Notes on Sweet orange (Citrus sinensis L. Osbeck) populations’ divergence: Landscape genetics, comparative phylogeny, and Niche modeling

doi:10.21203/rs.3.rs-3801400/v1

Download PDF

Research Article

Notes on Sweet orange (Citrus sinensis L. Osbeck) populations’ divergence: Landscape genetics, comparative phylogeny, and Niche modeling

https://doi.org/10.21203/rs.3.rs-3801400/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Sweet orange is one of the economically important plant species. The present study was conducted with the aim of generating genetic diversity data in Iranian sweet orange germplasm and investigating the landscape genetics of these plants in order to identify genetic regions compatible with environmental and climatic variables using SCoT molecular marker on 29 cultivars. The obtained results showed low to moderate genetic diversity in the sweat orange populations and indicated that the orange germplasm contains a complex genetic group of closely related individuals, but probably to some extent due to local breeding practices and artificial selection by orchard management. It is genetically differentiated. Also, some genetic kidneys were identified, especially in the southern regions of Iran. We also identified genetic regions that are significantly associated with environmental and climatic variables that can be used in the sweet orange conservation program in the country. This is especially true for the studied orange plants from southern Iran. The present study showed that global and local spatial variables affect the genetic structure of orange populations, and orange populations are separated by the phenomenon of distance, that is, as the geographical distance of the studied populations increases, the genetic distance increases. The analysis of species distribution modeling in the present study showed that both northern and southern regions of Iran are suitable habitats for orange cultivation, while temperature and precipitation are both important climatic factors for the cultivation and propagation of orange plants.

Dismo

LFMM

Maxent

Redundancy analysis

Spatial PCA

The genus Citrus L. contains important plant species that extensively contribute to human food and are cultivated in different parts of the world. The sweet orange (Citrus sinensis (L.) Osbeck) is the most important species among those belonging to the Citrus genus, representing about 50% of global Citrus production. Oranges are particularly appreciated for their organoleptic characteristics and the high nutraceutical value of the fruits (high content of antioxidants) (Seminara et al. 2023).

The Citrus genus originated from the Malay Archipelago and Southeast Asia, and the varieties of edible Citrus fruits on the market are generated by hybridizations of ancestral species, natural or artificial mutations, and human selection during domestication (Frydman et al. 2013; Wu et al. 2018; Li et al. 2019).

It has been suggested that Citrus diversified during the late Miocene epoch through a rapid southeast Asian radiation that correlates with a marked weakening of the monsoons. A second radiation enabled by migration across the Wallace line gave rise to the Australian limes in the early Pliocene epoch. Further identification and analyses of hybrids and admixed genomes provide insights into the genealogy of major commercial cultivars of Citrus. Among sweet oranges, Wu et al. (2018) find an extensive network of relatedness that illuminates the domestication of these groups.

Oranges originated in Asia in what is now called southeast China Cultivated for at least 7,000 years in India and in China since 2,500 BCE and documented in China since 340 BCE, sweet orange (Citrus x sinensis) is a hybrid between pomelo (Citrus maxima) and mandarin (Citrus reticulata) (Seminara et al. 2023).

The genetic and genomic studies in oranges have shown the diversification of C. sinensis and its relation with other Citrus species and have clarified that all the genetic diversity within the sweet orange species was derived from subsequent mutations starting from a single ancestor and was derived from complex cycles of hybridization and backcrossing between the mandarin (Citrus reticulata Blanco) and the pummelo (Citrus maxima (Burm.) Merr.) (Seminara et al. 2023).

Oranges are all intermediate between the two ancestors in size, flavor, and shape. The bitter orange and sweet orange both arose from mandarin-pomelo crosses, the former involving a pure mandarin, the latter with a mandarin already containing small amounts of pomelo.

Today oranges are cultivated in subtropical and tropical America, northern and eastern Mediterranean countries, Australia, and South Africa. Orange trees thrive in subtropical regions with warm temperatures and moderate humidity levels. Brazil is the leading country in orange production worldwide with 30% of world production. Within Brazil, the region with the largest production volume is Sao Paulo, with 94% of the total.

Oranges should be stored at about 38°F to 48°F. Warmer temperatures will result in a more rapid loss of quality while prolonged exposure to temperatures below 38°F may result in freeze damage. These trees grow well in slightly acidic, well-drained loam or sandy loam soils, but with proper irrigation that drains easily, could grow well in clay soils. At present, there are over 400 different varieties of oranges grown around the world.

Landscape genetics explores how the micro evolutionary processes of gene flow, genetic drift, and natural selection interact with environmental heterogeneity to shape population genetic structure and identify the movement corridors and barriers to gene flow, and the relative effects of current versus historical landscape factors on population genetic structure (Wang 2012).

This discipline also tries to identify the present-time adaptive genetic loci and investigate the adaptive potential of populations in response to future landscape and climatic changes (Kimberly 2019). Therefore, a combination of population and landscape genetics or genomics is likely to provide the best understanding of the molecular imprint of local adaptation and further guide the conservation strategies. In addition to local adaptation, migrating to new favorable locations is also a response pattern of plants to rapid climate changes, which is important for species conservation (Liu et al. 2022). Data obtained in landscape genetics analyses are intended to inform the conservation and management of the target species (Storfer et al. 2007; Epps and Keyghobadi 2015).

A variety of different molecular markers have been used to study Citrus species' phylogeny and genetic diversity. The markers used are simple sequence repeats (SSRs), nuclear LEAFY Second Intron and Plastid trnL-trnF Sequence, genome sequencing, expression sequence tags (ESts), restriction fragment length polymorphism (RFLP), randomly amplified polymorphic DNA (RAPD) and cleaved amplified polymorphic sequences (CAPS), EST-SSR and genomic SSR markers (see for example, Yingzhi et al. 2007; Omura et al. 2016; Kaur et al. 2022).

