Exploring elite alleles based on association analysis of economic traits will promote the breeding, development, and utilization of Prunus sibirica resources. This study used 113 P. sibirica clones with diverse provenances to perform association mapping of 20 economic traits based on 149 pairs of highly polymorphic simple sequence repeat (SSR) markers. Population structure analysis split the clones into three subgroups. General linear model (GLM) and mixed linear model (MLM) identified 117 SSR loci associated with 20 P. sibirica economic traits, specifically fruit (62 loci), nuclear (69), and kernel (46) traits. Phenotypic variation in these three sets of traits ranged 8.93–82.12% (GLM) or 0.23–84.82% (MLM), 10.96–56.65% (GLM) or 0.17–52.54% (MLM), and 5.17–59.41% (GLM) or 4.86–59.89% (MLM), respectively. Of the 117 marker SSR loci, 35 were simultaneously associated with two or more economic traits. In addition, phenotypic effects analysis identified 1842 elite alleles and seven typical materials related to key economic traits. We then designed elite parental combinations through crossing the typical materials. In conclusion, our study detected alleles and plant materials that allowed the identification of clones expressing desired traits of interest. These findings can be used for improved, targeted breeding of P. sibirica through marker-assisted selection.
Research Article
Association mapping of major economic traits and exploration of elite alleles in Prunus sibirica
https://doi.org/10.21203/rs.3.rs-1457343/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
Prunus sibirica
simple sequence repeats
microsatellite markers
association mapping
phenotypic effect analysis
Prunus sibirica (Rosaceae) is a deciduous shrub mainly distributed in Mongolia, eastern Siberia, and northern to northeastern China (Ma et al. 2020). The species has ecological importance, widely used for wind and sand fixation, as well as soil and water conservation (Yin et al. 2019). Furthermore, P. sibirica seed kernel is high in oil (50%) that can be used to produce cooking oil, biodiesel fuel, surfactants, almond milk, skin care, and healthcare products (Mai et al. 2020). The seeds are also an ingredient in traditional Chinese medicine (Yan et al. 2018). At present, P. sibirica covers approximately 1.7 million ha and yields nearly 192,500 tons of seeds per year in China (Niu et al. 2018).
With rich germplasm resources, P. sibirica breeding has the potential for generating improved phenotypes. Current P. sibirica yield and quality are significantly reduced by issues such as pistil abortion, vulnerability to freezing injury at the flowering stage, and the bad management. Therefore, the selection and cultivation of appropriate varieties is particularly important. To generate desired cultivars quickly, however, breeding efficiency must be improved, as traditional breeding methods have a very long working cycle and are strongly season-dependent.
Several types of DNA markers are currently available for breeding programs, including RFLP markers (Duval et al. 2001), PCR markers (Ruas et al. 2001; Paz et al. 2005; Feng et al. 2013), and new markers single nucleotide polymorphisms (SNP) (Ravelombola et al. 2017). The most common marker is simple sequence repeats (SSR) or microsatellites, DNA sequences consisting of repeated motifs of 1–6 bp that are widespread distributions biological genomes (Kozlowski et al. 2010). Multiple advantages make SSR markers widespread in research: it is abundant, highly polymorphic, multiallelic, inherited codominant, easy to detect, highly informative, stable, and comparatively inexpensive (Meng et al. 2018). Many SSR markers have been applied to research on Prunus spp., including DNA fingerprinting maps (Li et al. 2019), genetic diversity analysis (Samarina et al. 2022), and linkage analysis (Bhattarai and Mehlenbacher 2018).
With advancements in genomics and improvements to biostatistics software, association mapping has become an important approach for the genetic dissection of quantitative traits (Yeon et al. 2021). This analysis links molecular markers to a target phenotype based on linkage disequilibrium (Zondervan et al. 2004). Additionally, it detects the contributions that different alleles of germplasms to the target trait, facilitating fine mapping of target genes and dramatically improving trait detection efficiency (Xu et al. 2020). For example, Thornsberry et al. (2001) successfully applied association mapping to evaluate markers associated with maize flowering time, and similar studies have followed in corn (Kim et al. 2017), peas (Jha et al. 2020), olives (Ipek et al. 2015), and coconuts (Wei et al. 2014). Several other studies have used association mapping to locate trait-associated loci, then applied phenotypic effect analysis to find favorable alleles and typical materials, including in cotton (Du et al. 2016), rice (Edzesi et al. 2016), coconut (Zhou et al. 2020), poplar (Wei et al. 2010), and China fir (Zhang et al. 2015). However, association mapping and phenotypic effect analysis has not been performed in P. sibirica.
Thus, our study applied this combined analysis to identify elite alleles and typical material from 113 P. sibirica clones of different regions, using 149 SSR marker pairs. Our results will provide a theoretical basis for molecular breeding studies that aim to manipulate the quality of economic traits in P. sibirica, laying the foundation for subsequent development and utilization of plant resources.
