provided results comparable to stricter cutoffs while maximizing the number of retained samples and loci.

Genetic diversity parameters were estimated in GenAlEx v6.5 (Peakall and Smouse 2006), including observed and expected heterozygosity, allelic richness, effective allele number, pairwise genetic distance, and Wright's fixation indices (F ST). To examine population relationships, Principal Coordinates Analysis (PCoA) was performed in GenAlEx using pairwise genetic distances, while a neighbor‐joining network based on Nei's distance was generated in SplitsTree v4 (Huson and Bryant 2006). Branch support was evaluated with 1000 bootstrap replicates using the Populations tool v1.2.32 (Langella 1999).

Population structure was further analyzed through Bayesian clustering in STRUCTURE v2.3.4 (Pritchard et al. 2000), applying an admixture model with correlated allele frequencies. The analysis considered values of K ranging from 1 to 15 and 1 to 11 for Oreochromis populations and O. niloticus respectively, each replicated 15 times with a burn‐in of 10,000 followed by 100,000 MCMC iterations. The optimal K was inferred from the ΔK method using Structure Harvester tool (Li and Liu 2018).

Principal Component Analysis (Hotelling 1933) was performed in R studio version 4.5.0 using PCA‐maker Shiny application (https://www.r‐bloggers.com/2021/12/pca‐maker‐shiny‐app/) to reveal markers that strongly contributed to population structure among Oreochromis species.

Results

3

Amplification Success of EPIC Markers

3.1

A total of 50 EPIC markers produced high‐quality amplification in O. niloticus and other Oreochromis species (Table S1). Illumina sequencing yields a total of 38,909,070 paired reads across all samples, of which 37,303,579 successfully met quality control standards and were retained for genotyping. Five markers (Til11‐HCIIEPIC, Til17‐catBEPIC, Til44‐C4EPIC, Til49‐TRIFEPIC, and Til53‐cd40lgEPIC) and 24 samples showed a missing data rate above 50% and were filtered out, yielding a final matrix consisting of 45 markers and 272 individuals.

Genetic Diversity Among the Oreochromis Species and Populations

3.2

Interspecific difference in genetic diversity was evident across the studied populations. Generally, overall heterozygosity values in species (He = 0.33 & Ho = 0.35, Table 2) and populations (He = 0.3 & Ho = 0.36, Table 3) were moderate. Among the species, O. niloticus showed higher heterozygosity (He = 0.50) than other species ( O. jipe ; He = 0.29, O. leucostictus ; He = 0.28 and O. esculentus ; He = 0.25, Figure 2, Table 2). Among the wild populations, O. niloticus from lake Turkana (He = 0.49) in Kenya showed the highest heterozygosity compared to Lake Victoria (He = 0.40) in Uganda and Lake Hashenge (He = 0.14) in Ethiopia (Table 3). Across lakes, the heterozygosity in O. esculentus from lake Kanyaboli (He = 0.27) was higher than individuals from Lake Bisina (He = 0.20). Within the Victoria basin, Lake Victoria had a higher diversity (He = 0.40) than Kajjansi (He = 0.37) (Table 3). Between the farmed populations, the heterozygosity in Kyanamira (He = 0.39) was higher than Kajjansi (He = 0.37, Table 3). Among the Ethiopian populations, heterozygosity varied with Chamo (He = 0.24) having the highest followed by Tana (He = 0.15) and Hashenge (He = 0.14) with the least value (Table 3). The remaining diversity metrics showed similar trends. The overall fixation index (F) was negative across populations (Table 3).

Bar plot of Oreochromis species showing observed (Ho) and expected (He) heterozygosity measures.

Genetic Differentiation Among Oreochromis Species

3.3

Pairwise genetic differentiation (F ST) among the Oreochromis populations ranged between 0.02 and 0.73 (Table 4). As expected, the highest values were observed among populations of different species. Of note, there was high differentiation between the O. jipe and O. niloticus in Lake Jipe (F ST = 0.57) and O. niloticus and O. leucosticus in Kyoga (F ST = 0.56). At the intraspecific level, both populations of O. esculentus (Kanyaboli and Bisina) had a moderately low genetic differentiation (F ST = 0.12). Overall, populations of O. niloticus exhibited low differentiation among them, particularly between Victoria and Kajjansi (F ST = 0.03), Kajjansi and Kyanamira (F ST = 0.03), Kyanamira and Albert (F ST = 0.03), Kyoga and Victoria (F ST = 0.04), and Albert and George (F ST = 0.07). A moderate differentiation was observed between Hashenge and Chamo (F ST = 0.10) and a very high differentiation between Chamo and Tana (F ST = 0.5) and Tana and Hashenge (F ST = 0.58).