SCoT molecular markers are based on the short conserved region flanking the ATG start codon in plant genes (Collard and Mackill 2009), and targets the coding sequence in the genome, mostly the open reading frames (ORFs), which reveals the relationship between amplification results and phenotype (Li and Quiros 2001). It has the benefits of simple operation, low cost, and abundant polymorphisms. The SCoT marker has been used in many crop plant species including orange (Juibary et al. 2021).

Global warming and climate change is considered one of the greatest threats of the 21st century that may affect global biodiversity, with serious ecological consequences, like frequent droughts, wildfires, and invasive pest outbreaks, leading to the loss of plant species and lowered productivity, shortages of food crops, and a higher cost for consumers. Climate change can adversely affect genetic diversity and genetic connectivity of plant populations, ending in a greater homogenization in genetic content and a lower adaptive potential for the species in future environmental changes (Guan et al. 2021).

The species distribution models (SDMs) are used to study the present geographical distribution of species and predict their future occurrence in response to climate changes (Elith and Leathwick 2009). SDMs can be used to make inferences about the distribution of suitable habitats for species of interest, population demography, and genetic diversity (Lee-Yaw et al., 2021).

The concept of the fundamental niche versus the realized niche originally developed by Hutchinson (1957), comprises the central concept of species distribution models. Both abiotic and abiotic environmental conditions and the movement capacity of a species determine the geographic area suitable for the specie (Elith and Leathwick 2009).

Species distribution modeling (SDM), uses computer algorithms to predict the distribution of a species across geographic space and time using environmental data. SDMs use climate data (e.g. temperature, precipitation), and other variables such as soil type, water depth, and land cover. These studies are used in conservation biology, ecology, and evolution and try to illustrate how environmental conditions influence the occurrence or abundance of a particular species. Data obtained may be used for predictive purposes (ecological forecasting), a species’ future distribution under climate change, and predictions are made on the current and/or future habitat suitability of the target species (Elith and Leathwick 2009).

SDM has been also applied to landscape genetics to investigate the association of genetic variation with environmental gradients and make inferences about the role of gene flow and selection (Ortego et al. 2012; Poelchau and Hamrick 2012). These studies often use model predictions to describe habitat or climatic suitability as a single integrated measure of multiple complex environmental factors, which is then assessed in terms of its influence on genetic patterns.

To our knowledge and the literature review, no report is available on landscape genetic analysis of Citrus species including landraces of sweat orange in the country and elsewhere. Therefore, the aims of this study are to 1) investigate the population's genetic structure, and identify the geographical variables shaping it in sweat orange populations, 2) identify the climatic factors affecting the present and future distribution of these plants, and, 3) identify SCoT genetic regions potentially adapted to geographic variables in the studied populations. These findings may be used in future breeding and selection programs for orange trees in the country.

We used multiple computational methods in our study like, K-Means Clustering and linear discriminant analysis of principal components (DAPC), for grouping the accessions studied; spatial PCA (sPCA), for landscape genetic analysis, RDA (redundancy analysis), and LFMM (latent factor mixed model), for identifying SCoT regions with adaptive potentials to geographical and climatic variables studied, and species modeling by Maxent and Dismo package. Moreover, we performed random forest analysis to illustrate the importance of geographical variables on genetic polymorphism and genetic differentiation (Fst values) of the studied populations.

2.1. Plant material

We studied 130 plants of 29 sweet orange populations that are cultivated in different regions of the country. we used the information of 24 populations of oranges from the north of Iran in Lame et al. 2019, and we also studied four populations of oranges from the south of Iran for the first time (Table 1).

Table 1

the information of populations of oranges from the north and South of Iran
Number	Locality	pop	Locality	Long	Lat	Alt
1	North of Iran	Thompson ( Navel)	Joybar	36.63	52.9	1340
2	North of Iran	Thompson	Shahsavar	36.81	50.87	-13
3	North of Iran	Thompson	Joybar	36.63	52.9	1340
4	North of Iran	Thompson	Behshahr	36.69	53.54	29
5	North of Iran	Valencia	Joybar	36.63	52.9	1340
6	North of Iran	Khooni-Ablagh	Siavarz village of Shahsavar	36.81	50.87	-13
7	North of Iran	Camcuat	Siavarz village of Shahsavar	36.81	50.87	-13
8	North of Iran	Lamcuat	Siavarz village of Shahsavar	36.81	50.87	-13
9	North of Iran	Jaffa or Palestine orange	Joybar	36.63	52.9	1340
10	North of Iran	Jaffa or Palestine orange	Joybar	36.63	52.9	1340
11	North of Iran	Hamlin orange	Ramsar	36.92	50.64	21
12	North of Iran	Jaffa or Palestine orange	Joybar	36.63	52.9	1340
13	North of Iran	Moro	Siavarz village of Shahsavar	36.81	50.87	-13
14	North of Iran	Cara navel orange	Joybar	36.63	52.9	1340
15	North of Iran	local orange	Joybar	36.63	52.9	1340
16	North of Iran	Sangin Navel	Joybar	36.63	52.9	1340
17	North of Iran	local orange (kooche farhangiyan)	Joybar	36.63	52.9	1340
18	North of Iran	Local orange	Sari	36.56	53.05	42
19	North of Iran	Local orange	Joybar	36.63	52.9	1340
20	North of Iran	Local orange of Gilan	Gilan	37.11	49.52	800
21	North of Iran	Local orange( jadeh nezami)	Ghaemshahr	36.46	52.85	104
22	North of Iran	Local orange Fereydoon kenar	Fereydoon kenar	36.68	52.53	-23
23	North of Iran	Local orange (Paein rood dasht)	Joybar	36.63	52.9	1340
24	North of Iran	Local orange Behshar	Behshahr	36.69	53.54	29
25	North of Iran	Local orange ghaemshahr	Ghaemshahr	36.46	52.85	104
26	South of Iran	washington Navel Orange	Dezfoul	32.24	49.48	82
27	South of Iran	Salustiana Orange	Dezfoul	32.24	49.48	82
28	South of Iran	Marrs Orange	Dezfoul	32.24	49.48	82
29	South of Iran	Valencia Orange	Dezfoul	32.24	49.48	82