Plant materials
The study used 113 P. sibirica clones supplied by the National Forest Germplasm Resources Preservation Repository for P. sibirica in Kazuo, Liaoning Province, P.R. China. Of these, 16 were from Russia and 97 were from China (21 clones from Kazuo, Liaoning; 11 clones from Chaoyang, Liaoning; 19 clones from Beipiao, Liaoning; 12 clones from Aohan Banner, Inner Mongolia; and 34 clones from Zhalantun, Inner Mongolia) (Table 1).
|
Provenances |
Clone No. |
|---|---|
|
Kazuo, Liaoning |
1, 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 17, 18, 19, 20, 21, 22, 23 |
|
Aohan banner, Inner Mongolia |
24, 25, 27, 28, 31, 32, 33, 34, 35, 36, 37, 39 |
|
Chaoyang, Liaoning |
26, 29, 30, 41, 42, 43, 44, 45, 46, 48, 53 |
|
Beipiao, Liaoning |
47, 49, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 70, 71, 72, 73, 81, 95, 99 |
|
Zhalantun, Inner Mongolia |
322, 324, 328, 329, 332, 334, 339, 341, 345, 350, 354, 358, 366, 367, 368, 375, 381, 401, 404, 405, 406, 407, 408, 409, 442, 443, 449, 455, 457, 459, 460, 462, 463, 464 |
|
Russia |
94, 501, 502, 503, 504, 505, 506, 507, 508, 509, 510, 511, 513, 516, 517, 518 |
Measurement of economic traits
Twenty economic traits were measured during the 2020 growing seasons: fruit length (FL), fruit width (FW), fruit thickness (FT), flesh thickness (FLT), fruit shape index (FSI), single fruit weight (SFW), fruit yield (FY), nuclear length (NL), nuclear width (NW), nuclear thickness (NT), nuclear shell thickness (NST), nuclear shape index (NSI), single nuclear weight (SNW), nuclear rate (NR), kernel length (KL), kernel width (KW), kernel thickness (KT), kernel shape index (KSI), single kernel weight (SKW), and kernel rate (KR).
DNA amplification and SSR marker genotyping
In 2020, mature and fully expanded leaves were collected and stored at − 80°C. Genomic DNA was extracted from leaf tissue with CTAB. The P. sibirica genome was used to design 600 SSR primer pairs, which were synthesized by SBS Genetech Co., Ltd. For association mapping and population structure analysis, 149 highly polymorphic primers were selected. The PCR reaction solution (20 µL) comprised: 1 µL of DNA template, 0.15 µL of SSR primers, 2 µL of 10× PCR buffer (including Mg2+), 0.4 µL of Taq enzyme, 1.5 µL of dNTP, and 13.1 µL of ddH2O. Thermocycling conditions were as follows: 94°C for 5 min; 34 cycles of 94°C for 30 s, 56°C for 30 s, 72°C for 30 s; and extension at 72°C for 5 min. Amplicons were stored at 4°C, then separated using non-denaturing 12% polyacrylamide gel electrophoresis before silver staining and imaging.
Population structure analysis
Structure 2.3 (Pritchard et al. 2000) with the admixture model was used to analyze population structure and estimate the K value (optimal number of subgroups). Parameter settings were as follows: kinship (K) value between 1 and 10 with a replication of 10, burn-in period of 10 000 for the Markov chain Monte Carlo (MCMC) iteration, followed by 100,000 simulations; all other parameters were set to default. ΔK was calculated based on lnP (D), and an appropriate K value was selected to obtain the corresponding population structure (Q) matrix.
Association mapping
To reduce the influence of population structure on association mapping, mixed linear model (MLM) and general linear model (GLM) were used to determine associations between genotypes and phenotypes in TASSEL 2.1 (Bradbury et al. 2007). For MLM, multiple levels of population structure (Q-matrix) and relative kinship among individuals (K-matrix) were simultaneously considered. For GLM, the Q matrix was selected as the fixed covariate. The P-value and phenotypic variation explained by a given marker (R2) for each locus were calculated.
Exploration of elite alleles and germplasm clones
Based on the results of association mapping, alleles of loci significantly associated with target traits were further analyzed using the null allele method proposed (Breseghello and Sorrells 2006). Average phenotypic values were compared between clones with a given allele and all clones to obtain the phenotypic allele effect (ai). The calculation formula is as follows:
where ai is the mean of the th allelic phenotypic effect value, xij is the j th material phenotypic value of the i th allele, Nk is the number of materials with the th allele, is the th material phenotypic value of all materials, and nk is the number of materials. An allele is synergistic if and negative if .
Population structure analysis
Population structure analysis of the 113 P. sibirica clones using 149 polymorphic markers showed that when K = 3, ΔK peaked at its maximum value (Fig. 1), splitting the test population into subgroups 1, 2, and 3 (Fig. 2). Subgroup 1 contained 27 clones, with seven from Kazuo, two from Aohan Banner, four from Chaoyang, 11 from Beipiao, and three from Zhalantun. Subgroup 2 contained 43 clones: four from Kazuo, 10 from Aohan Banner, three from Chaoyang, 12 from Beipiao, and 14 from Russia. Subgroup 3 contained 38 clones: eight from Kazuo, four from Chaoyang, five from Beipiao, 19 from Zhalantun, and two from Russia. Based on the results when K=3, the corresponding Q matrix was used for marker-trait association mapping.
