Многие экологические и биологические факторы могут быть эффективными при оценке генетических параметров в исследованиях потомства. Количество блоков (также называемых репликами) является одним из наиболее важных факторов окружающей среды в этих оценках. Влияние блочных сценариев на генетические параметры было исследовано на основе высоты и диаметра на высоте груди в 23-летнего потомства, созданного с помощью 4 блоков семенного стенда и саженцев турецкой красной сосны (Pinus Brutia Ten.) . Средние значения составили 14,14 м для высоты дерева и 17,69 см для диаметра на высоте груди в исследованных образцах. В результате дисперсионного анализа были обнаружены общезначимые (р<0,05) различия между блоками и отличительными родственными признаками. Хотя дисперсия ошибок была одинаковой для обоих признаков, были отмечены большие различия между другими компонентами сценариев. Индивидуальная наследуемость, фенотипический коэффициент вариации, генетический коэффициент вариации и генетический прирост варьировались для отличительных признаков и сценариев блоков. Результаты исследования подчеркивают влияние количества блоков на генетические параметры
разведение, прирост, генетика, рост, наследуемость, изменчивость
Introduction
Turkish red pine (Pinus brutia Ten.) is classified as one of the most economically important tree species for Turkish forestry and the “National Tree Breeding and Seed Production Programme” [1] because of its valuable commercial wood product and the largest natural distribution by 5.2 million ha which of %26 of total forest area of Turkey [2]. One of the basic aims of tree breeding programme is to estimate genetic parameters such as heritability and genetic gain based on variance components. Progeny trial is the main tool for the estimations, and an important part of the breeding programme. For these purposes, progeny trials were established in the species based on the programme. However, there could be many environmental (e.g., number of blocks, establishment design, aspect, spacing, homogeneity of the site) and biological (e.g., numbers of families and seed sources, distribution of the species) effect on establishment of the trial. Number of block (also called as replicate) is one of the main environmental factors, which can be changed/balanced by establisher. It plays important roles in the estimations from decision on model to statistical methods [3-12], and also, establishment cost and organization of trials.
In this study, some genetic parameters were estimated for tree height, and diameter at breast height based on different block scenarios in a 23-year-old progeny trial of Turkish red pine to contribute establishment and management of the trials.
Material and methods
The progeny trial was established by one-year containerized seedlings grown 27 seed trees selected phenotypic from each seed stand (SS-1… SS-5), and their clonal seed orchards (SO-1…SO-5) (Table 1) consist of 31 (SO-1), 25 (SO-2), 25 (SO-3), 28 (SO-4) and 20 (SO-5) clones selected from the stands.
The seedlings of the seed sources were planted by 3x2 m spacing by four blocks (also replicate) at experiment site from southern part of Turkey (36°57'30'' N latitude and 30°36'39'' E longitude, 292 m altitude) in 1997. Tree height (H), and diameter at breast height (D) were measured by 6 trees from each family at each block at 23rd year at end of growth period of 2020.
Data analysis
The data was analyzed at five scenarios based on purpose of the study. They were scenario 0 (S0, standard model by four blocks), scenario 1 (S1, by 3 blocks except of first block), scenario 2 (S2, by 3 blocks except of second block), scenario 3 (S3, by 3 blocks except of third block), and scenario 4 (S4, by 3 blocks except of fourth block).
The families were compared by following model of multiple analyses of variance (MANOVA) by [13].
(1)
Where Yikl is the observation from the lth tree of ith block/replicate of jth family, μ is overall mean, Bi is the effect of the ith block (I=1…4), Fk is the effect of kth family (k= 1, 2….134), BFik is the effect of interaction between block and family, eikl is random error.
Individual (narrow sense) heritability (hi2) was estimated as [14]:
(2)
Where is the additive genetic variance,
is the genetic variance of families,
is the phenotypic variance, k is the covariance coefficient between half-sibs.