2.2. DNA extraction and PCR details

The genomic DNA was extracted from leaves following a modified CTAB-activated charcoal (cetyltrimethylammonium bromide) extraction procedure (Krizman et al. 2006). The extraction procedure was based on activated charcoal, polyvinylpyrrolidone (PVP), and 2-mercaptoethanol for the binding of polyphenolics during extraction and under mild extraction and precipitation conditions. This promoted high-molecular-weight DNA isolation without interfering with contaminants. The DNA concentration was measured by running aliquots on a 1% agarose gel. PCR reactions were performed in a 25 µL volume containing 13µl master mix, 8 µl H2O, 1 µl Primer (SCoT), and 3 µl DNA.

2.3. SCoT Assay

In this study, three SCoT primers, SCoT-1, SCoT-2, and SCoT-36 were used. The PCR was carried out with 45 cycles of 4 min for the initial denaturation step at 95°C, 1 min at 94°C, 1 min at 54–56°C, and 1 min at 72°C. The reaction was completed by a final extension step of 7 min at 72°C. The amplification products were examined using 2% agarose gel electrophoresis and fluorescence staining on KBC power load (Kowsar Biotech Company, Iran). The fragment size was estimated by using a 100-bp molecular size ladder (Fermentas, Germany).

2.4. Data analyses

2.4.1. Genetic grouping and gene flow

2.4.1.1. Genetic diversity and grouping of the populations

A binary data matrix indicating the presence (1) or the absence (0) of bands was made from SCoT profiles. Only strong, reproducible, and distinguished bands were used for the analysis. The software GenAlex 6.5 (Peakall and Smouse 2006) was used to estimate the genetic diversity parameters (Weising 2005).

Genetic differentiation of the studied populations was determined by AMOVA with 1000 permutations as performed in GenAlex 6.5 (Peakall and Smouse 2006), and the Just’ D parameter as determined in GenoDive v. 3.6.

The grouping of the studied populations was determined using UPGMA (unweighted paired group using average), and K-Means clustering, as well as discriminant analysis of principal components (DAPC).

The assignment test of DAPC and admixture plot were carried out to illustrate gene flow and admixture of the studied orange trees. These analyses were performed in Adegenet and LEA packages of R 4.2.

The clonality test of the GenoDive package ver.3.6 (Meirmans and Van Tienderen 2004), was used to test random mating among sweat orange accessions studied which is based on the concept of clonal diversity. Under asexual reproduction, the clonal diversity is lower than under sexual reproduction. Therefore, we can test the null hypothesis that the observed clonal diversity is due to sexual reproduction by randomizing alleles over individuals and comparing the observed clonal diversity with that of the randomized dataset (Gomez & Carvalho, 2000).

This test allows for the inclusion of the threshold concept used for the clone assignment (Meirmans 2020). For every permuted dataset, the distance matrix is recalculated, clones are assigned using the same threshold as for the original dataset, and a diversity index is calculated using the assigned clones. In this way, the test allows for mutations and genotyping errors within clones. There are several processes other than asexual reproduction that can lead to identical genotypes, such as inbreeding (especially self-fertilization) and undetected population structure (Halkett et al., 2005).

2.4.1.2. Association studies

The association between genetic data and geographical and climatic variables was determined by the RDA method by 999 permutations as performed in PAST program ver.4.

Similarly, LFMM (Latent factor mixed model) was used to identify the SCoT loci with significant association with geographical and climatic variables. This was done after utilizing FDR (False determined ratio), and Bonferrini test, as performed in the LFMM package in R. ver.4.2 (Frichot et al.2013).

LFMM is a Bayesian approach method for testing associations between loci and geographical gradients using latent factor mixed models. It performs a regression analysis in which the confounding variables are modeled with unobserved (latent) factors. The program estimates correlations between geographical and ecological variables and allelic frequencies and simultaneously infers the background levels of population structure (Frichot et al. 2013).

RDA is a form of constrained ordination that suits genomic data sets, where we are interested in understanding how the multivariate environmental factors shape the patterns of genomic composition across geographical areas. RDA is a direct gradient analysis technique that summarizes linear relationships between components of response variables that are "redundant" with (i.e. "explained" by) a set of explanatory variables. It is based on multivariate regression, (Legendre and Legendre 2012).

2.4.2. Landscape genetics

2.4.2.1. Spatial PCA (sPCA)

We used the spatial PCA method (sPCA) to investigate the spatial pattern of genetic variability based on allelic frequency data of individuals or populations (Jombart et al. 2008). This approach is independent of presumed Hardy–Weinberg expectations or linkage equilibrium among loci and uses statistical (Monte Carlo) tests to partition the spatial structure into random, local, and global variance patterns, where local patterns are taken to relate to highly negative spatial autocorrelation and global patterns are taken to relate to highly positive spatial autocorrelation.