TABLE 4: Pairwise Nei's Genetic Distance by pairwise FST values among Oreochromis populations.

Analysis of Molecular Variance demonstrated high species‐level divergence, with 47% genetic variation attributed to differences among species; 31% differentiation among individuals of the same population was observed. Among populations, 17% differentiation of the same species was indicated, and 5% differentiation among individuals not explained by either population or species groupings (Figure 3).

AMOVA output showing genetic differentiation among Oreochromis species.

Population Structure Among the Oreochromis Species and Their Respective Populations

3.4

Principal Coordinates Analysis revealed distinct clustering patterns. Considering all species, four major clusters were observed concordant with species boundaries (Figure 4a and Figure 5a). Zooming in on the O. niloticus population, five major distinct clusters were observed: (1) Chamo and Hashenge, (2) Turkana, (3) Jipe, (4) Tana, and (5) Ugandan populations (Figure 5b). PCoA of O. niloticus from Ugandan water bodies didn't reveal any distinct clustering pattern (Figure 5c). Further PCoA analysis on O. esculentus in Bisina and Kanyaboli indicated a separation between the two populations (Figure 4d). Individuals of different species co‐occurring in the same waterbody clustered according to species boundaries (Figure 4b,c). Principal Component Analysis (PCA) loadings indicate more than 50% of the EPIC markers contributed to the observed genetic variation in Oreochromis species (Figure 6, Table S2). The first two principal components accounted for 36% of the total genetic variation among Oreochromis species (Figure 6).

PCoA analysis showing population differentiation in: a = Oreochromis species, b =  O. niloticus and O. jipe from Lake Kanyaboli, c = Oreochromis species from L. Kyoga, d =  O. esculentus from Lake Kanyaboli and Bisina.

PCoA analysis showing population differentiation in: a = Oreochromis populations, b =  O. niloticus populations (of Ethiopia, Kenya and Uganda), c =  O. niloticus from Uganda.

PCA loading plot shows the contribution of the 45 EPIC markers to genetic differentiation among Oreochromis species.

The Neighbor‐Joining analysis (Figure 7a) and UPGMA (Figure 7d) show a clustering pattern that closely reflects the results of the PCoA analysis (Figure 4 and Figure 5). Species were distinctly separated from each other (Figure 7a). The O. niloticus from Uganda population formed one cluster, O. niloticus from Lake Turkana and O. jipe from Jipe formed independent clusters, while O. niloticus from Lake Chamo and Hashenge clustered together independent from Lake Tana population (Figure 7b,c, Figure 8).

Genetic differentiation among Oreochromis populations as shown by Neighbor‐Joining network (a = species, b =  O. niloticus , c = Oreochromis populations), and UPGMA with bootstrap values for Oreochromis populations showing branch support (d).

The genetic structure of Oreochromis populations inferred using STRUCTURE analysis. K = optimal number of clusters; the population codes: ABK = Albert, KNS = Kyoga, VMS = Victoria, KA = Kajjansi, KAY = Kyanamira, GKH = George, TU = Turkana, ON = Jipe, TA = Tana, H = Hashenge, CHA = Chamo, BSE = Kanyaboli, ESC = Bisina, JIPE = Jipe, NNL = Kyoga.*

Levels of Admixtures Among the Oreochromis Species and Their Populations

3.5

The Bayesian structure analysis revealed clear patterns of genetic clustering among the sampled Oreochromis populations. The optimal K values for analysis including all individuals was K8, followed by K10 and K12 (Figure 8). Structure analysis considering all individuals indicated seven clusters with admixture observed among the populations of Kajjansi, Kyanamira, and Turkana, from O. niloticus (Figure 8). Clustering was overall congruent with PCoA with no admixture among species (Figure 4a). Structure analysis was performed in O. niloticus to further show admixture levels. The optimal K values for O. niloticus were K9, followed by K7 and K2 (Figure S1). Five distinct groups were observed: (1) Chamo and Hashenge; (2) Albert, Kajjansi, Kyanamira, Kyoga, George and Victoria; (3) Jipe; (4) Tana; and (5) Turkana (Figure S1). This similarity is observed in Neighbor joining output (Figure 7b).

