Conversion of the Statistical Combining Ability into a Genetic Concept

Since the combining ability was proposed in 1942, efforts to uncover the genetic basis underlying this phenomenon have been ongoing for nearly 70 yr, with little success. Some breeding strategies based on evaluation of combining ability have been produced, and are still extensively used in hybrid breeding. In this review, the genetic basis underlying these breeding strategies is discussed, and a potential genetic control of general combining ability （GCA） is postulated. We suggested that GCA and the yields of inbred lines might be genetically controlled by different sets of loci on the maize genome that are transmitted into offspring. Different inbred lines might possess different favorable alleles for GCA. In hybrids, loci involved in multiple pathways, which are directly or indirectly associated with yield performance, might be regulated by GCA loci. In addition, a case of GCA mapping using a set of testcross progeny from introgression lines is provided.

Analysis of combining ability for stem-related traits and its correlations with lodging resistance heterosis in hybrid wheat

With the application of hybrid wheat, lodging is becoming one of the major factors limiting high yield in its production. However,few studies have focused on combining ability and heterosis analysis of stem-related traits. In this study, 24 crosses were made according to NCII genetic design, using the three(photo-sensitive male sterile lines)×eight(restorer lines) incomplete diallel crosses. The length of basal second internode(LBSI) and breaking strength of basal second internode(BSBSI)as well as other stem-related traits were used to perform the principal component analysis(PCA), combining ability and heterosis analysis. The PCA results showed that the variables could be classified into two main factors, which were named as the positive factor(factor 1) and the negative factor(factor 2), and accounted for 52.3 and 33.2%, respectively, of the total variance in different variables, combined with the analysis for index weight indicated that the factor 1-related traits play positive roles in lodging resistance formation of hybrids. Combining ability variance analysis indicated that its genetic performance was mainly dominated by additive gene effects, and the hybrid combinations with higher lodging resistance can be selected by using of 14 GF6085(R1), 14 GF6343-2(R4), 14 GF6937(R6), 14 GF7433-1(R7), and BS1086(M3),which are with the features with lower general combining ability(GCA) effects of factor 2-related traits whereas higher GCA effects of factor 1-related traits. The heterosis analysis showed that the wide range of heterosis varied with the traits and combinations, and GCA or specific combining ability(SCA) effects of factor 1-related traits except wall thickness of basal second internode(WTBSI) were positively and closely related to the heterosis of lodging resistance. Generally, the correlation coefficients of heterosis to GCA effects of sterile lines(GCAm) of factor 1-related traits are significantly higher than that to GCA of restorer lines(GCAr) and SCA, combined with the higher GCAm variance values of factor 1-related traits compared to GCAr, the GCAm of factor 1-related traits should be particularly considered when breeding hybrid combinations.The heritability analysis showed that the narrow-sense heritability of the diameter of basal second internode(DBSI) and the center of gravity height(TCGH) were obviously lower(<60%) than other traits, suggesting that these two traits were suitable for selection in higher generation for parental breeding. These could provide a theoretical basis for parental breeding and heterosis utilization of lodging resistance.

QTL mapping of general combining abilities of four traits in maize using a high-density genetic map

General combining abilities （GCAs） are very important in utilization of heterosis in maize breeding. However, its genetic basis is unclear. In the present study, a set of 118 doubled haploid （DH） lines were induced from F1 generations produced from the cross between the inbred line Zheng 58 and the inbred line W499 belonging to the Reid subgroup. Using the MaizeSNP50 BeadChip, a high-density genetic map was constructed based on the DH population which included 1 147 bin markers with an average interval length of 2.00 cM. Meanwhile, the DH population was crossed with three testers including W16-5, HD568, and W556, which belong to the Sipingtou subgroup. The GCAs of the ear height （EH）, the kernel moisture content （KMC）, the kernel ratio （KR）, and the yield per plant （YPP） were estimated using these hybrids in three environments. Combining the high-density genetic map and the GCAs, a total of 14 QTLs were detected for the GCAs of the four traits. Especially, one pleiotropic QTL was identified on chromosome 1 between the SNP SYN16067 and the SNP PZE-101169244 which was simultaneously associated with the GCAs of the EH, the KR, and the YPP. These QTLs pave the way for further dissecting the genetic architecture underlying GCAs of the traits, and they may be used to enhance GCAs of inbred lines under the fixed heterotic pattern ReidxSipingtou in China through a marker-assisted selection approach.