Table 1. Geographic details of the seed sources
|
Seed Sources |
Latitude (N) |
Longitude (E) |
Altitude (m) |
|
SS-1 |
36°45'00'' |
31°57'55'' |
650 |
|
SS-2 |
36°35'30'' |
30°28'00'' |
320 |
|
SS-3 |
36°44'34'' |
29°28'36'' |
600 |
|
SS-4 |
37°06'20'' |
29°07'30'' |
800 |
|
SS-5 |
37°04'30'' |
29°32'40'' |
1100 |
|
SO-1 (SS-1)* |
37°00'40'' |
30°50'19'' |
100 |
|
SO-2 (SS-2) |
36°58'00'' |
30°40'42'' |
257 |
|
SO-3 (SS-3) |
36°40'04'' |
29°10'39'' |
240 |
|
SO-4 (SS-4) |
37°01'13'' |
30°43'35'' |
275 |
|
SO-5 (SS-5) |
36°58'11'' |
30°32' 42'' |
260 |
*; Mother population of the orchard.
Standard error of the heritability (S.E. (hi2)) was estimated as [15]:
(3)
Coefficients of variations of phenotypic (CVf) and genetic (CVg) were estimated based on phenotypic ( ), and genetic (
) variances and average (
) of the traits as [16]:
,
(4a&b)
Phenotypic Pearson’ correlation (rp) between tree height and diameter at breast height were estimated by [17].
(5)
Where is the sum of the factors of the traits x and y,
and
are phenotypic variances of the traits x and y.
Genetic correlation (rg) between tree height and diameter at breast height were estimated as [18]:
(6)
Where is the genetic covariance between traits x and y,
and
are the additive genetic variances of traits x and y, respectively.
Standard error of genetic correlation ( ) was estimated by [18]:
(7)
Where and
are standard error of heritability of traits x and y,
and
are heritability of traits x and y, respectively.
Genetic gain ( %) was estimated as:
(8)
Where is the average of breeding value estimated by BLUP (Best Linear Unbiased Prediction) of the highest 30 families,
is the breeding value of the control material, and
is the absolute breeding value of the control material.
Results and Discussion
Averages of tree height (H), and diameter at breast height (D) were 14.14 m and 17.69 cm in polled seed sources, respectively (Table 2). Annual increments were 0.62 m for H and 0.77 cm for H. They were 0.29 m and 0.56 cm in 32-year-old provenance trial of the species [12]. Individuals of seed orchards showed 9% higher growth performance for volume in 14th year result of progeny trial of the species [19].
Table 2. Averages (x̄) and standard deviation (S) of the traits for the seed sources
|
Seed sources |
H (m) |
D (cm) |
||
|
|
x̄ |
S |
x̄ |
S |
|
SS |
14.07 |
16.50 |
17.02 |
3.86 |
|
SO |
14.15 |
18.06 |
17.73 |
4.20 |
|
Total |
14.14 |
18.02 |
17.69 |
4.19 |
Positive and significant (p<0.05) phenotypic (r=0.76) and genetic (r=0.85±0.03) relations were found between height and diameter at breast height. Similar relation was also reported in provenance and progeny trials of Turkish red pine [12].
Significant (p<0.05) differences among blocks, and families for the traits in the scenarios were found based on results of analysis of variance except of among blocks (p>0.05) in scenarios 1 and 2. Block x family interactions was also significant in the scenarios (Table 3). Significant (p<0.05) differences among families were also reported for height and diameter in the species in early results of progeny trials of Turkish red pine [20-22]. While homogenous experimental area was expected in progeny trial, it could not be under control because of large number of genotype [23]. It was suggested heterogeneity of the site could be balanced by number of blocks [4].