The spatial autocorrelation is measured using Moran's I (Moran 1950), which is incorporated within the sPCA algorithm. It differentiates between global structures (patches, clines, and intermediates) and local ones (which represent strong genetic differences between neighbors) and from random noise (Jombart et al. 2008). These analyses were performed in the Adegenet package in R 4.2.

2.4.2.2. Random Forest (RF) analysis

RF (Random Forest) is a machine learning method that is used in spatial analysis to illustrate the importance of spatial factors affecting genetic differentiation (Fst value), and genetic polymorphism in the studied populations. This is a supervised machine learning algorithm that is used in classification and Regression problems. It builds decision trees on different samples and takes their majority vote for classification and average in case of regression (Breiman 2001). This approach consists of many decision trees and uses bagging and feature randomness when building each tree to try to create an uncorrelated forest of trees whose prediction by committee is more accurate than that of any individual tree.

Random Forests illustrates the feature importance, based on the calculated Gini Index. A lower value of this index shows the importance of that variable and reveals the association between genetic data and landscape features (Pless et al. 2021). RF analysis was performed in the random Forest package of R. 4.2, with 500 bootstraps.

2.4.2.3. Species distribution modeling (SDM)

We used a combination of methods to perform SDM, such as Maxent (a maximum entropy modeling method tailored for presence-only species, Phillips et al., 2006), and Dismo analyses.

MaxEnt estimates the most uniform distribution (‘maximum entropyʼ) across the study area, given the constraint that the expected value of each environmental predictor variable under this estimated distribution matches its empirical average (average values for the set of speciesʼ presence records) (Phillips et al. 2006).

Models were run using linear, quadratic, and product features in MaxEnt. Predicted climatic habitat suitability maps were produced for sweat orange populations using MaxEnt’s logistic output, which provides an estimate of the probability of presence ranging from 0 (low suitability) to 1 (high suitability) in geographic space. We also used the dismo package in R ver 4. 2, for SDM analysis, which uses the Bioclim algorithm. This algorithm has been extensively used for species distribution modeling and is the classic’ climate-envelope model’.

SDMs require species absence point data, which we obtained using pseudo-absences (PAs) methods, to predict suitable species habitat. All the models were constructed with 80% training and 20% testing of occurrence data. The model evaluation was performed by both the threshold method and AUC determination (ROC curve).

In species distribution modeling we used projected climate data layers for the current (∼1950–2000) and the year 2050 period (average for years 2061–2050) based on 19 bioclimatic variables at a resolution of 5 minutes spatial resolution (of a longitude/latitude degree, this is about 9 km at the equator).

Data were downloaded from the World Clime database. To represent climate change impacts, future climate variables were projected for 2050, under Representative Concentration Pathways (RCP, 2.6). We used bioclimatic variables derived from the monthly temperature and rainfall values. These bioclimatic variables are coded as follows:

BIO1 = Annual Mean Temperature, BIO2 = Mean Diurnal Range (Mean of monthly (max temp - min temp)), BIO3 = Isothermality (BIO2/BIO7) (Å~100), BIO4 = Temperature Seasonality (standard deviation Å~100), BIO5 = Max Temperature of Warmest Month, BIO6 = Min Temperature of Coldest Month, BIO7 = Temperature Annual Range (BIO5-BIO6), BIO8 = Mean Temperature of Wettest Quarter, BIO9 = Mean Temperature of Driest Quarter, BIO10 = Mean Temperature of Warmest Quarter, BIO11 = Mean Temperature of Coldest Quarter, BIO12 = Annual Precipitation, BIO13 = Precipitation of Wettest Month, BIO14 = Precipitation of Driest Month, BIO15 = Precipitation Seasonality (Coefficient of Variation), BIO16 = Precipitation of Wettest Quarter, BIO17 = Precipitation of Driest Quarter, BIO18 = Precipitation of Warmest Quarter, BIO19 = Precipitation of Coldest Quarter.

2.4.2.4. Comparative Phylogenetic analyses

We assessed the phylogenetic signal of the studied environmental variables by computing Blomberg's K and lambda parameters as performed in the phytools package (Revell 2011) in R. These values indicate the effects of environmental variables on sweat orange populations’ divergence and similarity. Pagel's λ is the transformation of the phylogeny that ensures the best fit of trait data to a Brownian Motion model” (Pagel 1999). Blomberg's K is defined as the ratio between two other ratios. The first (observed) is the Mean Squared Error of tip data divided by the Mean Squared Error of data calculated using a variance-covariance matrix derived from the phylogeny. The second ratio (expected) is the same thing, but using data from a model under the assumption of Brownian motion of trait evolution (Blomberg et al. 2003).

2.4.2.5. Disparity analyses

In order to visualize the relationship between phylogeny and populations’ distribution in the character space, we constructed a ‘phylomorphospace’, which projects the branches of a phylogenetic tree into a 2D morpho-space defining trait variation among the species/ or populations (Ravell 2011). The phylomorphospace was constructed for the studied variables as well as the PCA-axes of these variables.

3.1. Genetic polymorphism and grouping of the accessions

We obtained 44 SCoT bands or loci in total. The genetic polymorphism obtained ranged from a very low magnitude (6.8%) in population number 1, to a moderate level (45%) in population number 21.

Different methods used to group sweat oranges studied are presented in Fig. 1. These methods include, K-means clustering and DAPC analysis for the genetic grouping of the accessions irrespective of their population number. All these methods suggest placing the accessions in three to four genetic groups with some degree of genetic admixture. However, in all analyses, accessions collected from the South-Iran comprised a distinct, separate group (plants number 90–130 in Fig. 1).