Marker-trait association mapping
Association mapping of 149 SSR markers and P. sibiricaeconomic traits detected 177 significant marker-trait associations (p< 0.05), with R2 ranging from 5.17% to 82.12% (average: 27.93%) for GLM and from 0.17% to 84.82% (average: 21.80%) for MLM. The locus that explainedthe largestproportion of phenotypic variation was X38 associated with SFW (82.12%, GLM, 84.82%, MLM), followed by X104 (59.41%, GLM) and X34H (59.89%, MLM). Thelocus that explained the least phenotypic variation was X37 associated with NR (5.17%, GLM, 0.17%, MLM), followed by X67 (8.92%, MLM) and P2 (0.62%, MLM). Among the 117 marker-trait associations, 62 markers were associated with fruit traits, 69 with nuclear traits,and 46with kernel traits. Their corresponding R2 ranges from GLM were 8.93%~82.12%, 10.96%~56.65%, 5.17%~59.41%, R2 ranges from MLM were 0.23%~84.82%, 0.17%~52.54%, 4.86%~59.89%. Among these significant SSR markers, 35 were simultaneously associated with two or more different traits, suggesting tightly linked genes or pleiotropy. We used all 117 marker-trait associations to explore elite alleles (Table 2).
Table 2 SSR loci associated significantly with economic traits
|
Trait |
Number of loci |
Marker loci |
R2 ranges |
|
|
GLM |
MLM |
|||
|
FL |
11 |
X112, X115, X16H, X18, X20, X27, X36, X69, X85, Y22, Y87, |
12.54%~43.74% |
0.23%~43.86% |
|
FW |
8 |
P2, X112, X16H, X18, X20, X27, X36, X40 |
16.47%~48.71% |
4.88%~46.59% |
|
FT |
8 |
L20H, P2, X112, X16H, X18, X20, X36, X69 |
19.18%~53.16% |
10.40%~49.89% |
|
FLT |
10 |
L2, X112, X115, X15H, X16H, X17H, X18, X20, X27, X40H |
14.52%~53.92% |
0.91%~37.67% |
|
FSI |
7 |
L34, L49H, X10, X11, X116, X20, X30H |
8.93%~34.21% |
7.80%~30.81% |
|
SFW |
9 |
L20H, L39, P2, X15H, X18, X20, X36, X38, X69 |
11.67%~82.12% |
11.46%~84.82% |
|
FY |
9 |
L49H, P3, X103, X106, X20H, X40H, X85, X94, X98 |
13.83%~58.36% |
7.46%~39.14% |
|
NL |
11 |
P2, X115, X136, X140, X15H, X16H, X17H, X18, X20, X64, X86 |
12.88%~45.56% |
0.62%~23.17% |
|
NW |
12 |
L39, L75, P3, X136, X17H, X19H, X40, X64, Y26, Y28, Y50, Y86 |
13.90%~47.16% |
5.83%~40.27% |
|
NT |
5 |
L20H, L34, X19H, X74, X98 |
16.38%~28.61% |
14.86%~27.54% |
|
NST |
10 |
L25, L33, L49, X116, X134, X17H, X19H, X36, X37, Y22 |
13.61%~56.65% |
11.58%~52.54% |
|
NSI |
19 |
L20H, L49H, L73H, L75, P56H, X116, X134, X138, X17H, X19H, X27, X37, X38H, X40, X64, X98, Y22, Y50, Y87 |
10.96%~51.81% |
8.32%~45.25% |
|
SNW |
5 |
L68, P2, X20, X69, X8H |
18.02%~34.41% |
17.93%~24.92% |
|
NR |
7 |
L49, L70H, X11, X18, X27, X37, X76 |
18.51%~42.33% |
0.17%~41.95% |
|
KL |
13 |
L39, L7, P2, X104, X115, X116, X17H, X18, X20, X36, X39, X41, Y87 |
10.84%~59.41% |
7.80%~42.15% |
|
KW |
5 |
X104, X41H, X58, X92, Y87 |
17.37%~51.14% |
17.23%~49.85% |
|
KT |
4 |
L20H, X12, X42H, X93 |
14.48%~30.05% |
9.14%~27.83% |
|
KSI |
9 |
L17, L49H, X10H, X11, X20, X27, X30H, X54H, X67 |
8.92%~28.39% |
7.79%~28.26% |
|
SKW |
5 |
L45H, P2, X20, X34H, Y65 |
5.17%~57.37% |
4.86%~59.89% |
|
KR |
10 |
L28, L75, X101, X20, X29H, X42H, X47, X64, X70, X83 |
18.12%~33.22% |
16.37%~34.46% |
Exploration of elite alleles
We investigated 1842 elite alleles, 820 of which were negative and 1022 were synergistic. Phenotypic effects ranged from -245.07 to 803.11. We found 1401 elite alleles that were simultaneously associated with more than two traits. For example, eight elite alleles were associated with 12 traits, 12 with eight traits, nine with seven traits, and 61 with six traits. (All allele effects (ai) andtypical materials are listed in S1 Table.) Thenumber of elite alleles for fruit traits FL, FW, FT, FLT, FSI, SFW, and FY were 106, 79, 89, 95, 58, 121, and 70, respectively.The number of elite alleles for nuclear traits NL, NW, NT, NST, NSI, SNW, and NRwere 93, 124, 52, 149, 228, 50, and 99, respectively.The number of elite alleles for kernel traits KL, KW, KT, KSI, SKW, and KRwere 139, 46, 38, 74, 50, and 82, respectively. Table 3 presents the largestphenotypic effect and most typical material for each trait.
