Construction of a Core Collection of Korean Pine (Pinus koraiensis) Clones Based on Morphological and Physiological Traits and Genetic Analysis

Abstract

To improve the ecological and economic benefits of Korean pine (Pinus koraiensis), we analysed and evaluated its germplasm resources. This promotes in-depth research and utilisation of germplasm resources, providing excellent genetic resources for Korean pine breeding. We performed genetic analysis based on morphological and physiological traits and nuclear SSR molecular marker data was performed by collecting 314 clones (5 ramets of each clone) of Korean pine from eight (8) locations within the Korean pine range. The core collection underwent testing and evaluation for representativeness using variable rate (VR), coincidence rate (CR), variance difference percentage (VD), mean difference percentage (MD), Shannon index (I), and other indicators. The results indicated significant differences in morphological and physiological traits among the populations. All traits had a coefficient of variation (CV) greater than 10%, except for the water content of the needles (WC), which had an average CV of 17.636%. The populations showed high overall genetic diversity, with the HL (Helong) population exhibiting the highest genetic diversity, with an Ne (number of effective alleles), I, and He (expected heterozygosity) of 3.171, 1.103, and 0.528, respectively. Genetic variation mainly originated from individuals within populations, while the variation between populations was relatively small, at only 3%. The population did not exhibit any distinct subpopulation structures and was mainly derived from two admixed gene pools. Six core sets were obtained using different sampling strategies, and subset 6 was identified as the core collection, consisting of 114 individuals, representing a selection rate of 36.31%. In conclusion, the most appropriate method for constructing the core collection of Korean pines is the M-strategy (maximizing the number of alleles), based on both phenotypic and molecular data. The resulting core collection effectively represents the genetic diversity of the entire population effectively.

1. Introduction

Germplasm resources are essential for genetic research, enabling the exploration and utilisation of economically and ecologically important genes and traits [1]. Therefore, it is crucial to collect, preserve, evaluate, and utilise germplasm resources for breeding improved cultivars and new varieties [2,3]. Germplasm work can also be applied in theoretical studies related to plant origin, evolution, classification, and other aspects. However, the management and utilisation of germplasm resources require significant resources for appropriate assessment, particularly regarding redundant and duplicated germplasm resources [4,5,6]. Due to their large quantity, diverse structure, and incomplete information, the available potential of collected germplasm may not be fully utilised effectively. To tackle these issues, the concept of a “core collection” was further expanded in the early 1980s, introducing additional principles and methodologies for construction [7]. A core collection is a selected subset of the entire germplasm resource, chosen through certain methods, with the objective of representing the genetic diversity of the entire germplasm using the minimum number of resources required. They facilitate the rapid capture of germplasms possessing target traits and promote the effective utilisation of these resources [8,9]. Core collection resources are a novel approach to the utilisation and preservation of germplasm resources. Core collections perform a crucial function in the management and utilisation of genetic resources [10].

Initially, researchers used geographic origin and phenotypic characters to establish core collections, since these characters visually represent plant differences and are easy to measure, e.g., pomegranate Punica granatum L. [11], Ziziphus mauritiana Lam. [12], Prunus persica L. Batsch [13], etc. Nevertheless, there are some dilemmas with the method, such as the loss, incompleteness, and unreliability of germplasm and genetic data, as well as the sensitivity of phenotypic data to environmental conditions. These dilemmas mean that a core collection does not correctly represent the diversity of the original population [14,15]. Microsatellites or simple sequence repeats (SSRs) are tandemly repeated motifs of 1–6 bases found in all prokaryotic and eukaryotic genomes analysed to date [16]. These help to reveal the population structure and are often used as a tool for core collection development, with the advantage of exhibiting high polymorphisms that lead to population-specific alleles. A core collection has been developed using SSR markers in quite a few species, such as Corylus avellana L. [17]. In practice, breeders often use multiple data types to improve the quality of a core collection, which can greatly prevent the loss of key germplasm and improve the accuracy and comprehensiveness of a core collection. To name only a few, Krichen L et al. established a core collection of Prunus armeniaca L. through molecular markers and morphological traits [18]. Sun et al. developed a core collection of Litchi chinensis Sonn. based on genotypic data and agronomic traits [19]. Kumar et al. developed a core collection of Carthamus tinctorius L. using molecular, phenotypic, and geographic diversity [20]. Wang et al. used a combination of molecular markers and different phenotypic data to construct a core collection of Pinus yunnanensis [21].