Correlation Between Parents and F_1 and Combining Ability of Parents on Seed Dormancy in indica Rice (Oryza sativa)

Dormancy indices of hulled and dehulled seeds were investigated by using 19 cytoplasmic male sterile （CMS） lines, 9 restorer lines and their 109 F1 hybrids of indica hybrid rice. The seeds of each F1 and the parents were harvested on 35 days after flowering. Combining ability was analyzed in 25 combinations made by 5 CMS lines and 5 restorer lines （North Carolina II mating design）. The seed dormancy index of F1 was positively and highly significantly correlated with those of their parents and mid-parent value. Out of the 109 combinations, 82 combinations showed mid-parent heterosis, and 43 heterobeltiosis. Seed dormancy indices of F1s and their parents declined dramatically in dehulled seeds compared with hulled seeds, indicating that the hull played an important role in seed dormancy. However, the trends were similar in hulled seeds and dehulled seeds in terms of relationships between the seed dormancy indicices in F1 and their parents. The influence of hull on seed dormancy mainly depended on F1 genotype, not on the hull from maternal parent. The variances of general combining ability （GCA） in female and male parents occupied 59.2% and 31.1% of total variance, respectively. The variance of specific combining ability （SCA） in combinations occupied 9.7% of total variance, indicating that gene additive effects were principal. Among the 5 CMS lines, II112A had the highest GCA effect for seed dormancy, followed by D62A. Among the 5 restorer lines, IRl12 had the highest GCA effect for seed dormancy, followed by 2786. These lines are elite parental materials for breeding F1 hybrid rice with stronger seed dormancy.

Combining ability and heterosis analysis over two environments in a diallel cross of three families of the clam Meretrix meretrix

A 3×3 complete diallel cross comprising three families of the clam Meretrix meretrix（P1, P2 and P3） was used to determine the combining ability of parental families and heterosis of F1 under indoor and openair environments for growth traits. Analysis of variance for shell length and whole body weight indicated highly significant cross effects, environment effects and the interaction of cross by environment. General combining ability（GCA） and specific combing ability exhibited great variation among crosses and between two environments. Pooled over environments, P2 was the top combiner among the three parental families for both traits studied. The cross of P1 and P3 had the highest SCA. Additionally, significant reciprocal effects were observed. For individual environment, about half of the crossbred combinations showed favorable Mid-parent heterosis（MPH）（〉1%） for the shell length and whole body weight. Our data has shown that non-additive genetic and reciprocal effects constituted the major sources of genetic variation for both shell length and whole body weight, which indicates that crossbreeding among selective families could further explore the heterotic effects.

Study of Yield Combining Ability and Genetic Relationship Among Exotic Tropical,Subtropical Maize(Zea mays L.) Inbreds and Temperate Maize Inbreds in China

Information on the genetic relationship between tropical maize (Zea mays L), germplasm and temperate maize germplasm is of great value to maize breeding. The objective of this study was to determine the combining ability and genetic relationship of 25 inbreds extracted from five tropical maize populations and a land race, with four temperate maize inbreds (Huangzaosi, Mol7, B73 and Dan 340). The 25 tropical inbreds were crossed with the four temperate inbreds and evaluated. Lines from Suwanl and POP28 had high general combining ability (GCA) for grain yield. The lines from POP32 (ETO) had the highest special combining ability (SCA) with B73; the average SCA value of the 5 lines was 879 kg/ha. The lines from Suwanl had the second-highest SCA (584 kg/ha) with Huangzaosi. The lines from Suwanl had the greatest relative heterosis (20%) with B73, followed by the lines from POP32 (ETO) with B73 (19%). Five heterotic patterns have been identified from this study: Suwanl × Reid, ETO × Reid, POP28× Reid, POP28× Ludahong-gu, and Suwan1× Lancaster.

Analysis of Combining Ability for Resistance of Sugarcane Parents and Cross Combinations to Brown Stripe Disease