3.2. Admixture plot and assignment test

Admixture analysis and assignment tests to study the extent of genetic affinity in sweat orange accessions are presented in Fig. 2. Both analyses produced similar results and placed sweat orange plants of the South-Iran in a distinct genetic group with a lower magnitude of gene flow with the rest of plants studied. These orange plants form a more genetically uniform group compared to the North-Iran accessions studied.

A higher magnitude of genetic admixture was observed in the orange plants of North Iran. However, there exists a good level of genetic differentiation in the orange populations studied, as evidenced by significant paired-AMOVA (p = 0.01), obtained among these populations.

AMOVA also revealed that about 63% of total genetic variation is among orange populations, while 37%. is due to within-population genetic variability.

The clonal diversity test at different thresholds revealed a different number of clones in the orange populations studied, which were in accord with random mating (p = 0.30).

These results are in agreement with admixture and assignment plots reported before, and that many orange populations are genetically alike probably due to the high level of genetic admixture.

DAPC analysis identified SCoT loci that contribute the most in differentiating sweat orange accessions (Fig. 3). SCoT loci 11, 12, 15, 33, and 37, are differentiating loci of the first LDA axis, while SCoT loci 8, 22, 30, 35, and 38, are important loci of the second LDA axis. These markers may be used in the genetic fingerprinting of sweat oranges as they can differentiate different populations and also reveal gene flow among different accessions.

3.3. Landscape Genetics and Association studies

We used both RDA and LFMM analysis to identify the SCoT loci with adaptive potential in response to geographical and environmental variables.

RDA analysis produced a significant association (p = 0.001) between SCoT loci and geographical variables (Fig. 4, A). The first RDA axis comprised about 20% of the total variance, while the second axis comprised about 3% of this variance.

LFMM analysis followed by Bonferroni and FDR tests also revealed a significant association between some of the SCoT loci and geographical variables (Fig. 4, B-D). These genetic regions may be selected for orange accessions for cultivation in different geographical regions and also be used in hybridization programs.

3.3.1. Spatial PCA (sPCA) analysis

The results of sPCA analysis for both pairs of longitude and latitude, as well as humidity and temperature showed significant effects of these spatial variables on the sweat orange populations studied. We obtained both significant positive and negative eigenvalues for all these variables, showing that orange populations are genetically structured by both global scale and local spatial features.

IBD analysis and Moran I index after 999 times permutations produced significant results for spatial variables (p < 0.001). Significant Mantel test and IBD indicate that with an increase in geographical distance of sweat orange populations, the genetic difference increases too. Similarly, Morans' I index was significant for both global and local spatial features studied. Moreover, the global and local tests (m-tests in the sPCA package), produced significant p < 0.001, after 999 permutations. Therefore, both global and local structures affect the genetic structure in orange populations.

The neighbors' connectivity graph and the scores of entities in space, as well as the genetic clines of the populations are provided in (Fig. 5, A-C). These figures also support populations' genetic isolation in space. Orange plants of South Iran are placed in a separate genetic cline group on the left corner of these plots.

SCoT loci that contribute to global and local spatial genetic structuring (genetic clines) are presented in Fig. 6. These loci are loci No. 4, 12, 15, 21, 23, 26, 28, 30, 33, 37, 39, and 44.

3.3.2. Random Forest results

We used genetic polymorphism and Fst values of the studied populations and categorized them in distinct levels from A-D classes; depending on the degree of differentiation, i.e. from lowest to highest value of Fst, and based on the level of genetic polymorphism.

RF analysis with 500 bootstraps revealed 90 percent accuracy for both train predictive data, which shows a very good fit of results to the model provided. The error rates obtained were very low and less than 0.1, and therefore precise results are obtained from RF analysis.

Based on the Ginni value, the importance of the studied spatial feature for genetic diversity starts with latitude (Ginni value = 0.99), followed by altitude (Ginni value = 1.0), and temperature (Ginni value = 1.1). Therefore, these variables affect the level of genetic diversity in sweat orange populations.

The same analysis for Genetic differentiation (Fst value), starts with altitude (Ginni value = 0.82), humidity (Ginni value = 0.96), followed by longitude (Ginni value = 1.18). The lowest importance value was obtained for latitude (Ginni value = 4.60), followed by longitude (Ginni value = 1.51). Therefore, orange populations are mainly differentiated in their genetic content (Fst), mainly in response to altitude, humidity, as well as longitude (Fig. 7, A, and B).

Comparative phylogeny and Disparity analyses of sweat orange populations, divergence also revealed significant signals for humidity (K = 1.20, P = 0.007; lambda = 0.99, p = 0.0006), temperature (K = 1.20, P = 0.004; lambda = 0.99, p = 0.0006), and longitude (K = 1.30, P = 0.007; lambda = 0.99, p = 0.0001). However, no significant signal was obtained for latitude, and altitude (p = 0.3). Therefore, these results revealed the importance and effect of both climatic and geographical variables for the sweat orange population divergence (Fig. 7, C, and D).

3.4. Species distribution modeling

The model’s probability of occurrence for orange populations produced by the dismo package showed a maximum probability of 0.8. Therefore, it shows that we have almost a good collection of these plants from occurring areas in the country, but there are potential areas that were not covered in our collection of plant materials. After the estimation of pseudo-absence points, we took the post hoc evaluation of the model as stated before in the material and methods section. Finally, the threshold was used to paint a map with the predicted range of the orange plants. The predicted distribution area map reveals the areas in which oranges have the potential of growing and may be considered for planning and conserving this species. The same analyses based on climate data for the year 2050 produced a prediction of the presence of orange plants, which shows a little lower occurrence of these plants in southern parts of Iran. AUC obtained for this

The model was very low (0.30).