Discussion

4

The Oreochromis species in Eastern Africa water bodies continue to be threatened by various anthropogenic activities especially overfishing, habitat destruction, climate change and pollution (Eknath and Hulata 2009; Kwikiriza et al. 2023). These pressures have subsequently led to reduced fish stocks and sometimes extinction which has affected genetic diversity. Therefore, effective management for conservation of important traits and selective breeding strategies to sustain resilient stocks is very important. Molecular tools are used to monitor population structure and genetic diversity in Oreochromis species in East Africa (Tibihika et al. 2019; Kariuki et al. 2021; Kwikiriza et al. 2024). Amidst this, no study has applied EPIC markers to assess population structure of the Eastern Africa Oreochromis species. Therefore, Exon regions result in much more conserved primer binding sites than non‐coding microsatellites, explaining low drop out in the current study making EPIC markers a better fit for intraspecific comparisons (Li et al. 2002, 2010). In the current study, forty‐five (45) EPIC markers successfully cross amplified Oreochromis species in Eastern Africa which five markers exhibited amplification failure rates exceeding 50%. This is relatively high success rate especially when compared with previous analogous studies using microsatellites to compare the same set of species and populations (Kariuki et al. 2021; Kwikiriza et al. 2024). Amplification failures are influenced by several factors including, sequence variation, marker selection, intron length and mismatch of primer binding sites (Dudov and Perry 1984; Quattro et al. 2001; Li et al. 2007; Thomson et al. 2010).

Genetic Diversity and Population Structure of Oreochromis Species

4.1

Genetic Diversity of Oreochromis Species

4.1.1

Existence of genetic variation in the populations is highly important for the adaptation to environmental stressors and survival of a species (Bezault et al. 2011). Many factors influence genetic diversity among populations. Notably, hybridization, sample size, effective population size and environmental adaptations (Li et al. 2009; Mwanja et al. 2010; Bezault et al. 2011; Abdel‐kader et al. 2013; Tibihika et al. 2019; Kwikiriza et al. 2024). In addition, Wahlund effect, inbreeding, isolate‐breaking effect and admixture in populations also contribute to genetic diversity (Garnier‐Géré and Chikhi 2013; Kanaka et al. 2023). In this study, the unbiased heterozygosity estimate was 0.50, 0.26, 0.32 and 0.29 for O. niloticus , O. esculentus , O. jipe and O. leucostictus respectively (Table 2). The high unbiased heterozygosity in O. niloticus indicates that O. niloticus maintains higher genetic diversity than the other Oreochromis species. In addition, a small number of effective alleles was exhibited in the three species, O. jipe (Ne = 1.77), O. esculentus (Ne = 1.56), O. leucostictus (Ne = 1.73) (Table 2). In a study by Yoshida et al. (2019) using SNP markers to assess Latin American Oreochromis niloticus breeding populations, reported that the observed heterozygosity was lower than the expected heterozygosity and they attributed it to founder effects and small effective population sizes. Founder effects and small effective population sizes reduce heterozygosity through the loss of alleles and increased genetic drift (Glover 2010; Yoshida et al. 2019).

In addition, Castric et al. (2002) found a very strong correlation between small lake size and reduced diversity due to a higher probability of mating with relatives when compared with fish in larger lakes. In this study, the low genetic diversity in O. jipe , O. esculentus , and O. leucostictus is attributed to the relatively small lake sizes that tend to support small and often isolated fish populations (Ogutu‐Ohwayo et al. 1999; Abila 2005). In addition, intense fishing pressure further accelerates population declines (Smith et al. 1991; Kariuki et al. 2021).

Population Structure of Oreochromis Species

4.1.2

Multivariate analyses, Principal Coordinates Analysis (PCoA), Neighbor Joining network (NJ), Unweighted Pair Group Method with Arithmetic Mean (UPGMA), and STRUCTURE revealed similar clustering patterns, forming four main clusters. Each species formed a distinct cluster; however, O. niloticus populations formed subclusters, reflecting the fact that samples were obtained from different water bodies. At the species level, O. esculentus from Lake Bisina and Lake Kanyaboli clustered together; however, separate PCoA analyses for each population revealed distinct genetic patterns corresponding to their respective lakes.