Korean pine (Pinus koraiensis) is a valuable tree species in Northeastern China, serving as both an excellent source of timber and a valuable edible dry fruit and oilseed tree species [22]. The distribution area of Korean pine includes Northeastern China, the Korean Peninsula, and far southeastern Russia, with intermittent distribution in Honshu and Shikoku, Japan. In China, it is mainly found in Changbaishan, Zhangguangcailing, Laoyeling, Wandashan, and Xiaoxinganling [23]. Korean pine is a tall and straight tree with a perfectly shaped stem, high longevity, good physical quality of wood, and high productivity levels [24]. Additionally, the seeds, resin, and bark of Korean pine hold high economic value [25]. Korean pine is one of the main species of pine nuts and has a similar nutrient composition to other pine family cones such as Pinus pinea, Pinus sibirica, Pinus sinensis, Pinus cembra, Pinus edulis, and Pinus monophylla. These cones have significant food value and are highly beneficial to human health [26]. Since the 1960s, Heilongjiang Province has focused on the production and utilisation of Korean pine resources. The establishment of numerous seed orchards has provided a strong foundation for the development of Korean pine nut resources and the cultivation of nut forests [27]. Breeding research has been conducted based on Korean pine seed source zoning [28], resulting in the accumulation of abundant germplasm resources and a theoretical basis [29,30,31,32]. However, the direction and method of seed orchard construction has mainly been determined by geographical origin and growth traits. The genetic relationships and diversity of germplasm were overlooked. The repetitive and redundant materials in the germplasm resources of seed orchards has impeded the effective conservation of germplasm resources and may have also impede the germplasm resources’ evaluation and efficient use. Therefore, it is necessary to establish a core collection of Korean pine seed orchard germplasms.

The aim of this study is to enhance the effective management and utilisation of Korean pine resources. To achieve this, we constructed a core collection consisting of 314 clones from eight (8) populations, based on 11 SSR markers and nine (9) morphological and physiological traits, using various data types and strategies. The core collection was selected based on an analysis of genetic diversity parameters. The characteristics of the core collection have been described and compared to those of the entire collection. The results of this study will lay the foundation for research on good genes and molecular markers and provide the basis and valuable materials for the effective use and conservation of Korean pine resources. This study is of theoretical importance for the development of conservation strategies and breeding programs for Korean pine germplasm.

2. Materials and Methods

2.1. Plant Material

A total of 314 Korean pine clones were collected from six (6) Korean pine clones seed orchards. The assessment of genetic diversity is the basis of breeding research. Morphological variation is an important component of genetic variation. The clones were grafted from superior trees (refers to individuals with excellent growth, timber, and resistance adaptations in natural or planted forests with similar environmental conditions) from eight (8) Korean pine seed source natural forests in 1975–1980 (

Table 1. Information on the origin of clones.

Population	Regions	Source	Seed Orchard	Longitude (°)	Latitude (°)	Number of Clones
LSH	Changbaishan	Lushuihe	Lushuihe	E127.783	N42.467	21
HL		Helong	Sanchazi	E126.848	N42.626	25
SCZ		Sanchazi		E126.848	N42.626	24
XBH	Zhangguangcailing	Xiaobeihu	Bohai	E129.094	N44.057	45
HB	Xiaoxinganling	Hebei	Weihe	E128.028	N44.676	81
WY		Wuying	Hegang	E130.233	N47.066	51
LX		Langxiang	Tieli	E128.333	N47.110	40
SGL		Shangganling		E128.333	N47.110	27

). The above eight (8) natural forests cover the distribution area of Korean pine in China in the same climate zone and their germs have been well preserved. Sampling trees were planted during the installation of orchards using specific grid planting diagrams. The annual and biennial disease-free needles of five ramets from each clone were randomly collected and stored at −20 °C, protected from light.

2.2. Evaluated Traits

Needle traits were measured in biennial needles from five (5) ramets of 314 clones (

Table 2. Investigated traits information of Korean pine.

Traits	Abbreviation
Needle length (cm)	CL
Needle width (mm)	CW
Needle length/width	CL/CW
Fresh weight of needles (g)	FWC
Dry weight of needles (g)	DWC
Water content of needles (%)	WC
Chlorophyll a content in the needles (mg/g)	Chla
Chlorophyll b content in the needles (mg/g)	Chlb
Total Chlorophyll content in the needles (mg/g)	Chl