The results we obtained in landscape genetic analysis showed the presence of genetic structuring of orange plants in a few genetic clines located in the North as well as in Southern Iran. As evident, due to climate change in the year 2050, we probably lose some of the orange cultivation areas in the genetic cline area of South Iran.

Maxent modeling results, produced almost similar distribution maps as in the case of dismo package (Fig. 8, A, and B). The gain and importance of bioclimatic variables are presented in Fig. 8, E. The most important and influential variables of the present-time distribution of orange plants are those related to temperature and precipitation. We obtained the same result in the forecast future time of the year 2050 too.

The AUC and ROC curves of the Maxent model in both the present time and forecast time in the year 2050, showed that this model is highly accurate in performance (Fig. 8, A, and B). The AUC for the present time and the year 2050 were almost equal to 1 indicative of a perfect mode.

Due to the effects of future climate change on orange cultivation in the country, we performed an LFMM analysis to identify the potentially adaptive SCoT loci for climate change in the year 2050. As evident in Fig. 9, a few SCoT loci are potentially adaptive to both temperature and precipitation conditions in the year 2050, and therefore, for the conservation program we should conserve the plants with these genetic regions and use them in breeding plans.

4.1. Genetic diversity and gene flow

The present study revealed a low to moderate genetic variability within sweat orange populations, but a significant genetic difference is present among the studied populations. This holds true, particularly for the orange plants studied from South Iran. The low within-population variation may be due to the asexual reproduction usually by grafting of the orange plants.

We observed a high rate of clonality in the studied orange populations, which may be due to the above-said reason.

Despite the genetic admixture and clonality of the orange populations studied, they can be placed in three to four broad genetic groups which differ significantly from each other based on AMOVA. Therefore, we have a good genetic variability in the orange germplasm of the country, which can be managed for the future breeding of these plants.

In a similar study, Malik et al. (2012), also reported a low magnitude of genetic variability in Indian accessions of Citrus sinensis. The morphological diversity observed in these plants was mainly due to somatic mutations. The low level of genetic variability has been noticed in orange plants in Turkey (Uzun et al. 2009), as well as in other parts of the world (see for example, Ben-Abdelaaliet al. 2018, Lame et al., 2021). Rao and Hodgkin (2002), suggest that we need a deep and clear insight into the genetic diversity of orange plants and their distribution throughout geographical distribution for the planning conservation and breeding purposes, and hence, the present findings may be of use in future conservation and breeding of sweet oranges in the country.

The present study reports the presence of colons within sweat orange populations due to a high degree of genetic similarity and admixture.

Aydin and Turgut (2012), studied Turkey’s orange cultivars and identified a large number of clones for the cultivars,’Washington Navel’, ‘Valencia’, ‘Moro’, ‘Shamouti’, ‘Pineapple’, ‘Parson Brown’, ‘Salustiana’, Sanguinello’, ‘Tarocco’, and ‘Yafa’, which were either introduced from other countries or selected in Turkey. The presence of a low level of intraspecific genetic diversity within orange germplasm may expose the cultivars to many harmful pests and diseases due to a lack of sources of resistance (Seminara et al., 2023).

4.2. Association studies

We reported the association between genetic markers and geographical as well as climatic variables which may be used in the future selection and breeding of orange plants in the country.

Cultivation and propagation of suitably adapted genotypes to a particular environmental condition can tune future plantations of orange. For example, nowadays, Valencia oranges are among the most widely cultivated varieties worldwide, with fruits used both for fresh consumption and for processing. The success of the “Valencia” orange has been largely determined by its adaptability to different climatic conditions, its high productivity, and its good fruit conservation both on the plant and during postharvest, and the fruits are usually seedless (Seminara et al., 2023)

A similar association study should be planned between genetic structure and agronomic and morphological characteristics in different orange cultivars. Sweet orange breeding programs are carried out worldwide by applying conventional breeding approaches, like selection and mutation breeding, and hybridization, as well as non-conventional methods (somatic hybridization and genetic engineering). These methods are applied with the aim to improve orange plant traits, such as tolerance or resistance to biotic and abiotic stresses, the ripening period, postharvest behavior, yield, improved fruit quality for fresh consumption (i.e., peel and flesh color, flavor, seedlessness, and beneficial compound content), and industrial transformation (i.e., juice yield, color, etc.). Moreover, the clonal selection is also performed in oranges which are relatively simple since it allows for the improvement of one of a few characteristics in a well-defined phenotype. The generation of sweet-orange-like hybrids is more challenging compared to clonal selection, but it is now needed to introgress favorable genes from other Citrus species or relatives.

4.3. Landscape genetics and species distribution modeling

The present study revealed that global and local spatial variables affect the genetic structuring of orange populations in the country. The presence of heterogeneous environmental conditions brings about changes in the genetic diversity of plant species which in turn results in local adaptations (Zhang et al. 2019; Garot et al. 2021).

Orange populations showed isolation by distance phenomenon, i.e. by an increase in geographical distance of the studied populations, an increase in genetic distance occurs. This may be due to some limitations to gene flow between populations that are placed far from each other. In this situation, we may have local adaptation, particularly, at the margin of plant distribution, as also revealed in the sPCA analysis of orange populations studied. In fact, the LFMM analysis identified some of the genetic loci which are potentially adapted to environmental and climatic variables.