The variance of error was similar for both traits, while large differences were estimated among other variance components for the scenarios (Table 4). The differences were about three times for the variance among blocks in S2 (1.21 %) and S3 (3.93 %), and the variance among families between S2 (3.97 %) and S1 (10.28 %) for height (Table 4). Variances among blocks were lower than 1% for diameter at breast height (Table 4). Variance of block x family interaction decreased from 19.80% (S0, four blocks) to 14.29% (S1, three blocks except of block 1) for height (Table 4). It showed heterogeneity of first block. Bian et al. [24] reduced 13% and 19% of error variance by missing plot for height and diameter in progeny trial of China fir (Cunninghamia lanceolata (Lamb.) Hook.). The results emphasized importance of number of blocks, traits and selection of experimental area for accurate estimations.
Large differences were estimated for individual heritability (hi2), phenotypic coefficient of variations (CVf), genetic coefficient of variations (CVg), and genetic gain ( %) for the traits and scenarios (Table 5). For instance, individual heritability ranged from 0.16 (S2) to 0.42 (S1) for height, and also genetic gain increased from 1.52 (S2) to 5.38 (S1) for the trait (Table 5). It was reported that in a progeny trial of Turkish red pine, the plot error was eliminated, that was, it was reduced to zero, for tree height in interlocked model and for diameter at breast height in twelve blocks model. The accuracy of predicted breeding values in height and diameter were increased significantly [12]. Similar results were also found that genetic parameters could change in different forest tree species based on analysis procedure [5, 7, 8, 24, 25]. For instance, heritability increased 35% and 51% for height and diameter in progeny trial of China fir [24] by different number of plots. Individual heritability, phenotypic and genetic coefficient of variations and genetic gain were generally higher in diameter at breast height than height in scenarios (Table 5). The results emphasized importance of number of blocks, traits and selection of experimental area for accurate estimations. Besides, there could be many environmental and biological factors together with block number. It was reported than heritability of growth traits changed for ages in the species [ i.e., 19, 26].
Table 3. Results of analysis of variance
|
|
Source of |
Scenarios |
||||
|
Traits |
variation |
S0 |
S1 |
S2 |
S3 |
S4 |
|
Height |
Block (B) |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
|
|
Family (F) |
p <0.001 |
p <0.001 |
p <0.05 |
p <0.05 |
p <0.05 |
|
|
FxP |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
|
Diameter |
Block (B) |
p <0.05 |
p>0.05 |
p>0.05 |
p <0.05 |
p <0.05 |
|
|
Family (F) |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
p <0.001 |
|
|
FxP |
p <0.001 |
p <0.05 |
p <0.001 |
p <0.05 |
p <0.05 |
Table 4. Variance components of the traits
|
Variance |
Scenarios |
||||
|
components* |
S0 |
S1 |
S2 |
S3 |
S4 |
|
Height |
|||||
|
|
868.2 (2.64 %) |
438.5 (1.40 %) |
368.2 (1.21 %) |
1336.6 (3.93 %) |
1332.7 (3.75 %) |
|
|
1911.4 (5.82 %) |
3224.1 (10.28 %) |
1208.6 (3.97 %) |
1416.4 (4.16 %) |
1795.1 (5.05 %) |
|
|
6505.1 (19.80 %) |
4483.1 (14.29 %) |
7787.8 (25.58 %) |
7118.8 (20.93 %) |
6669.8 (18.75 %) |
|
|
23561.8 (71.73 %) |
23227.8 (74.04 %) |
21075.9 (69.24 %) |
24144.1 (70.98 %) |
25772.0 (72.46 %) |
|
Diameter |
|||||
|
|
5.57 (0.32 %) |
2.31 (0.13 %) |
2.46 (0.15 %) |
9.37 (0.52 %) |
8.79 (0.47 %) |
|
|
109.98 (6.26 %) |
134.52 (7.69 %) |
96.40 (5.93 %) |
98.81 (5.50 %) |
113.07 (6.09 %) |
|
|
72.54 (4.13 %) |
60.23 (3.44 %) |
98.54 (6.06 %) |
69.89 (3.89 %) |
62.32 (3.36 %) |
|
|
1568.70 (89.29 %) |
1551.80 (88.73 %) |
1428.40 (87.86 %) |
1618.80 (90.09 %) |
1673.20 (90.08 %) |
*;
is the variance among blocks, 
is the variance among families, 
is the variances of block x family interaction, is the variance of error.