). CL was measured using a ruler with an accuracy of 0.1 cm and CW was measured using an electronic caliper, respectively. Both FWC and DWC were measured as 15-needle weights using a 10,000-point balance. Three (3) randomized samples were taken from each ramet and replicated. DWC was the weight of fresh needle dried to constant weight. Chl in needles was determined using the acetone–ethanol method [33]. The 1:1 acetone–ethanol solution was used to dissolve chlorophyll and the absorbance of the solution was determined at 665 mm and 649 mm using INESA 722s (Shanghai, China). Chlorophyll content was calculated using the following formula:

$$\mathrm{C}\mathrm{h}\mathrm{l}\mathrm{o}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{l}\mathrm{l} \mathrm{a}=\left(12.7{\mathrm{D}}_{665}-2.69{\mathrm{D}}_{649}\right)\times \frac{\mathrm{V}}{1000\times \mathrm{W}}$$

(1)

$$\mathrm{C}\mathrm{h}\mathrm{l}\mathrm{o}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{l}\mathrm{l} \mathrm{b}=\left(22.88{\mathrm{D}}_{649}-4.76{\mathrm{D}}_{665}\right)\times \frac{\mathrm{V}}{1000\times \mathrm{W}}$$

(2)

$$\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{C}\mathrm{h}\mathrm{l}\mathrm{o}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{l}\mathrm{l}=\left(20.29{\mathrm{D}}_{649}-8.04{\mathrm{D}}_{665}\right)\times \frac{\mathrm{V}}{1000\times \mathrm{W}}$$

(3)

where D₆₄₉ and D₆₆₅ are the optical density values at the corresponding wavelengths, V is the volume of the extract, and W is the fresh weight of the needles.

2.3. DNA Extraction and Genotyping

DNA of all samples were extracted from 0.2 g annual needles using the DP-320 Plant Genome Extraction Kit (Tiangen Biochemical Technology (Beijing) Co. Ltd., Beijing, China). The integrity of genomic DNA was examined using a 1% agarose gel and DNA concentration and quality were examined using MicroDrop 3000 (Shanghai BIO-DL Science Instrument Co., Ltd., Shanghai, China) after extraction. The concentration of each DNA sample was diluted to 10 ng/mL and stored at −20 °C.

SSR primers from published studies in the field and synthesised by Sangon Biotech Co. Ltd. (Shanghai, China) [29] were used. The information is shown in

Table A1. SSR primer information for estimating Genetic diversity in eight (8) populations.

Locus	Primer Sequence	Motif	Tm (°C)	Size (bp)	Fluorescent Dye
p49	F: GAGATGAGCGAATCTGGG	(AAG)7	52	261	FAM
	R: TACAAGTTCCACCTACGG
p70	F: CAACATCGCCAATGACTC	(CTCA)6	54	294	FAM
	R: CCTACCTACGCTCTGCTC
p72	F: TGGGTTACCACCTTTAGC	(GCT)6	52	193	HEX
	R: CAATCAGAGTCTGGAGCA
p79	F: CCACCGCCAAGTCCATTA	(CAA)7	55	190	HEX
	R: GCTTTGTTAGCCGTCCAG
p82	F: GGAAGATGAATCGCAAACC	(GCG)6	54	280	ROX
	R: ACACCCGCCTGAAGAGCA
EPD11	F: GTGGATGCAATGAAGAAAAACAT	(AGG)6	60	139	TAM
	R: ACGAATTGCAAAACTGCATAACT
NFPK-34	F: AACCCACAGAAAGCTGAGGA	(TAA)6	60	221	TAM
	R: CACCCCTGAACAGAGAGGAG
P6*	F: TCAAATTACCAGACAATAA	(TA)3 (GT)15	55	125	FAM
	R: GAATTCGCCAATGAAATCA
P45*	F: CTTACATTTTGCTGCTTTTC	(TG)16 (AG)17	55	173	HEX
	R: TTGTCAGTTTTAGGTTGGAT
P51*	F: CCTAAGAGCAATGTAAAATG	(AG)15	55	204	TAM
	R: AGCTTGACAACGACTAACT
P52*	F: CCATCCTTCAAATTTTCCT	(AG)26	56	138	ROX

. Each forward primer was labelled at the 5′ end with fluorescent dyes FAM, HEX, ROX, and TAMRA, respectively [34]. The PCR system consisted of 2U DNA polymerase (TransGen Biotech Co., Ltd. Beijing, China), 400μM Dntp (TransGen Biotech Co., Ltd. Beijing, China), 2.0μL of 10× buffer (TransGen Biotech Co., Ltd. Beijing, China), 0.5 μmol each of forward and reverse primer (Sangon Biotech Co. Ltd., Shanghai, China), and 10 ng DNA. The PCR procedure was as follows: 94 °C for 3 min, 35 cycles (30 s at 94 °C, 30 s at Tm, and 15 s at 72 °C), with a final extension at 72 °C for 7 min. The above procedures were executed under light-avoidance conditions. PCR products were assayed on an ABI 3730 Genetic Analyzer (Applied Biosystems, Inc., Foster City, CA, USA) by Shagon Biotech Co. Ltd. (Shanghai, China) and diploid genotype data were collected using DataCollection 3.0 software (Applied Biosystems, Inc., Foster City, CA, USA).