The distinct species‐level clustering observed across these species and their respective populations suggests reproductive isolation among the studied Oreochromis species (Angienda et al. 2011). This is further supported by the clear separation of O. niloticus and O. leucostictus from Kyoga, and O. jipe and O. niloticus from Lake Jipe, which demonstrates that these sympatric species maintain distinct gene pools, likely due to ecological niche partitioning (Nyingi et al. 2009). Oreochromis niloticus primarily feeds on phytoplankton and dominates in open water rich in algal biomass (Ndiwa 2014; Kwikiriza et al. 2023). While O. esculentus and O. leucostictus tend to occupy more sheltered niches characterized by papyrus fringes, shallow muddy bays, and lake inlets (Twongo 1995; Laurent et al. 2020). Similarly, O. jipe is associated with shallow, vegetated littoral habitats featuring reed beds and papyrus fringes that offer refuge and breeding sites (Orina et al. 2023). These distinct habitat preferences among Oreochromis species might have contributed to ecological separation.

The consistent separation of O. niloticus populations into clear regional clusters further suggests that natural geographic barriers and historical isolation have maintained population structure in Turkana, Jipe, Tana, Chamo, and Hashenge (Bezault et al. 2011; Tesfaye et al. 2021). The study by Tesfaye et al. (2021) on the genetic diversity of O. niloticus populations in Ethiopia, using nuclear DNA microsatellites, provides strong evidence that distinct O. niloticus populations are structured according to their geographic locations. The authors found clear genetic differentiation among populations from different water bodies, supporting the idea that natural geographic barriers have historically limited gene flow between populations.

Further, the lack of distinct clustering among Ugandan O. niloticus populations indicates substantial gene flow or a shared origin, likely reflecting widespread translocations, stocking, and possible mixing of hatchery and wild stocks (Ogutu‐Ohwayo et al. 1997; Mwanja et al. 2016). The slight separation of the Lake George O. niloticus population from the main Ugandan cluster in the NJ analysis suggests subtle local differentiation that may be due to restricted gene flow or local adaptation (Angienda et al. 2011). The low genetic differentiation between O. esculentus of Lake Kanyaboli and Bisina, despite their different geographical locations, can be attributed to translocations or a shared stocking source that has maintained similar gene pools (Abila 2005). In previous work investigating genetic structure patterns of O. niloticus in Uganda using microsatellites, the genetic structure among some of the studied populations was much more pronounced (Tibihika et al. 2019). Thus, the low intraspecific structure found in this study may be a result of the more conserved nature of EPIC markers, which are less variable than microsatellites and may therefore overlook some intraspecific differentiation patterns.

Population Differentiation and Admixture Levels Among Oreochromis Species

4.2

Genetic Differentiation in Oreochromis Species

4.2.1

Genetic differentiation quantifies the extent of allele frequency divergence among populations or species and provides insights into evolutionary processes such as gene flow, drift, and selection (Holsinger and Weir 2009). Genetic differentiation among populations is commonly quantified using F‐statistics, particularly the fixation index (F ST), which ranges from 0 to 1 and measures the proportion of genetic variation that can be attributed to population differences (Weir and Cockerham 1984; Holsinger and Weir 2009). Wright (1978) categorized F ST values into four levels of genetic differentiation: < 0.05 (low), 0.05–0.15 (moderate), 0.15–0.25 (high), and > 0.25 (very high). In the present study, pairwise F ST values ranged from 0.02 to 0.73. Genetic differentiation was considered low at F ST values of ≤ 0.09, moderate at 0.10–0.30, high at 0.31–0.40, and very high at values above 0.40, with these categories reflected in the light‐to‐dark color patterns.

Similar studies have reported a wide F ST range; for instance, Romana‐Eguia et al. (2004) observed F ST values ranging from 0.004 to 0.007 among farmed O. niloticus and red tilapia hybrids in the Philippines, attributing this low differentiation to the use of O. niloticus as the founder stock in the breeding program. Likewise, Vähä et al. (2007) identified within‐river subpopulations of Atlantic salmon exhibiting average F ST values as low as 0.09, which they attributed to substantial gene flow between populations. Similarly, Kwikiriza et al. (2025) reported low to moderate genetic differentiation (F ST = 0.04–0.40) among farmed O. niloticus populations, consistent with mixing and frequent stock movements. Such patterns may reflect recent gene flow, admixture, or non‐random mating practices (Kwikiriza et al. 2024).