Table 5. Some parameters for the traits
|
|
Scenarios |
||||
|
Parameters |
S0 |
S1 |
S2 |
S3 |
S4 |
|
Height |
|||||
|
|
0.24±0.08 |
0.42±0.10 |
0.16±0.09 |
0.17±0.09 |
0.21±0.09 |
|
CVg |
6.18 |
7.96 |
4.96 |
5.31 |
5.99 |
|
CVf |
12.64 |
12.32 |
12.37 |
12.76 |
13.08 |
|
|
3.16 |
5.38 |
1.52 |
2.57 |
2.58 |
|
Diameter |
|||||
|
|
0.25±0.06 |
0.31±0.07 |
0.24±0.07 |
0.22±0.07 |
0.24±0.07 |
|
CVg |
11.85 |
13.02 |
11.17 |
11.21 |
12.04 |
|
CVf |
23.65 |
23.46 |
22.91 |
23.84 |
24.35 |
|
|
7.66 |
9.16 |
5.29 |
8.08 |
6.13 |
Conclusions
The results of the study showed importance of number of blocks in estimations. However, number of block is a changeable environmental factor by trial managers. Homogenous test area should be selected in establishment stage of trials.Future studies should be carried out for accurate conclusions on the trials based on rotation age of the species. Results of the study could be combined with other progeny trials in the species.
Blocks and their number should be checked during data collection and estimations.
1. Koski V, Antola J. National tree breeding and seed production programme for Turkey 1994-2003. Ankara, 1993; 52 p. www.ortohum.gov.tr
2. Anonymous . Forest inventory of Turkey. General Directorate of Forestry of Turkey, Ankara, 2021; https://www.ogm.gov.tr (acess: 07.07.2021) (in Turkish).
3. Brownie C., Bowman D.T., Burton J. W. Estimating spatial variation in analysis of data from yield trials: a comparison of methods. Agronomy Journal. 1993; (85):1244-1253. DOIhttps://doi.org/10.2134/agronj1993.00021962008500060028x
4. Neter J., Kutner M., Wasserman W., Nachtsheim C. Applied linear statistical models (4th ed.). New York: 1996. 1408 p. ISBN 978-0256117363.
5. Costa e Silva J., Dutkowski G. W., Gilmour A. R. Analysis of early tree height in forest genetic trials is enhanced by including a spatially correlated residual. Canadian Journal of Forest Research. 2001; (31):1887-1893. https://doi.org/10.1139/x01-123
6. Cargnelutti Filho A., Storck L., Dal'Col Lúcio A. Ajustes de quadrado médio do erro em ensaios de competição de cultivares de milho pelo método de Papadakis. Pesquisa Agropecuária Brasileira. 2003; (38):467-473. https://doi.org/10.1590/S0100-204X2003000400004
7. Zas R. Iterative kriging for removing spatial autocorrelation in analysis of forest genetic trials. Tree genetics & genomes. 2006; (2):177-185. https://doi.org/10.1007/s11295-006-0042-4
8. Gezan S.A., Huber D.A., White, T.L. Post hoc blocking to improve heritability and precision of best linear unbiased genetic predictions. Canadian journal of forest research. 2006; (36):2141-2147. https://doi.org/10.1139/x06-112
9. White T.L., Adams W.T., Neale D.B. Forest Genetics. Cabi. Athens. 2007: 702p. ISBN 9789608002524
10. Welham S.J., Gezan S.A., Clark S.J., Mead A. Statistical methods in biology: design and analysis of experiments and regression. CRC Press, Boca Raton. 2014; 24-27. ISBN 9780429112980
11. Silva F.H., Muñoz P.R., Vincent C.I., Viana A.P. Generating relevant information for breeding Passiflora edulis: genetic parameters and population structure. Euphytica. 2016; (208):609-619. https://doi.org/10.1007/s10681-015-1616-8
12. Ozbey AA. Investigation on plot error in provenance and progeny trials of Turkish red pine. Graduate School, Isparta Universityof Applied Sciences, PhD. Thesis, Isparta. 2022. 137 p. www.yok.gov.tr (in Turkish).