2.4. Data Analysis

The mean value, standard deviation (SD), coefficient of variation (CV), and analysis of variance (ANOVA) of the traits were calculated using SPSS 27.0 software (IBM SPSS Inc., Chicago, IL, USA). The diversity index (H) and principal coordinates analysis (PCoA) was performed. Morphological and physiological properties were classified into 10 levels, with values less than $\overline{\mathrm{X}}$ −2σ ($\overline{\mathrm{X}}$ is the mean value and σ is the standard deviation) as level 1 and values greater than or equal to $\overline{\mathrm{X}}$ +2σ as level 10, where every 0.5σ was classified as a level.

The Shannon-Weaver diversity index (H) of morphological and physiological traits was used, which was calculated as:

$$H=-\sum Pi\left(lnPi\right)$$

(4)

where H is the diversity index of a trait; Pi is the number of germplasm copies of the itch level of a trait as a percentage of the total germplasm material; and ln denotes the natural logarithm [35,36].

The concept of Kataring Lindgren’s (1984) equivalent latitude was adopted, which was calculated as:

$$Le=L+(A-300)\times a^{-1}$$

(5)

where Le is equivalent latitude; L is latitude; and A is altitude. When the altitude ≥300 m, a = 140. When the altitude <300 m, a = 140 [37].

GenAIEX6.51b2 (Australian National University, Acton ACT, Australian) was applied to calculate the number of alleles (Na), number of effective alleles (Ne), Shannon index (I), observed heterozygosity (Ho), expected heterozygosity (He), unbiased expected heterozygosity (uHe), and fixation index (F). Analysis of molecular variance (AMOVA), principal coordinate analysis (PCoA), and Nei genetic distances between populations were also performed in GenAIEX 6.51b2 (Australian National University, Acton ACT, Australian) [38]. Genetic structure was analysed using the admixture model in Structure 2.3.4 software (Stanford University, Palo Alto, CA, USA), with the number of populations (K) set from 1 to 10 and each K value was simulated 10 times. MCMC was set to 100,000 times for both the uncounted iterations and the MCMC after uncounted iterations. The results of the operations were imported into the Structure Harvester website (https://taylor0.biology.ucla.edu/structureHarvester/, accessed on 25 January 2024). The optimal K value was selected according to the method of Evanno et al. [39]. Significant deviations from both the Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium (LD) between all pairs of SSR loci were identified using Genepop v4.2 (Laboratiore de Genetique et Environment, Montpellier, France) [40].

To improve the retention of alleles of the original germplasm, we performed random sampling (R-) and simulated annealing sampling (S-) using different sampling strategies, as follows: M-strategy (maximizing the number of alleles), E-strategy (Entropy), and D-strategy (Diversity), respectively. We used a sampling ratio of 1%–99% by PowerMarker 3.25 (NC State University, Raleigh, NC, USA) to determine the optimal sampling method and quantity [41]. M-strategy sampling was performed using PowerCore 1.0 (National Institute of Agricultural Biotechnology, Suwon, Republic of Korea) [42] based on an improved heuristic algorithm that considers different data types. Each program was run independently 10 times. GeneAlEx 6.501 (Australian National University, Canberra, Australia) was used to calculate the genetic parameters (Na, Ne, Ho, He, and I) for each collection. To evaluate the representativeness of the core collection, mean difference percentage (MD), variance difference percentage (VD), coincidence rate (CR), and variable rate (VR) were calculated according to the method of Hu et al. [14]. The Brekin Comprehensive Evaluation Method was used to evaluate the collections, as follows:

$$Qi=\sqrt{\sum _{j=1}^n{a}_i } a_i=X_{ij}/X_{j\mathrm{m}\mathrm{a}\mathrm{x}}$$

(6)

where Qi is the comprehensive score, X_ij is the average value of an indicator, X_jmax is the optimal value of an indicator, and n is the number of evaluation indicators [43].

3. Results

3.1. Variation of Traits

Table 3. Morphological and physiological traits of Korean pine in different populations.