Table 3 Phenotypic effects of alleles of loci associated
|
Traits |
Elite allele |
ai |
Typical materials (Clone No.) |
|
FL |
X18-126 |
-0.5843 |
1 |
|
FL |
X36-144 |
-0.0010 |
11,49,95,99,442,443,449,457,511 |
|
FL |
Y22-111 |
0.0008 |
17,24,26,339,341,345,381,404 |
|
FL |
X36-174 |
1.3446 |
94 |
|
FW |
X18-126 |
-0.6243 |
1 |
|
FW |
X20-140 |
-0.0010 |
322,324,334,381 |
|
FW |
X27-135 |
0.0211 |
18,19,21,53,62,71,339,345,350,516,517,518 |
|
FW |
X27-162 |
1.3186 |
464 |
|
FT |
X69-116 |
-0.5300 |
324 |
|
FT |
X69-144 |
-0.0067 |
42,48,375,408,443,455,457,459,460,462,464,511 |
|
FT |
X18-118 |
0.0018 |
59,60,61,407 |
|
FT |
X36-174 |
1.4377 |
94 |
|
FLT |
X18-126 |
-1.8357 |
1 |
|
FLT |
X27-132 |
-0.0491 |
22,328,334 |
|
FLT |
X27-159 |
0.0105 |
14,95,332,367,375,401,405,406 |
|
FLT |
X40H-138 |
5.5750 |
94 |
|
FSI |
X10-144 |
-0.1316 |
464 |
|
FSI |
X10-165 |
-0.0016 |
341 |
|
FSI |
X30H-147 |
0.0002 |
43,46,48,95,322,328,367,368,381,401,404,406,408,442,443 |
|
FSI |
L49H-129 |
0.2100 |
409 |
|
SFW |
X69-116 |
-4.3900 |
324 |
|
SFW |
X15H-126 |
-0.0320 |
4,6,7,20,39,41,43,58,71,81 |
|
SFW |
X15H-148 |
0.0458 |
8,10,14,18,25,27,31,32,322,324,329,334,339,341,345,354, 366,381,405,406,407,408,443,501,505,506 |
|
SFW |
X36-174 |
18.8200 |
94 |
|
FY |
X20H-124 |
-245.0905 |
18,20 |
|
FY |
X40H-128 |
-0.7700 |
24 |
|
FY |
X98-138 |
2.9100 |
509 |
|
FY |
X98-158 |
803.1100 |
95 |
|
NL |
X18-126 |
-0.4229 |
1 |
|
NL |
X136-144 |
-0.0071 |
2,11,13,28,35,49,53,339,341,404,459,463,510,516 |
|
NL |
X17H-142 |
0.0026 |
28,42,53,58,61,71 |
|
NL |
X115-111 |
0.4377 |
43 |
|
NW |
X19H-148 |
-0.3890 |
12,45 |
|
NW |
X136-156 |
-0.0008 |
1,5,17,18,19,21,22,25,27,31,32,33,34,35,36,37,39,41, 46,47,62,71,73,322,324,328,334,345,354,358,381,501, 502,503,504,505,507,509,513,517,518 |
|
NW |
X17H-134 |
0.0016 |
17,30 |
|
NW |
X40-180 |
0.3258 |
43 |
|
NT |
X98-147 |
-0.1450 |
2,55 |
|
NT |
X74-134 |
-0.0033 |
19,358 |
|
NT |
L34-144 |
0.0021 |
1,339,345,507,509,518 |
|
NT |
L20H-158 |
0.3120 |
22 |
|
NST |
X37-146 |
-0.5929 |
407 |
|
NST |
L33-132 |
-0.0077 |
2,4,15,17,26,27,32,36,37,42,43,44,58,59,62,95,401,513,517 |
|
NST |
X19H-150 |
0.0090 |
1,4,6,8,10,11,13,14,18,19,20,21,22,23,27,28,31,32,33,34, 35,37,41,42,43,46,49,53,55,62,71,72,81,95,99,322,324,328, 329,332,334,339,341,345,350,354,358,366,375,381,401, 404,405,406,407,408,409,442,443,455,457,459,460,462, 463,464,501,502,503,504,505,506,507,508,509,510,513, 516,517,518 |
|
NST |
X116-99 |
1.0168 |
45 |
|
NSI |
L20H-158 |
-0.1580 |
22 |
|
NSI |
X138-126 |
-0.0007 |
30,61,63 |
|
NSI |
L20H-132 |
0.0010 |
1,6,10,13,21,25,28,31,32,33,34,35,37,72,81,95,366,381,462,463,464 |
|
NSI |
X38H-162 |
0.4400 |
12 |
|
NSW |
X69-118 |
-0.5133 |
25,32 |
|
NSW |
X69-146 |
-0.0100 |
29,366,367 |
|
NSW |
P2-138 |
0.0144 |
4,23,42,328,358,405 |
|
NSW |
X8H-159 |
0.4461 |
5 |
|
NR |
X11-111 |
-0.1400 |
73 |
|
NR |
X37-136 |
-0.0007 |
24,95 |
|
NR |
L70H-156 |
0.0009 |
7,13,18,21,22,23,25,26,27,31,32,33,34,35,37,39,41,42,43, 44,45,48,49,53,55,72,94,322,324,329,332,334,339,341,345, 354,358,375,381,401,406,407,408,442,443,455,457,459, 460,462,464,501,502,509 |
|
NR |
L70H-117 |
0.1400 |
58,70 |
|
KL |
X36-104 |
-0.2603 |
63 |
|
KL |
X41-123 |
-0.0003 |
5,7,9,12,15,17,20,24,26,29,45,47,63,70,73 |
|
KW |