Therefore, the studies concerned with the genetic basis of local adaptation and identifying adaptive genetic loci or SNPs can improve the knowledge of the genetic mechanism of local adaptation and probably species diversification within a genus, as well as population divergence within a single species (Zhang et al. 2019). These data can be utilized for proper conservation programs of important plant species including Citrus species in different regions of the world.

Species distribution may occur in the sparsely distributed populations that occupy habitat fragments to continuously distributed individuals of the same species that cover tens to thousands of hectares. Human activity, like, building agricultural fields and urban and suburban development, and foot trails, roads, and highways, may act as either barriers or conduits for dispersal (Cruzan and Hendrickson, 2020). In cases like orange populations, genetic fragmentation occurs across multiple habitat fragments, and as the distance between sample sites increases, the effects of gene flow on genetic similarity decrease, and genetic differentiation becomes governed primarily by mutation and genetic drift.

Species distribution modeling analysis in the present study shows that both northern as well as southern Iran regions are suitable habitats for orange cultivation. In fact, dispersal within and among plant populations is determined by habitat suitability, as well as the behavior of seed and pollen dispersal vectors. If a seed is dispersed outside the habitat of its parental plant, the probability of encountering high-quality habitat sufficient for germination and seedling recruitment is low and the majority of dispersed propagules will fail to establish in novel locations (Cruzan and Hendrickson, 2020).

We also reported that both temperature and precipitation are important climatic factors for orange plants’ cultivation and propagation. It has been stated that regardless of dispersal distance, the establishment, reproductive success, and fitness of plants depend on the habitat suitability, like plant–soil feedback loops, temperature, and precipitation shifts, and light availability, among other conditions (Cruzan and Hendrickson, 2020).

In conclusion, the present study produced data on genetic diversity within the sweet orange germplasm of Iran and revealed gene flow and genetic admixture among these populations. The results showed that orange germ-plasm comprises a complex genetic group of closely related individuals, but yet somewhat genetically differentiated probably due to local propagation practices and artificial selection by orchard management. Moreover, some genetic clines were identified particularly in southern regions of Iran. We also identified some genetic regions which are significantly associated with environmental and climatic variables which can be used in the conservation program of orange in the country.

Ethical Approval

Not applicable.