Traits		Population
		LSH	HL	SCZ	XBH	HB	WY	LX	SGL	Total
CL (mm)	Mean	96.0 c	117.5 a	114.7 a	104.3 b	118.4 a	105.1 b	116.8 a	114.2 a	111.8
	H	1.8	1.7	1.8	2.0	2.1	2.1	1.9	1.9	2.1
	CV/%	9.3	6.7	6.7	12.1	7.3	7.2	9.3	11.9	10.8
CW (mm)	Mean	0.7 c	0.8 b	0.8 b	0.8 b	0.8 a	0.8 ab	0.8 b	0.8 ab	0.8
	H	1.8	1.9	1.8	1.9	2.0	1.9	1.8	2.0	2.1
	CV/%	7.6	5.5	6.4	16.8	8.0	14.2	8.5	13.0	12.0
CL/CW	Mean	139.4 bc	149.3 ab	148.2 ab	139.3 bc	141.0 bc	134.0 c	154.7 a	146.5 ab	142.9
	H	1.9	1.9	2.0	1.9	2.0	1.9	2.0	1.8	2.0
	CV/%	10.1	8.4	6.5	16.5	9.6	15.5	14.5	19.1	13.8
FWC (g)	Mean	0.6 c	0.9 a	0.9 a	0.7 c	0.9 a	0.8 b	0.9 a	0.9 a	0.9
	H	2.0	1.8	1.9	1.9	2.0	2.1	1.8	1.7	2.1
	CV/%	16.3	11.8	11.7	21.7	13.2	15.3	13.6	24.9	19.6
DWC (g)	Mean	0.3 c	0.4 a	0.5 a	0.3 b	0.4 a	0.4 b	0.4 a	0.5 a	0.4
	H	2.0	1.9	1.9	1.9	2.1	2.0	1.8	1.9	2.1
	CV/%	17.8	11.8	10.3	20.0	13.0	22.5	13.6	25.6	21.3
WC (%)	Mean	56.5 a	51.5 cde	49.4 f	50.6 def	52.3 bcd	53.6 b	52.7 bc	50.1 ef	52.1
	H	1.8	1.8	1.8	1.8	1.9	1.5	2.0	1.8	2.0
	CV/%	2.4	6.8	7.9	9.5	6.6	8.7	4.1	5.8	7.7
Chla (mg/g)	Mean	1.0 ef	1.4 a	1.2 bc	1.1 cd	1.2 b	1.0 fg	1.1 de	0.9 g	1.1
	H	1.9	1.8	2.0	2.0	2.0	1.9	1.9	1.9	2.1
	CV/%	11.1	9.5	11.3	16.7	16.5	17.0	17.9	23.0	20.0
Chlb (mg/g)	Mean	0.4 bc	0.5 b	0.5 a	0.3 d	0.4 bc	0.4 c	0.4 c	0.3 d	0.4
	H	1.8	1.9	2.0	2.0	2.1	1.9	1.8	2.0	2.0
	CV/%	9.5	12.7	20.3	39.7	31.9	33.1	28.2	38.8	33.5
Chl (mg/g)	Mean	1.4 c	1.8 a	1.7 b	1.4 c	1.7 b	1.3 c	1.4 c	1.2 d	1.5
	H	2.0	1.9	1.9	2.0	2.0	2.0	1.9	2.0	2.1
	CV/%	9.8	9.3	10.9	19.0	16.6	19.5	19.3	16.8	20.0
Mean	H	1.9	1.9	1.9	1.9	2.0	1.9	1.9	1.9	2.1
	CV/%	10.4	9.2	10.2	19.1	13.6	17.0	14.3	19.9	17.6

Kolmogorov–Smirnov test prove that 1. the data are normally distributed; 2. there is homogeneity of variance; and 3. the observations are independent of each other. There is a significant difference at α = 0.05 between the mean value without same letters (a–f).

shows the results of genetic variation for nine (9) traits. The coefficient of variation (CV) for each trait was on average 17.64%, with all traits having a CV greater than 10%, except for water content (WC). The highest level of variation was found in Chlb (33.45%). Similarly, the value of H for all traits was greater than 2.00, except for water content. The average H for each trait was 2.052, with the highest H found in CW (2.089). The highest average H of each trait was found in HB (2.015), while all other populations had values less than 2. The CV was less than 20% for all groups, with the highest CV (19.87%) found in SGL and the lowest H and CV in HL, which were 1.851 and 9.170%, respectively. Morphological and physiological traits from different populations in the same orchard tended to behave similarly. There were no significant differences in CL, CW, CL/CW, FWC, or DWC between HL and SCZ. However, there were significant differences in CL, FWC, and DWC between LX and SGL. The given statement indicates that the morphological and physiological traits were influenced not only by genetic factors but also by the environment in which they were cultivated.

