Comparison of Statistical Models for Regional Crop Trial Analysis

Based on the review and comparison of main statistical analysis models for estimating variety-environment cell means in regional crop trials, a new statistical model, LR-PCA composite model was proposed, and the predictive precision of these models were compared by cross validation of an example data. Results showed that the order of model precision was LR-PCA model > AMMI model > PCA model > Treatment Means (TM) model > Linear Regression (LR) model > Additive Main Effects ANOVA model. The precision gain factor of LR-PCA model was 1.55, increasing by 8.4% compared with AMMI.

Cultivar Selection and Test Site Evaluation of Cotton Regional Trials in Jiangsu Province Based on GGE Biplot

This study was to evaluate the high yielding and stability of candidate cultivars, depict the adaptive planting region, analyze trial location discrim-ination ability and representativeness, as wel as identify the ideal cultivar and trial location, with the aim to provide theory background for cultivar selection and rea-sonable scheme of test location in Jiangsu Province. [Method] The GGE biplot method was used to analyze the lint cotton yield of 12 experimental genotypes in the 6 test locations （three replicates in each） of the cotton regional trial in Jiangsu Province in 2013. [Result] The effects of genotype （G）, environment （E）, and geno-type by environment interaction （G＆#215;E） on lint cotton yield were al highly significant （P〈0.01）, which made it necessary to further explore the specific pattern of geno-type by environment interaction. Jinmian118 （G4） and SF3303 （G5） were the best ideal genotypes screened by the ＂ideal cultivar＂ and ＂ideal location＂ view of GGE biplot, and the ordination of test sites based on the ideal index were in the order of Dafeng （DF）, Yanliang （YL）, Liuhe （LH）, Dongtai （DT）, Yancheng （YC）, and Nantong （NT）, among which NT was relatively weak in representing of the whole target cot-ton planting region in Jiangsu Province. The ＂similarity among locations＂ view of GGE biplot clustered al trial locations into one group, showing that the test sites in the cotton planting region in Jiangsu Province were in the same mega-environment. The ＂which-won-where＂ view of GGE biplot indicated that cotton cultivar Jinmian118 （G4） was the most appropriate cultivar in the homogeneous cotton planting region in Jiangsu Province. [Conclusion] Among the candidate cultivars, Jinmian118 and SF3303 were identified as the most ideal cultivars in this set of conventional cotton regional trial in Jiangsu Province; the test site of Dafeng ranked the first out of al locations in terms of discrimination and representativeness, and al test locations were clustered into the same mega-environmet, which indicated the high efficiency of cultivar selection in the cotton regional trial in Jiangsu Province.

Experimental Precision and Variety Comparison Precision in Regional Crop Trials

According to the basic concepts of precision and the principles of analysis of variance (ANOVA), precision types for experiments and variety comparison in regional crop trials (RCT) were studied and developed; expected variety comparison precision (EVCP) and realized variety comparison precision (RVCP) and the corresponding statistical indexes of them were proposed. It was explained that experimental precision (EP) and variety comparison precision (VCP) are two kinds of precision of RCT; EP includes error precision and variety mean precision, which can be measured respectively by the coefficient of variation of single observation's error (CVe) and the coefficient of variation of variety mean's error (CVy); VCP includes EVCP and RVCP, which can be measured respectively by the detectable least relative difference (DLRD) and the relative least significant distance (RLSD); EP is an important factor of VCP but not identical to it; RVCP is the realization of EVCP. Besides error, experimental design and GE interaction and ANOVA model affect VCP. Several application examples for these precision indexes were presented, and the precision of regional cotton trials in the Yellow River Valley and the Changjiang Valley were investigated through the historical data of RCTs from 1980 to 1999.

Evaluation of yield,stability and adaptability of national winter rapeseed regional trials in the upper Yangtze River region in 2017-2018

Rapeseed variety needs to be tested by regional trial in multiple sites for many years before being applied in market in China.Performants of rapeseed were affected by the interaction of sites and varieties.Evaluation of regional trials is very important for guiding rapeseed breeding.GGE(genotype main effects and genotypeenvironment interaction)biplot was used to evaluate yield,stability,adaptability,representativeness and discrimination of national winter rapeseed trials in the upper reaches of Yangtze River in 2017-2018.Results showed that the main effects of genotype(G),environment(E)and genotypeenvironment interaction(GE)were significant(P<0.01)for yield.Among them,E accounted for 46.95%total variation on average,G and GE accounted for 19.34%and 33.71%respectively.Eight varieties were found with high-yield,excellent stability and adaptability,including‘Yiyou 29’,‘Xiwang 920’,‘Liyouza 108’,‘Nanyou 546’,‘Dadi 195’,‘Jiayou 1’,‘Huayouza 28’and‘Yuhua 2’.Test sites included Nanchong,Mianyang,Wanzhou,Shuangliu and Chengdu in Sichuan Province and Zunyi together with Guiyang in Guizhou Province were selected for their excellent representativeness and discrimination.These results would provide theoretical basis for rapeseed breeding.