Funding

Not applicable

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Aydin U, Turgut Y (2012) Genetic diversity in citrus. Genetic Diversity in Plants, pp 213–230.
Ben Abdelaali S, Saddoud O, Abdelaali NB, Hajlaoui MR, Mars M (2018) Fingerprinting of on-farm conserved local
tunisian orange cultivars (Citrus sinensis (L.) osbeck) using microsatellite markers. Acta Biol Cracov Bot 60:83–93.
Blomberg SP, Garland JR, Ives AR (2003) Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57(4): 717-745.
Breiman, L (2001) Random forests. Mach Learn 45: 5-32.
Cruzan MB, Hendrickson EC (2020) Landscape Genetics of Plants: Challenges and Opportunities. Plant Communications 1 (6) : 100100.
Debbabi OS, Mezghani N, Madini M, Abedelaali NB, Bouhlel R, Ksia A, Mars M (2014) Genetic diversity of orange
fruit (Citrus sinensis L.) cultivars in Tunisia using AFLP markers. IJAAR 5(1):7–15.
Elith J, Leathwick JR (2009) Species distribution models: ecological explanation and prediction across space and time. Annu Rev Ecol Evol Syst. 40 (1): 677–697.
Epps CW, Keyghobadi N (2015) Landscape genetics in a changing world: disentangling historical and contemporary influences and inferring change. Mol Ecol. Dec. 24(24):6021-40.
Frichot E, Schoville SD, Bouchard G, François O (2013) Testing for associations between loci and environmental gradients using latent factor mixed models. Mol Biol Evol. 30(7), 1687-1699.
Frydman A, Liberman R, Huhman D.V, Carmeli-Weissberg M, Sapir-Mir M, Ophir R, Sumner LW, Eyal Y (2013). The molecular and enzymatic basis of bitter/non‐bitter flavor of Citrus fruit: evolution of branch‐forming rhamnosyltransferases under domestication. Plant J. 73(1), 166-178.
Guan B, Gao J, Chen W, Gong X, Ge G (2021) The Effects of Climate Change on Landscape Connectivity and Genetic Clusters in a Small Subtropical and Warm-Temperate Tree. Front. Plant Sci.
Garot E, Dussert S, Domergue F, Joët T, Fock-Bastide I, Combes MC, Lashermes P (2021) Multi-approach analysis reveals local adaptation in a widespread forest tree of reunion Island. Plant Cell Physiol. 62(2): 280-292.
Gomez A, Carvalho GR (2000) Sex, parthenogenesis and genetic structure of rotifers: microsatellite analysis of contemporary and resting egg bank populations. Mol Ecol. 9:203-214.
Hutchinson GE (1957) Concluding remarks. Cold Spring Harbor Symposia on Quantitative Biology. 22:415–427.
Halkett F, Simon JC, Balloux F (2005) Tackling the population genetics of clonal and partially clonal organisms. Trends Ecol Evol. 20:194-201.
Jombart T, Devillard S, Dufour AB, Pontier D (2008) Revealing cryptic spatial patterns in genetic variability by a new multivariate method. Heredity 101(1): 92-103.
Juibary P L, Seyedmehdi F, Sheidai S, Koohdar F (2021) Genetic structure analysis and genetic finger printing of sweet orange cultivars (Citrus sinensis (L.) Osbeck) by using SCoT molecular markers. Genetic Resources and Crop Evolution 68(1):1-10.
KimberlyAW (2019) Landscape Genetics: Landscape Effects on Gene Flow and Population Genetic Structure. Essentials of Landscape Ecology (Oxford; online edn, Oxford Academic, 22 Aug. 2019), accessed 21 Apr. 2023.
Lamine M, Chebaane A, Mliki A (2015) Genetic diversityanalysis in Tunisian Maltaise orange (Citrus sinensis L.).
J New Sci 14(2): 438–448.
Lee-Yaw JA, McCune JP, Pironon S, Sheth S N (2021) Species distribution models rarely predict the biology of real populations. Ecography 2022 (6): E4.
Li L J, Tan W S, Li WJ, Zhu YB, Cheng YS, Ni H (2019) Citrus taste modification potentials by genetic engineering. Int J Mol Sci 20(24): 6194.
Liu K, Qi M, Du FK (2022) Population and Landscape Genetics Provide Insights into Species Conservation of Two Evergreen Oaks in Qinghai–Tibet Plateau and Adjacent Regions.Front. Plant Sci., Sec. Plant Bioinformatics. 13
Legendre, P., & Legendre, L. (2012). Numerical ecology. Elsevier.
Malik SK, Rohini MR, Kumar S. et al (2012) Assessment of Genetic Diversity in Sweet Orange [Citrus sinensis (L.) Osbeck] Cultivars of India Using Morphological and RAPD Markers. A gric Res 1: 317–324.
Meirmans PG, Van Tienderen PH (2004) GENOTYPE and GENODIVE: two programs for the analysis of genetic diversity of asexual organisms. Mol Ecol Notes 4: 792-794.
Meirmans, P.G. 2020, GENODIVE version 3.6: Easy-to-use software for the analysis of genetic data of diploids and polyploids. Mol Ecol Resour. 20: 1126-1131.
Moran PA (1950). Notes on continuous stochastic phenomena. Biometrika 37(1/2): 17-23.
Omura M, Shimada T (2016) Citrus breeding, genetics and genomics in Japan. Breed Sci 66(1): 3-17.
Ortego J, Riordan EC, Gugger PF, Sork VL (2012) Influence of environmental heterogeneity on genetic diversity and structure in an endemic southern Californian oak. Mol Ecol. 21: 3210 – 3223.
Pagel M (1999) Inferring the historical patterns of biological evolution. Nature 401: 877-884.
Peakall ROD, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6(1): 288-295.
Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecol Modell. 190z231-259.
Pless E, Saarman NP, Powell, JR, Caccone A, Amatulli, G (2021) A machine-learning approach to map landscape connectivity in Aedes aegypti with genetic and environmental data. Proc Natl Acad Sci USA 118(9): e2003201118.
Poelchau MF, Hamrick JL (2012) Differential effects of landscape-level environmental features on genetic structure in three codistributed tree species in Central America. Mol Ecol. 21: 4970– 4982.
Revell LJ (2011) Phytools: An R package for phylogenetic comparative biology (and other things). Evolution 3(2):217-223.
Seminara S, Bennici S, Di Guardo M, Caruso M, Gentile A, La Malfa S, Distefano G (2023) Sweet Orange: Evolution, Characterization, Varieties, and Breeding Perspectives. Agriculture. 13:264.
Storfer A, Murphy MA, Evans JS, Goldberg CS, Robinson S, Spear SF, Dezzani R, Delmelle E, Vierling L, Waits LP (2007) Putting the 'landscape'in landscape genetics. Heredity 98(3): 128-142.
Uzun A, Yesiloglu T, Aka-Kacar Y, Tuzcu O, Gulsen O (2009). Genetic diversity and relationships within Citrus and related genera based on sequence related amplified polymorphism markers (SRAPs). Scientia Horticulturae, 121:306–312.
Wang IJ (2012) Environmental and topographic variables shape genetic structure and effective population sizes in the endangered Yosemite toad. Divers Distrib 18 (10): 1033-1041.
Weising K, Nybom H, Pfenninger M, Wolff K, Kahl G (2005) DNA fingerprinting in plants: principles, methods, and applications. CRC press.
Wu GA, Terol J, Ibanez V, López-García A, Pérez-Román E, Borredá C, ... & Talon M (2018) Genomics of the origin and evolution of Citrus. Nature 554(7692): 311-316.
Zhang XX, Liu BG, Li Y, Liu Y, He YX, Qian ZH, Li JX (2019) Landscape genetics reveals that adaptive genetic divergence in Pinus bungeana (Pinaceae) is driven by environmental variables relating to ecological habitats. BMC Evol. Biol 19(1):1-3.

No competing interests reported.

supplementary.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Notes on Sweet orange (Citrus sinensis L. Osbeck) populations’ divergence: Landscape genetics, comparative phylogeny, and Niche modeling

Status:

Version 1

Abstract

Figures

1. Introduction

2. Material and methods

2.1. Plant material

2.2. DNA extraction and PCR details

2.3. SCoT Assay

2.4. Data analyses

2.4.1. Genetic grouping and gene flow

2.4.1.1. Genetic diversity and grouping of the populations

2.4.1.2. Association studies

2.4.2. Landscape genetics

2.4.2.1. Spatial PCA (sPCA)

2.4.2.2. Random Forest (RF) analysis

2.4.2.3. Species distribution modeling (SDM)

2.4.2.4. Comparative Phylogenetic analyses

2.4.2.5. Disparity analyses

3. Results

3.1. Genetic polymorphism and grouping of the accessions

3.2. Admixture plot and assignment test

3.3. Landscape Genetics and Association studies

3.3.1. Spatial PCA (sPCA) analysis

3.3.2. Random Forest results

3.4. Species distribution modeling

4. Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1