PCoA of the morphological and physiological traits for each group produced comparable results to those of the CV (Figure 1). HL had the smallest distribution range, while SGL, XBH, and WY, which had a large CV, had a large distribution range.

3.2. Genetic Diversity

The results of the genetic diversity analysis of the populations revealed that 11 SSR markers identified 93 alleles in 314 Korean pine clones, with an average Ne of 3.067. It is important to note that these results are objective and based solely on the data presented in

Table 4. Genetic diversity of eight (8) Korean pine populations, resulting from 11 loci.

Pop	N	Na	Ne	I	Ho	He	uHe	F
LSH	21	52	2.704	1.009	0.455	0.518	0.531	0.106
HL	25	59	3.171	1.103	0.418	0.528	0.538	0.169
SCZ	24	52	2.778	1.006	0.428	0.501	0.512	0.082
XBH	45	62	2.695	1.049	0.459	0.499	0.504	0.056
HB	81	74	2.859	1.066	0.479	0.495	0.498	−0.002
WY	51	61	2.94	1.07	0.467	0.513	0.518	0.057
LX	40	65	2.785	1.07	0.457	0.507	0.513	0.112
SGL	27	54	2.244	0.956	0.424	0.482	0.491	0.128
Mean	314	93	3.067	1.153	0.456	0.522	0.523	0.11

. The genetic diversity analysis of populations revealed that 11 SSR markers identified 93 alleles in 314 Korean pine clones, with an average Ne of 3.067. The highest I was HL (1.103) and the I of the total population was 1.153. HL also had the largest Ne (3.171), He (0.528), uHe (0.538), and F (0.169), indicating that HL had the highest level of genetic diversity with heterozygosity deficiency. SGL exhibited the lowest level of genetic diversity, with the smallest Ne of 2.244, I of 0.956, He of 0.482, and uHe of 0.491. The smallest F of −0.002 was observed in HB, which indicates that the HB had the most reasonable mating system and did not have inbreeding. Positive values of FIS indicate that mates are more closely related on average than expected, and also indicate a deficit of heterozygotes relative to Hardy–Weinberg proportions. This is most likely due to the mean values shown in Table 5. This could be an artificial effect from a single locus that shows high linkage disequilibrium.

3.3. Analysis of Genetic Structure

The AMOVA results (Table 5) indicate that only 3% of the variation occurred among the eight (8) populations, while 87% of the variation was due to differences among individuals, and the remaining 10% was due to differences among the materials within each population at the significance level of p ≤ 0.001. The Nei genetic distance between populations was unequal (Figure 2), with the largest genetic distance being 0.161 between LSH and HL, and the smallest genetic distance being 0.015 between WY and XBH.

Table 5. Results of AMOVA of Korean pine populations.

Source	df	SS	MS	Est. Var.	%
Among Pops	7	66.152	9.450	0.084	3%
Among Indiv.	306	950.221	3.105	0.298	10%
Within Indiv.	314	788.000	2.510	2.510	87%
Total	627	1804.373		2.891	100%

df: degrees of freedom. SS: sum of squares. MS: mean of the squares. Est. Var.: estimated variance of components. %: percentage of total variance contributed by each component. p ≤ 0.001.

Table 5. Results of AMOVA of Korean pine populations.

Source	df	SS	MS	Est. Var.	%
Among Pops	7	66.152	9.450	0.084	3%
Among Indiv.	306	950.221	3.105	0.298	10%
Within Indiv.	314	788.000	2.510	2.510	87%
Total	627	1804.373		2.891	100%

df: degrees of freedom. SS: sum of squares. MS: mean of the squares. Est. Var.: estimated variance of components. %: percentage of total variance contributed by each component. p ≤ 0.001.

Figure 3 displays the results of the genetic structure analysis. The analysis revealed that the germplasm mainly originated from two gene pools, with a maximum at K = 2 in the range of 2–10 for ΔK. Gene pool 1 was the main contributor to LSH and HL, while gene pool 2 was the main contributor to SGL, consistent with the results of the genetic distance analysis. The other populations did not exhibit any significant subpopulation structures and were genetically characterized by a mixed composition of the two gene pools. This finding aligns with the results of AMOVA, indicating that there is only a minor genetic variation between populations, which indicates the frequent gene flow among the populations. This may be related to the small range of distribution and the long evolutionary history of Korean pine.