All comparisons among populations and between species showed either very high genetic or high differentiation values highlighting substantial reproductive isolation. Nevertheless, similar F ST ranges were also obtained for some comparisons among O. niloticus populations from Ethiopia. Results by Tesfaye et al. (2021) also found high level of genetic differentiation among the Ethiopian O. niloticus . They attributed the high genetic differentiation to possible existence of sub‐species. Contrasting with these results are the low F ST values among O. niloticus populations from Uganda. These ongoing gene flow caused both by natural water body connectivity and extensive humanmediated translocations known for the region. Meanwhile, O. esculentus populations from Kanyaboli and Bisina showed moderately low genetic differentiation, suggesting partial isolation.

The analysis of molecular variance (AMOVA) showed high genetic variation occurred among species and populations of the same species rather than within them. This pattern is consistent with previous studies showing that tilapiine species often exhibit clear species boundaries and a pronounced spatial structure shaped by geographic barriers and historical isolation yet maintain high within‐population (Romana‐Eguia et al. 2004; Tesfaye et al. 2021). This is supported by clear structure patterns retrieved by F ST and clustering analysis results.

Admixture in Oreochromis Species

4.2.2

Oreochromis niloticus populations from Lakes, Turkana, Victoria and farms (Kyanamira and Kajjansi) indicated higher levels of admixture. O. esculentus , O. jipe and O. leucostictus did not show any admixture. The high admixture among O. niloticus indicates potential historical gene flow among populations. These can reflect cumulative impact of human‐mediated movements, including deliberate stocking for aquaculture and fisheries enhancement (Kwikiriza et al. 2024). Kajjansi and Kyanamira, being major aquaculture centres in Uganda, serve as a source of fish seed, increasing the risk of mixing locally adapted gene pools (Matthew et al. 2015). The limited admixture observed in Albert, George, and Kyoga suggests these populations remain relatively undisturbed by translocations and retain more distinct local genetic identities (Ndiwa 2014). The detection of admixture signals in isolated basins like Turkana could be due to connectivity with the Omo river system and possible translocation from tilapia regional stocking programs (Shechonge et al. 2019).

The absence of admixture in O. jipe , O. esculentus , and O. leucostictus is linked to geographic isolation, strong reproductive barriers and limited gene flow with sympatric species. For example, phylogeographic studies have shown that O. jipe populations are confined to the Pangani basin, with minimal connectivity to other catchments (Shechonge et al. 2019). Similarly, O. leucostictus , although widely translocated, often maintains a separate gene pool when co‐occurring with O. niloticus due to differences in habitat preference and breeding behavior (Angienda et al. 2011; Ndiwa 2014). Likewise, populations in Chamo, Hashenge, and Tana show limited connectivity to other water bodies, which likely contributes to the absence of admixture and the maintenance of local genetic distinctiveness (Tesfaye et al. 2021).

Applicability of EPIC Markers for Conservation

4.3

The Exon Primed Intron Crossing (EPIC) markers used in this study demonstrated high cross‐species amplification, enabling comparative analyses across O. niloticus , O. jipe , O. esculentus , and O. leucostictus . This study provides one of the first comprehensive applications of EPIC markers to assess genetic diversity and population structure among African Oreochromis species. The successful amplification of 45 out of 50 markers across multiple species highlights their value for both cross‐species amplification and functional relevance in non‐model taxa (Chow and Hazama 1998; Hassan et al. 2002; Jennings and Etter 2011; Curto et al. 2012; Chow Seinen et al. 2015). Similar studies have demonstrated the utility of EPIC markers for cross‐amplification and detecting genetic differentiation across diverse taxa. For example, Jennings and Etter (2011) developed EPIC markers for protobranch bivalves and successfully cross‐amplified them in various invertebrate species. Likewise, Silva et al. (2017) designed EPIC primers for Lutjanus purpureus and achieved successful cross‐amplification in other marine teleosts, including lutjanids, sciaenids, and anablepids, with six out of eight markers amplifying consistently and revealing clear patterns of genetic differentiation.

EPIC markers are designed from conserved exon regions flanking introns (Chow and Hazama 1998; Hassan et al. 2002), allowing them to target variable intronic sequences (Chow and Takeyama 1998; Li et al. 2010; Chow Seinen et al. 2015), supporting genetic differentiation analysis (Li et al. 2010; Chenuil et al. 2010). EPIC markers are more likely to behave as non‐neutral when compared for example with microsatellites. Unlike neutral markers that mainly describe background population structure (Funk et al. 2012), non‐neutral markers capture adaptive variation and evolutionary potential shaped by selective pressures (Holderegger et al. 2006; Kirk and Freeland 2011).