13. SAS. Statistical analysis system. SAS Institute, Inc. Cary., N.C.: 2004.
14. Zobel B., Talbert J. Applied forest tree improvement. John Wiley & Sons. 1984: 505 p. ISBN 9780471096825
15. Isik F., Holland J., Maltecca C. Genetic data analysis for plant and animal breeding (Vol. 400). Cham, Switzerland: 2017.
16. Sun O. Statistical methods. Forest Research Institute, Ankara. 1980: 376 p. https://www.ogm.gov.tr (in Turkish).
17. Rohlf F.J., Sokal R.R. Statistical tables. Macmillan: 1995. 28 p. ISBN 9780716706649
18. Falconer D.S. Introduction to quantitative genetics. Harlow Longman Sci Tech. 1898: 438 p. ISBN 0 470 21162 8.
19. Calikoglu M., Sabuncu, R., Alan, M., Cosgun S., Yolcu H.I. Genetic gain in seed orchards of Turkish red pine. South-Western Anatolia Forest Research Institute, Antalya. 2011. (41): 29p. www.baoram.ogm.gov.tr (in Turkish).
20. Ozturk H., Siklar S., Alan M., Ezen T., Korkmaz B., Gulbaba A.G., Sabuncu R., Derilgen S.I., Caliskan B. 8th year results of in progeny trial of Turkish red pine in Mediterranean region. The Research Directorate of Forest Tree Seeds and Tree Breeding, Ankara. 2008. (18):122 p. www.ortohum.gov.tr (in Turkish).
21. Ozbey A.A., Calikoglu M., Cobanoglu A., Erpulat M. (2020). Final report projects of selection age in progeny and progeny trial of Turkish red pine. The South-Western Anatolia Forest Research Institute, Antalya. 2020. www.baoram.ogm.gov.tr (in Turkish).
22. Ozyalcin, K. (2021). Estimation of breeding value of some wood characters for superior Turkish red pine (Pinus brutia Ten.) families in breeding zone (200-600 m) of Marmara region. Graduate School, Ankara Universityof Applied Sciences, PhD. Thesis, Ankara. 2021. 266 p. www.yok.gov.tr (in Turkish).
23. Xing L., Gezan S., Kenworthy K., Unruh J.B., Munoz P. Improved genetic parameter estimations in zoysiagrass by implementing post hoc blocking. Euphytica. 2017; (213):1-10. https://doi.org/10.1007/s10681-017-1984-3
24. Bian L., Zheng R., Su S., Lin H., Xiao H., Wu H.X., Shi J. Spatial analysis increases efficiency of progeny testing of Chinese fir. Journal of Forestry Research. 2017; (283):445-452. https://doi.org/10.1007/s11676-016-0341-z
25. Dutkowski G.W., Silva J.C.E., Gilmour A.R., Lopez G.A. Spatial analysis methods for forest genetic trials. Canadian Journal of Forest Research. 2002; (32):2201-2214. https://doi.org/10.1139/x02-111
26. Ozyalcin K., Ezen T., Sabuncu R., Polat O., Ozbey A.A., Calikoglu M., Unal C., Derilgen S.I., Acet M., Kaplan A. Final report projects of progeny trial in Turkish red pine:20th year results. The Research Directorate of Forest Tree Seeds and Tree Breeding, Ankara. 2021. www.ortohum.gov.tr (in Turkish).