X104-208 |
-0.1280 |
63 |
|
KW |
X58-116 |
-0.0035 |
23,29,42,55,405,464 |
|
KW |
X41H-120 |
0.0040 |
21,22,23,25,27,28,32,34,35,37,41,42,43,502,506,507 |
|
KW |
X58-122 |
0.3632 |
366 |
|
KT |
X42H-156 |
-0.0622 |
405,409 |
|
KT |
X12-132 |
-0.0011 |
5,7,9,15,17,20,26,29,30,36,45,47,54,56,57,58,59,60,63,70,73,329,368,407 |
|
KT |
X12-145 |
0.0002 |
11,18,27,31,35,39,49,53,503,504,508,510,513 |
|
KT |
L20H-140 |
0.1350 |
61,63 |
|
KSI |
X10H-120 |
-0.2339 |
339 |
|
KSI |
L17-108 |
-0.0002 |
30,36,59,61,63,70,73 |
|
KSI |
X54H-110 |
0.0036 |
17,24,47,70,73,367,368,381 |
|
KSI |
X11-123 |
0.4900 |
409 |
|
SKW |
X34H-154 |
-0.0981 |
39 |
|
SKW |
Y65-128 |
-0.0013 |
15 |
|
SKW |
Y65-116 |
0.0002 |
1,6,8,11,13,18,19,21,28,34,41,42,43,44,46,49,53,55,56,57, 58,59,60,61,62,63,71,95,328,339,345,350,354,358,404,408, 501,502,503,504,505,508,509,513,516,517,518 |
|
SKW |
P2-159 |
0.2422 |
5 |
|
KR |
X42H-156 |
-0.0727 |
405,409 |
|
KR |
X42H-122 |
-0.0047 |
22,27,33,39,41,43,44,48,62,94 |
|
KR |
X29H-164 |
0.0005 |
1,4,8,11,13,14,18,20,23,25,27,28,31,32,33,34,35,39,42,43,44, 46,49,53,55,71,81,95,99,322,324,328,329,334,339,345,350,354, 358,366,367,368,375,401,404,406,407,408,443,457,459,502,503, 504,505,506,507,508,509,510,511,513,516,517,518 |
|
KR |
X29H-168 |
0.1450 |
6,10 |
Prediction for novel parental combination
To develop clones with higher economic value, we simulated crossings for molecular breeding. We assumed that clones would have higher economic value with the inclusion of morefavorablealleles. Here, we considered synergistic alleles to be favorable for P. sibirica breeding programs. Based on the number of favorable alleles and phenotypic effects, we selected seven typical materials as receptor parents to conduct molecular breeding. We then proposed seven optimal cross combinations for improvingtheeconomic value of the fruit, nucleus, and kernel (Table 4). Table 5 lists all favorable alleles carried by the parents inthese crosses.
Table 4 Cross combinations with excellent alleles for economic traits
|
Traits |
Parental combinations |
No. of favorable alleles predicted |
|
Big fruit and High yield |
94*95 |
278 |
|
Excellent nuclear |
45*43 |
225 |
|
45*5 |
221 |
|
|
45*70 |
225 |
|
|
Excellent kernel |
70*15 |
209 |
|
15*94 |
231 |
|
|
70*94 |
258 |
Table 5 Information of parents
|
Clone No. |
Largest phenotypic effect carried |
No. of favorable alleles carried |
|
94 |
FL, FT, FLT, SFW, KR |
140 |
|
95 |
FY |
138 |
|
5 |
NSW |
114 |
|
43 |
NL, NW |
118 |
|
45 |
NST |
107 |
|
70 |
NR, KT, KSI |
118 |
|
15 |
SKW |
91 |
The fruit, nucleus, and kernel of P. sibirica have a high economic value and varied uses, necessitating the breeding of different seeds for distinct purposes. In this study, we examined more traits than prior research (Du et al. 2016; Edzesi et al. 2016), selecting 20 economic traits (seven fruit, seven nuclear, and six kernel traits). This increase in quantity correspondingly increases the power to evaluate the economic value of clones. We were also able to identify SSR markers associated with these traits, allowing for the use of a molecular breeding approach of marker-assisted selection (MAS) that significantly improves breeding efficiency over traditional methods of based on phenotypic selection.