Evaluation on the High Yielding, Stability and Comprehensive Performances of Check Cultivars in National Cotton Regional Trials in the Yangtze River Valley

[Objective] This study was to analyze the high yielding, stability, adaptabil-ity and main characteristics of check cultivars in national cotton regional trials in the Yangtze River Val ey in 2000-2013, with the aim to provide theoretical basis for the reasonable selection of check cultivars for cotton regional trials. [Method] GGE biplot analysis method was used to carry out the intensive analysis and comprehensive comparisons on the performances and differences of the 9 major characters of check cultivars, namely, lint cotton yield , stability, and other main traits including bol weight, bol number, lint percentage, fiber length, fiber strength, micronaire val-ue, plant height and seed index. [Result] Ezamian 10 was suitable in the majority of cotton planting regions in the Yangtze River Val ey, with the best high yielding abili-ty and perfect yield stability. The yielding ability of Xiangzamian 8 ranked the sec-ond, but its stability was the poorest. Xiangzamian 8 was of the specific adaptive cultivar with relatively good performance in high yielding but the poorest stability. The suitable planting areas of Xiangzamian 8 were limited to the cotton planting re-gion around Dongting Lake in Hunan province, Poyang Lake in Jiangxi Province, the Coastal region in Jiangsu Province and Wuhan district in Hubei Province. Xi-angzamian 2 performanced poorer in both yielding and stability, while Simian 3 ranked first in stability but the last in yielding ability. The dominant character of Xi-angzamian 8 and Xiangzamian 2 were high cotton plants and big bol s. Simian 3 performanced super in lint percentage and fiber length. Ezamian 10 was strong in cotton bol formation ability, lint cotton yielding, fiber strength and seed size, as wel as high micronaire value. [Conclusion] The outstanding performance of Ezamian 10 in high and stable yield in the Yangtze River Val ey in recent years has effectively promoted the high yielding ability of candidate varieties in cotton regional trials in the area.

Comparison of Several Statistical Analysis Models for Genotypic Stability of Saccharum officinarum

[Objective] The study aimed to compare several statistical analysis models for estimating the sugarcane （Saccharum spp.） genotypic stability. [Method] The data of sugarcane regional trials in Guangdong, in 2009 was analyzed by three models respectively： Finlay and Wilkinson model： the additive main effects and multiplicative interaction （AMMI） model and linear regression-principal components analysis （LR- PCA） model, so as to compare the models. [Result] The Finlay and Wilkinson model was easier, but the analysis of the other two models was more comprehensive, and there was a bit difference between the additive main effects and multiplicative inter- action （AMMI） model and linear regression-principal components analysis （LR-PCA） model. [Conclusion] In practice, while the proper statistical method was usually con- sidered according to the different data, it should be also considered that the same data should be analyzed with different statistical methods in order to get a more reasonable result by comparison.

Review and Prospects on Rice Breeding and Extension in China

A large number of new varieties, especially hybrid rice, have been developed and released in China since 1998. The yield potential and grain quality of hybrid rice have also been improved significantly both for three-line and two-line hybrid rice including indica and japonica types as well. The trends on extension of rice varieties in China since 1986 has been marked by the increasing number of new varieties, decreasing annual planting area per variety, decreasing number of varieties with a large-scale extension area, while both the renewal rate and the proportion of core varieties with cross-provincial extension have remained unchanged. This current major problems and future goals of rice breeding and extension in China were also discussed.

Development and Validation of National Cotton Cultivar Registration Index Model in Yangtze River Valley

[Objective] Based on the cotton variety high yielding potential, fiber quality traits, disease resistance, and early maturity characters, a cultivar registration index model was developed to simplify the tedious calculation process in national cotton registration procedure, and thus to enhance the practical application of cultivar regis- tration index in cotton breeding ancl cotton recommending. [Method] By means of correlation analysis, partial correlation analysis and path analysis methods, the cor- relation of cotton main properties and their effects on cultivar registration index were explored using the dataset of national cotton regional trials in Yangtze River Valley during 1996-2013. The cultivar registration index model was constructed with step- wise regression statistical technique to ascertain the quantitative relationship of main characters with cultivar registration index, and the regional cotton trial dataset in 2013 was used to validate the model. [Result] Several characters with larger deter- minants to cultivar registration index were screened out,ie. lint yield increase ratio, pro-frost yield ratio, verticillium wilt index, fiber strength, fusarium wilt index and mi- cronaire value. The cultivar registration index model defined the functional relation- ship of cultivar registration index with the selected main characters, among which lint yield increase ratio, fiber strength and micronaire value contributed most to culti- var registration index. The model validation with regional cotton trials in 2013 indi- cated the root mean square error, RMSE was only 2.77, and the variation coeffi- cient was 6.77%, which confirmed the model prediction effect was quite perfect. [Conclusion] The developed cultivar registration index model was reliable enough to simulate the complicated scoring system in cultivar registration procedure, also sim- plified cotton registration process, and enhanced the practicability of the cultivar reg- istration index.