Furthermore, the clear population structure revealed by EPIC markers reflects both geographic factors and human‐mediated influences, demonstrating their relevance for conservation genetics, aquaculture management, and evolutionary research (Berrebi et al. 2006). Importantly, markers with high absolute loadings on principal components (Figure 6, Table S2) explain the most variation captured among species and are therefore informative for distinguishing populations (Wilkinson et al. 2011). Their demonstrated amplification success in species with limited reference genomes, highlights their broader potential for biodiversity monitoring and sustainable resource management (Heuertz et al. 2023). Additionally, future studies comparing EPIC markers with other marker types (e.g., mitochondrial DNA and microsatellites) will be essential for distinguishing the roles of genetic drift, gene flow, and natural selection in shaping population structure of the studied species (Kirk and Freeland 2011).

Conclusion

4.4

The cross‐species amplification of EPIC markers enabled fine‐scale resolution of population structure and admixture patterns. High cross‐species transferability together with their ability to accurately retrieve inter and intra‐specific patterns emphasize their importance in conservation and aquaculture breeding programs. This study reveals high genetic diversity and differentiation among Oreochromis species across Eastern Africa, with clear population structure shaped by geographic isolation, historical connectivity, and human‐mediated activities.

The low to moderate genetic differentiation observed among many O. niloticus populations reflects ongoing gene flow and admixture, likely driven by natural dispersal and extensive aquaculture‐related translocations. The high differentiation between species of O. jipe , O. leucostictus , and O. esculentus indicates strong reproductive barriers and limited hybridization.

Overall, the results highlight the need for better aquaculture management practices. Conservation efforts should prioritize genetically distinct and isolated populations to preserve biodiversity and maintain ecosystem health. Future studies combining neutral and epic markers will further inform sustainable management strategies for East African tilapiine species. Furthermore, the ability of these markers to reveal adaptive variation in immune genes at the population level should be studied.

Author Contributions

Catherine Agoe: conceptualization (equal), data curation (equal), funding acquisition (supporting), investigation (lead), methodology (equal), project administration (equal), software (supporting), visualization (lead), writing – original draft (lead), writing – review and editing (lead). Gerald Kwikiriza: data curation (lead), investigation (equal), visualization (equal), writing – review and editing (equal). Peter Akoll: conceptualization (equal), supervision (supporting), writing – review and editing (equal). Papius Dias Tibihika: investigation (supporting), visualization (supporting), writing – review and editing (equal). Manuel Curto: conceptualization (equal), formal analysis (supporting), supervision (supporting), writing – review and editing (equal). John Walakira: conceptualization (supporting), supervision (supporting), writing – review and editing (equal). Thapasya Vijayan: methodology (equal), software (supporting), writing – review and editing (equal). Elizabeth Nyauchi: methodology (supporting), writing – review and editing (equal). John Kariuki: methodology (supporting), writing – review and editing (equal). Eva Dornstauder‐Schrammel: data curation (equal), methodology (lead), writing – review and editing (equal). Rose Basooma: methodology (supporting), writing – review and editing (equal). Sebastian Sonnenberg: data curation (equal), software (lead), writing – review and editing (equal). Paul Meulenbroek: conceptualization (supporting), supervision (supporting), writing – review and editing (equal). Harald Meimberg: conceptualization (lead), funding acquisition (lead), project administration (equal), resources (lead), supervision (lead), writing – review and editing (equal).

Funding

This study was partly financed by the Austrian Partnership Programme in Higher Education and Research for Development—APPEAR, a programme of the Austrian Development Cooperation (ADC) and implemented by Austria's Agency for Education and Internationalisation (OeAD), OEZA Project number: 0894‐01/2020, under which Catherine Agoe (Ref: MPC‐2023‐06640), Gerald Kwikiriza, Rose Basooma received support.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Genetic structure of Oreochromis niloticus populations inferred using STRUCTURE analysis. K* is the optimal number of clusters. Population codes are as follows; ABK = Albert, KNS = Kyoga, VMS = Victoria, KA = Kajjansi, KAY = Kyanamira, GKH = George, TU = Turkana, ON = Jipe, TA = Tana, H = Hashenge, CHA = Chamo. Table S1: The list of 50 EPIC markers that were genotyped, accession number, primer sequences (forward and reverse) and the gene descriptions. The five (5) EPIC markers that failed to amplify are indicated with an asterisk (*). Table S2: Principal Component Analysis (PCA) loadings of the 45 successfully amplified EPIC markers on principal components 1–45 across Oreochromis species.