The population structure is an important factor in association mapping, to avoid false marker trait associations obscuring the loci that truly affect target traits (Ca et al. 2017). Thus, regardless of population type, population structure analysis should be performed prior to association. A relatively complicated population structure occurs when allelic frequency varies within different sub-populations. This unbalanced allelic distribution causes false positive associations between genotypes and phenotypes, complicating association mapping (Cardon et al. 2003). However, based on simple population structure, false positives are minimized and association mapping is highly effective (Cardon et al. 2003). Here, our analysis in Structure clustered the test population of 113 P. sibirica clones into three subpopulations, with clones from the same geographic region having a high probability of grouping together. This relatively simple group structure was conducive to the subsequent correlation analysis.
Association mapping has many advantages in improving MAS efficiency. In particular, performing this analysis on a population with numerous genotypes and alleles decreases breeding time because parental crosses are no longer necessary (Xu et al. 2020). In this study, we used both GLM and MLM, the two most prevalent statistical models in association mapping. The MLM (Q + K) model considers both population structure and genetic relationships, reducing the risk of false positives but reducing true positives at the same time. As a result, complex models may miss some important loci (Xie et al. 2017). In contrast, the GLM(Q) model has high detection power and can find more associations (Colonges et al. 2021). Thus, different association mapping models should be used to make in-depth comparisons, as we have done in this study, to improve the accuracy of the analysis and reduce spurious false positives.
Here, we detected 117 marker-trait associations between 149 SSR and 20 economic traits. Of these associations, the highest phenotypic explanation rates reached 82.12% (GLM) and 84.82% (MLM), while the lowest were 5.17% (GLM) and 0.17% (MLM), respectively. These data show that the tested materials had rich genetic diversity. The larger the phenotypic explanation rate, the stronger the correlation between marker and trait. In this study, the average explained phenotypic variation of the 117 marker-trait associations was 27.93% (GLM) and 21.80% (MLM). These values were higher than those reported for upland cotton (9.5%) (Yuan et al. 2018), wheat (8.7%) (Quan et al. 2019), and poplar (21.4%) (Li et al. 2019).
Most of the biological traits are regulated by multiple loci. The trait with the most association loci (19) was NSI. Additionally, 35 markers were associated with multiple traits. Of these, six loci (X18, X20, X36, X27, X112, and X16H) were related to FL and FW, four (X18, X20, X36, and X69) were related to FL and SFW, and three (P2, X20, and X69) were related to SFW and SNW. These results agree with similar findings in mangoes (Shiv et al. 2017), grapes (Vafaee et al. 2017), and olives (Kaya et al. 2016), suggesting that the traits may exhibit linked inheritance. In terms of breeding, multiple traits can be improved simultaneously by these markers.
After we identified alleles associated with economic traits, we determined their phenotypic effects, and then identified how alleles in different germplasms contributed to the target trait. Through this process, we successfully found resources that carry multiple favorable alleles and bred hybrids with improved target traits by selecting those with the maximum number of favorable alleles. Breseghello and Sorrells et al. (Breseghello et al. 2006) first proposed the use of comparing allelic effects with null alleles to analyze phenotypic effects in 2006, a method that has been widely used to mine elite varieties of coconuts (Zhou et al. 2020), cotton (Wei et al. 2014), and rice (Du et al. 2016). Here, we used the same method to identify 1842 elite alleles of P. sibirica, including 1022 with synergistic effects and 820 with negative effects. Except for FY (range − 245.07–803.11), the phenotypic effect of other traits ranged from 10 to -10, similar to cotton (0.0918–2.1461) ((Wei et al. 2014)) and coconuts (-1.02 ~ 3.12) (Zhou et al. 2020). The phenotypic effect value of FY ranged from − 245.07 to 803.11. We speculated that value of FY was large for field measurement, after comparing this value with null allele, the result will also be large.
Peleman and van der Voort first proposed molecular breeding in 2003 (Peleman and van der Voort 2003), This method simulates and optimizes the breeding process before field experiments, thus increasing predictability over conventional breeding. If we want to increase a trait of interest, we should select alleles that have synergistic effects with the elite allele; if we want to reduce a trait, we use alleles with negative effects on the elite allele. For example, to develop stronger salt-tolerant accessions suitable for different saline-alkali soil, Lei et al. (Du et al. 2016) designed a simulated hybridization method to select 13 varieties carrying excellent alleles for molecular breeding. Likewise, Edzesi et al. (Edzesi et al. 2016) selected the five best rice varieties for reducing disease grade and proposed five optimal cross combinations. Our study identified clones that possess multiple desired elite alleles and considered them core germplasms for breeding. We then designed seven hybrid combinations to improve the economic value of P. sibirica.
Association mapping is an efficient tool for unravelling complex quantitative traits in plants and helps identify valuable alleles for breeding through MAS. This study identified 117 stable SSR markers, 1842 elite alleles, and seven typical materials. With these, we generated the seven optimal cross combinations for breeding economically valuable traits. Our findings provide empirical evidence of optimal crosses that can contribute to traditional parental breeding and MAS in the future.
Funding This work was supported by National Key R&D Program Project of China (Grant numbers SQ2019YFD100071)
Competing Interests The authors have no relevant financial or non-financial interests to disclose.