3.4. Construction of a Core Collection

Figure 4 shows the results of the varying sampling strategies and ratio using PowerMaker 3.25. Regardless of the strategy, R- had smaller values for Na (number of alleles), GE (gene entropy), and GD (gene diversity) than S-. Na increased with the sampling rate in different strategies, reaching a maximum at a sampling ratio of 32% (100) in the E-strategy. The E-strategy repeated the construction 1000 times at a 32% sampling proportion. Based on the frequency of each sample, the 100 samples with the highest frequency were selected as Core set 1. The maximum GE occurred at a sampling rate of 5% (16) in the E-strategy, while the maximum GD occurred at a sampling rate of 3% (9) in the D-strategy. Coresets 2 and 3, containing 16 and 9 individuals, respectively, were constructed in the same manner. Core sets 4, 5, and 6 were constructed using molecular markers, morphological and physiological traits, or both, after 10 independent repetitions using the M-strategy of PowerCore 1.0. The sample sizes for core sets 4, 5, and 6 were 103, 18, and 114, respectively.

3.5. Confirmation and Evaluation of Core Collection

The study initially selected six (6) core sets. The aim was to identify a core collection that could represent the entire population. To verify the validity of the core sets and identify the core collection, a comprehensive comparison of the six (6) core sets and the entire population was carried out. This was carried out following the principle of restoring the genetic level of the entire population as much as possible (

Table 6. Comparison of six (6) core sets with original germplasm.

Core Sets	1	2	3	4	5	6	Entire Pop
Na	93.000	75.000	45.000	93.000	52.000	93.000	93.000
Ne	3.197	4.479	2.424	3.749	2.596	3.687	3.067
I	1.156	1.453	0.881	1.287	0.971	1.274	1.153
Ho	0.460	0.568	0.465	0.493	0.444	0.494	0.456
He	0.528	0.668	0.434	0.571	0.479	0.567	0.522
MD/%	1.521	3.746	6.529	0.935	4.468	1.224	0.000
CR/%	92.609	52.091	49.569	83.724	94.364	98.788	100.000
VD/%	8.766	37.769	33.710	11.537	128.377	12.810	0.000
VR/%	105.602	77.714	101.338	94.532	148.626	105.131	100.000
H	2.006	1.876	1.523	2.038	1.891	2.027	2.051
Score	2.537	2.549	2.072	2.622	2.551	2.652

). The study found that core sets 1, 4, and 6 retained all alleles of the entire population (93). Core set 4 had the smallest MD (0.953%) and the largest value of H (2.038). Coreset 2 had the highest value of Ne (4.479), I (1.453), He (0.568), and Ho (0.668), but its CR and VR were only 52.091% and 77.714%, respectively. The largest CR and VR were observed in set 6 (98.788%) and set 5 (148.626%), respectively. The highest composite score was obtained for core set 6 (2.652), thus identifying set 6 as the core collection. The genetic diversity parameters of core set 6, such as Ne, I, and Ho, were comparable and slightly greater than those of the entire population. The means, variances, CV, and range of morphological and physiological traits of core set 6 were also similar to those of the entire population. Therefore, core set 6 can effectively restore the genetic diversity and morphological and physiological variation of the entire population.

The PCoA of the core collection and the entire population revealed that the individuals of the core collection were uniformly distributed in the entire population. Both the morphological and physiological traits and the molecular markers of the core collection accurately represent the entire population. The variation and genetic diversity of the entire population could be restored (Figure 5a). The core collection accounted for 36.31% of the total, comprising 114 individuals from 8 populations (Figure 5b). The inner ring displays the proportion of each population in the core collection, while the outer ring shows the proportion of each population in the entire population. In the core collection, HB accounted for the largest proportion at 20.18%, while SGL accounted for the smallest proportion at 6.14%. However, both proportions were smaller than those in the entire population. The proportions of HL, LX, LSH, SCZ, and XBH in the core collection were larger than those in the entire population. The results above indicate that the HL and other populations exhibit a high level of genetic variation, whereas the SGL and HB populations exhibit a lower level of genetic variation. This finding is consistent with the data.

4. Discussion

Germplasm resources are essential for breeding. It is crucial to understand the genetic diversity and variation of germplasm resources for the conservation and sustainable use of plant resources [44,45]. The greater the range of tree species, the higher the genetic variation, leaf phenology, and physiological variation [46,47]. Among the populations, Chlb exhibited the highest CV. The average CV for each trait was highest in SGL among the populations. The degree of genetic diversity is higher when the CV of quantitative traits in germplasm resources is greater. The results of PCoA indicate that there is a high degree of overlap between different populations from the same orchard. Therefore, it is necessary to use multiple site materials to build the core collection.