Author Contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Jianhua Chen, Yongqiang Sun, Quangang Liu, Ling Jin and Shengjun Dong. The first draft of the manuscript was written by Jiaxing Wen and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Data Availability The datasets generated during and analyzed during the current study are available from the corresponding author on reasonable request.
- Bhattarai G and Mehlenbacher SA (2018) Discovery, characterization, and linkage mapping of simple sequence repeat markers in hazelnut. Journal of the American Society for Horticultural Science 143:347-362
- Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES (2007) TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23:2633–2635
- Breseghello F, Sorrells ME (2006) Association mapping of kernel size and milling quality in wheat (Triticum aestivum L.) cultivars. Genetics 172:1165–1177
- Cai CP, Zhu GZ, Zhang TZ, Guo WZ (2017) High-density 80 K SNP array is a powerful tool for genotyping G. hirsutum accessions and genome analysis. BMC Genomics 18:654
- Cardon LR, Palmer LJ (2003) Population stratification and spurious allelic association. The Lancet 361:598-604
- Colonges K, Jimenez JC, Saltos A, Seguine E, Solorzano RGL, Fouet O, Argout X, Assemat S, Davrieux F, Cros E, Boulanger R, Lanaud C (2021) Two main biosynthesis pathways involved in the synthesis of the floral aroma of the Nacional cocoa variety. Front Plant Sci 12:681979
- Du L, Cai CP, Wu S, Zhang F, Hou S, Guo WZ (2016) Evaluation and Exploration of Favorable QTL Alleles for Salt Stress Related Traits in Cotton Cultivars (G. hirsutum L.). PLoS ONE 11: e0151076
- Duval MF, Noyer JL, Perrier X, Coppens d’Eeckenbrugge & Hamon P (2001) Molecular diversity in pineapple assessed by RFLP markers. Theor Appl Genet 102:83–90
- Edzesi WM, Dang XJ, Liang LJ, Liu EB, Zaid IU & Hong DL (2016) Genetic diversity and elite allele mining for grain traits in rice (Oryza sativa L.) by association mapping. Frontiers in Plant Science7:787
- Feng S, Tong H, Chen Y, Wang J, Chen Y, Sun G, He J and Wu Y (2013) Development of pineapple microsatellite markers and germplasm genetic diversity analysis. Biomed Res Int 2013:317912
- Ipek M, Ipek A, Seker M, Gul M (2015) Association of SSR markers with contents of fatty acids in olive oil and genetic diversity analysis of an olive core collection. Genetics and molecular research 14:2241-52
- Jha AB, Gali KK, Zhang HX, Purves R W, Tar’an B, Vandenberg A & Warkentin TD (2020) Folate profile diversity and associated SNPs using genome wide association study in pea. Euphytica 216: e60-782
- Kaya HB, Cetin O, Kaya HS, Sahin M, Sefer F & Tanyolac B, Genetics B (2016) Association Mapping in Turkish Olive Cultivars Revealed Significant Markers Related to Some Important Agronomic Traits. Biochemical Genetics 54:506-533
- Kim JY, Sa KJ, Ha YJ, Lee JK (2021) Genetic variation and association mapping in the F2 population of the Perilla crop (Perilla frutescens L.) using new developed Perilla SSR markers. Euphytica. https://doi.org/10.1007/s10681-021-02867-z
- Kim OG, Sa KJ, Lee JR & Lee JK (2017) Genetic analysis of maize germplasm in the Korean Genebank and association with agronomic traits and simple sequence repeat markers. Genes & Genomics 39:843–853
- Kozlowski P, de Mezer M, Krzyzosiak WJ (2010) Trinucleotide repeats in human genome and exome. Nucleic Acids Res 38:4027-39
- Lal S, Singh AK, Singh SK, Srivastav M, Singh BP, Sharma N & Singh NK (2017) Association analysis for pomological traits in mango (Mangifera indica L.) by genic-SSR markers. Trees 31:1391-1409
- Li D, Sun YH, Zhao KQ, Zhang J, Zhang YW, Li XY, Xun SH, Zhao JT, Wang SM & Li Y (2019) Development and application of EST-SSR markers for DNA fingerprinting and genetic diversity analysis of the main cultivars of black locust (Robinia pseudoacacia L.) in China. Forests 10:644-644
- Li XJ, Xu X, Liu WH, Li XQ, Yang XM, Ru ZG, Li LH (2020) Dissection of superior alleles for yield-related traits and their distribution in important cultivars of wheat by association mapping. Front Plant Sci 11:175
- Ma YX, Wang SX, Liu XJ, Yu HY, Yu D, Li GT, Wang LB (2020) Oil content, fatty acid composition and biodiesel properties among natural provenances of Siberian apricot (Prunus sibirica L.) from China. GCB Bioenergy 13:112-132
- Mai YT, Huo KS, Yu HY, Zhou N, Shui LY, Liu Y, Zhang CX, Niu J, Wang LB (2020) Using lipidomics to reveal details of lipid accumulation in developing Siberian apricot (Prunus sibirica L.) seed kernels. GCB Bioenergy 12:539-552
- Meng YS, Zhao N, Li Hui, Zhai H, He SZ, Liu QC (2018) SSR fingerprinting of 203 sweetpotato (Ipomoea batatas (L.) Lam.) varieties. Journal of Integrative Agriculture 17:86-93
- Niu J, Bi QX, Deng SY, Chen HP, Yu HY, Wang LB, Lin SZ (2018) Identification of AUXIN RESPONSE FACTOR gene family from Prunus sibirica and its expression analysis during mesocarp and kernel development. BMC Plant Biology. 18:21
- Paz EY, Gil K, Rebolledo L, Rebolledo A, Uriza D, Martínez O, Isidrón M and Simpson J (2005) AFLP characterization of the Mexican pineapple germplasm collection. Journal of the American Society for Horticultural Science 130:575–579
- Peleman JD, van der Voort JR (2003) Breeding by design. Trends Plant Sci 8:330–334
- Pritchard JK, Stephens M, Donnelly P (2000) Infernece of population structure using multilocus gentoype data. Genetics 155:945–959
- Quan MY, Du QZ, L, Lu WJ, Wang LX, Xie JB, Song YP, Xu BH, Zhang DQ (2019) Genetic architecture underlying the lignin biosynthesis pathway involves noncoding RNAs and transcription factors for growth and wood properties in Populus. Plant Biotechnology Journal 17:302-315
- Ravelombola W, Qin J, Shi A, Lu WG, Weng YJ, Xiong HZ, Yang W, Bhattarai G, Mahamane S, Payne WA, Miller Jr. JC & Scheuring D (2017) Association mapping revealed SNP markers for adaptation to low phosphorus conditions and rock phosphate response in USDA cowpea (Vigna unguiculata (L.) Walp.) germplasm. Euphytica 213:183
- Ruas CF, Ruas PM & Cabral JRS (2001) Assessment of genetic relatedness of the genera Ananas and Pseudananas confirmed by RAPD markers. Euphytica 119:245–252
- Samarina LS, Matskiv AO, Shkhalakhova RM, Koninskaya NG, Hanke MV, Flachowsky H, Shumeev AN, Manakhova KA, Malyukova LS, Liu SR, Zhu JY, Gvasaliya MV, Malyarovskaya VI, Ryndin AV, Pchikhachev EK and Reim S (2022) Genetic Diversity and Genome Size Variability in the Russian Genebank Collection of Tea Plant [Camellia sinensis (L). O. Kuntze]. Frontiers in Plant Science. https://doi.org/10.3389/fpls.2021.800141
- Thornsberry JM, Goodman MM, Doebley J, Kresovich S, Nielsen D, Buckler ES (2001) Dwarf8 polymorphisms associate with variation in flowering time. Nature Genetics 28:286–289
- Vafaee Y, Ghaderi N, Khadivi A (2017) Morphological variation and marker-fruit trait associations in a collection of grape (Vitis vinifera L.). Scientia Horticulturae 225:771-782
- Wei ZZ (2021) Evaluation of genetic resources and identification of SSRs associated with important traits in Populus simonii. Dissertation, Beijing Forestry University
- Xia W, Xiao Y, Liu Z, Luo Y, Mason AS, Fan HK, Yang YD, Zhao SL & Peng M (2014) Development of gene-based simple sequence repeat markers for association analysis in Cocos nucifera. Molecular Breeding 34:525-535
- Xie D, Dai Z, Yang Z, Sun J, Zhao D, Yang X, Zhang L, Tang Q, Su J (2017) Genome-Wide Association Study Identifying Candidate Genes Influencing Important Agronomic Traits of Flax (Linum usitatissimum L.) Using SLAF-seq. Front Plant Sci 8:2232
- Xu L, Jiang S, Wu J, Zou Q (2020) An in silico approach to identification, categorization and prediction of nucleic acid binding proteins. Briefings in Bioinformatics. https://doi, https://doi.org/10.1093/bib/bbaa171
- Yin MY, Wuyun T, Jiang ZM & Zeng J (2019) Amino acid profiles and protein quality of Siberian apricot (Prunus sibirica L.) kernels from Inner Mongolia. Journal of Forestry Research 31:1391–1397
- Yuan YC, Wang XL, Wang LY, Xing HX, Wang QK, Saeed K, Tao JC, Feng W, Zhang GH, Song XL, Sun XZ (2018) Genome-Wide Association Study Identifies Candidate Genes Related to Seed Oil Composition and Protein Content in Gossypium hirsutum L. Front Plant Sci 9:1359
- Zhang SL (2015) Variation regularity analyzation of material properties and association SSR locus mining for Cunninghamia lanceolata. Dissertation, Zhejiang Agriculture and Forestry University
- Zhang Y, Wu YN, Wang W, Yao GD, Jiang XF, Song SJ (2018) Flavones from a natural tea (the leaves of Armeniaca sibirica L.) prevent oxidative stress‐induced neuronal death. Journal of Food Biochemistry 42: e12657.
- Zhou LX, Yarra RJ, Cao HX (2020) SSR based association mapping analysis for fatty acid content in coconut flesh and exploration of the elite alleles in Cocos nucifera L. Current Plant Biology 21:2214-6628
- Zondervan KT, Cardon LR (2004) The complex interplay among factors that influence allelic association. Nature Reviews Genetics 5:89-100