Molecular markers are based on DNA polymorphisms, which are not biologically active and are rarely affected by the growth period of the plants or the external environment. They are therefore more suitable than morphological markers for constructing core collections and assessing genetic diversity. SSRs are co-dominant markers that provide richer allelic information than other molecular markers. In this study, Ne and I were higher than those in the study of Feng Fujuan et al. [48]. This study found that the use of SSR markers resulted in a higher efficiency compared to other types of molecular markers. The population with the highest genetic diversity was HL, while the lowest was SGL. The HB population had the lowest F (−0.002), indicating the presence of inbreeding in the seed orchard. All other populations had an F greater than zero, suggesting heterozygous deficiencies. These deficiencies may be related to the size of the original population and the procedures and criteria used to select superior trees. AMOVA revealed that despite the populations having different origins, interpopulation genetic variation accounted for only 3% of the total genetic variation. These results suggest that the interpopulation variation is insignificant and there is no clear genetic differentiation. The most genetically distant relationships are between HL and SGL.

The construction of a core collection typically involves four steps, data collection and organization, grouping of materials, determination of sampling strategy, and testing and evaluation of the core collection [49]. Observing genetic diversity is crucial for conserving germplasm resources and for breeding work. Sets 1, 2, and 3 compared the characteristics of collections constructed using different strategies. Both the E-strategy and M-strategy were able to maintain the Na, but only the E-strategy significantly improved the Ne, I, Ho, and He of the core collection. The D-strategy was able to restore the Ho of the original germplasm by using the least amount of germplasm, making it more suitable for the construction of breeding populations. Collections 4, 5, and 6 compared the quality of the core collection constructed using different data types. The results show that the core collection, constructed using multiple data types not only retained the Na, but also the morphological and physiological traits of the entire population. Previous studies have also demonstrated that constructing a more robust and reliable core collection using multiple data types can significantly prevent the loss of critical germplasm and improve the accuracy and comprehensiveness of the core collection [6]. Genetic analyses suggest that there is no clear genetic structure or significant genetic differentiation between geographic sources of Korean pine clones. Therefore, an appropriate sampling strategy is required to construct a core collection. The construction of a core collection depends on its purpose [17]. In recent years, the M-strategy has been widely used for this purpose. The M-strategy selects materials with high allele abundance and low redundancy by maximizing the number of alleles at each locus. This approach can preserve rare and local alleles [50,51]. The core collection constructed using M-strategy was able to restore more alleles and also restore the level of genetic diversity of the original germplasm under the same sample size.

The study found that the sampling rate of the core collection of Korean pine was at 36.31%, which is higher than that of Schima superba (15.3%), but with a comparable core collection size (115) [52], and lower than that of Juglans regia L. (44.23%) [53]. These results suggest that the sampling rate of the core collection varies depending on the special characteristics. The evaluation of the core collection depends on various factors, including the size and accessibility of the original germplasm, the similarity of the germplasm, the number and relevance of traits investigated, and the sampling strategy [54]. In addition to parameters such as MD, VD, Na, and Ne, the evaluation of morphological and physiological traits has introduced H. The core collection was determined using the Brekin Comprehensive Evaluation Method, which calculates composite scores for multiple traits in superior germplasm [43,55]. This method was also found to be suitable for the comprehensive evaluation of the core set. To compare distributional characteristics between the core collection and the original population, PCoA was used [56]. The distribution of molecular markers and morphological and physiological traits was consistent between the core and original groups, indicating no significant differences between the two. The highest proportion of the core collection was from HB, while the lowest was from SGL, which is related to the population size. The proportion of the core collection in each population of the original population reflects the redundancy of the population. SGL, which has the smallest proportion, exhibits the lowest genetic diversity. The higher proportion observed in LSH may be attributed to its limited population size, in addition to its rich genetic variation.

5. Conclusions

Establishing a core collection of Korean pine based on molecular and phenotypic traits can enhance the efficiency of its genetic resources. This study evaluated 314 clones from eight (8) Korean pine populations in China for genetic diversity using SSR markers and morphological and physiological traits. A core collection of 114 Korean pine clones was constructed, covering all eight (8) populations and representing the diversity of the entire population. The technology has wide-ranging applications in future breeding work. Firstly, it can be used for genomic research and molecular marker development to promote the conservation and utilization of germplasm resources. This includes genetic diversity evaluation, identification and classification of germplasm, genetic map construction, and the development of specific genes. Secondly, it can be used for genome-wide association research, such as the identification of disease-resistant genes and the discovery of superior genes. It also plays an important role in diversifying the genetic basis of germplasm materials and developing a breeding program.