[Objectives] The paper was to screen resistant sugarcane varieties against brown stripe disease,and to breed disease-resistant germplasm resource.[Methods]The combining ability for resistance to sugarcane brown stripe disease was analyzed based on 23 female parents,21 male parents and 29 cross combinations. [Results]The average heritability of resistance to sugarcane brown stripe disease successively were female parents（ 95. 3%）,cross combinations（ 93. 0%）and male parents（ 79. 1%）. The general combining ability of 12 female parents showed negative effect,including Pma 98-40,Yacheng 93-26,Yunrui 05-283,Yuetang 91-976,Chuanzhe 19,ROC10,Yunzhe 06-80,ROC26,Zhanzhe 74-141,K86-110,Yunzhe 03-194 and ROC25. The general combining ability of 10 male parents showed negative effect,including Q 199,Yunrui 06-649,Yunrui 05-733,CP 84-1198,CP 88-1762,Yacheng 84-125,Yunrui 05-784,Yuetang 00-236,CP72-3591 and CP 94-110. The special combining ability of 16 cross combinations showed negative effect,including Pma 98-40 × Yunrui 05-649,Yacheng 93-26 ×Yunrui 05-733,Yunrui 05-283 × Q199,Yuetang 91-976 × CP 84-1198,Chuanzhe 19 × CP 88-1762 and ROC10 × Yuenong 73-204. [Conclusions] There were significant differences in combining ability among female parents,male parents and cross combinations,which were mainly controlled by additive and non-additive gene.

Yarn Quality Prediction for Small Samples Based on AdaBoost Algorithm

In order to solve the problems of weak prediction stability and generalization ability of a neural network algorithm model in the yarn quality prediction research for small samples,a prediction model based on an AdaBoost algorithm(AdaBoost model) was established.A prediction model based on a linear regression algorithm(LR model) and a prediction model based on a multi-layer perceptron neural network algorithm(MLP model) were established for comparison.The prediction experiments of the yarn evenness and the yarn strength were implemented.Determination coefficients and prediction errors were used to evaluate the prediction accuracy of these models,and the K-fold cross validation was used to evaluate the generalization ability of these models.In the prediction experiments,the determination coefficient of the yarn evenness prediction result of the AdaBoost model is 76% and 87% higher than that of the LR model and the MLP model,respectively.The determination coefficient of the yarn strength prediction result of the AdaBoost model is slightly higher than that of the other two models.Considering that the yarn evenness dataset has a weaker linear relationship with the cotton dataset than that of the yarn strength dataset in this paper,the AdaBoost model has the best adaptability for the nonlinear dataset among the three models.In addition,the AdaBoost model shows generally better results in the cross-validation experiments and the series of prediction experiments at eight different training set sample sizes.It is proved that the AdaBoost model not only has good prediction accuracy but also has good prediction stability and generalization ability for small samples.

SNWPM:A Siamese Network Based Wireless Positioning Model Resilient to Partial Base Stations Unavailable

Artificial intelligence(AI)models are promising to improve the accuracy of wireless positioning systems,particularly in indoor environments where unpredictable radio propagation channel is a great challenge.Although great efforts have been made to explore the effectiveness of different AI models,it is still an open problem whether these models,trained with the data collected from all base stations(BSs),could work when some BSs are unavailable.In this paper,we make the first effort to enhance the generalization ability of AI wireless positioning model to adapt to the scenario where only partial BSs work.Particularly,a Siamese Network based Wireless Positioning Model(SNWPM)is proposed to predict the location of mobile user equipment from channel state information(CSI)collected from 5G BSs.Furthermore,a Feature Aware Attention Module(FAAM)is introduced to reinforce the capability of feature extraction from CSI data.Experiments are conducted on the 2022 Wireless Communication AI Competition(WAIC)dataset.The proposed SNWPM achieves decimeter-level positioning accuracy even if the data of partial BSs are unavailable.Compared with other AI models,the proposed SNWPM can reduce the positioning error by nearly 50%to more than 60%while using less parameters and lower computation resources.

Customer Churn Prediction Framework of Inclusive Finance Based on Blockchain Smart Contract

In view of the fact that the prediction effect of influential financial customer churn in the Internet of Things environment is difficult to achieve the expectation,at the smart contract level of the blockchain,a customer churn prediction framework based on situational awareness and integrating customer attributes,the impact of project hotspots on customer interests,and customer satisfaction with the project has been built.This framework introduces the background factors in the financial customer environment,and further discusses the relationship between customers,the background of customers and the characteristics of pre-lost customers.The improved Singular Value Decomposition(SVD)algorithm and the time decay function are used to optimize the search and analysis of the characteristics of pre-lost customers,and the key index combination is screened to obtain the data of potential lost customers.The framework will change with time according to the customer’s interest,adding the time factor to the customer churn prediction,and improving the dimensionality reduction and prediction generalization ability in feature selection.Logistic regression,naive Bayes and decision tree are used to establish a prediction model in the experiment,and it is compared with the financial customer churn prediction framework under situational awareness.The prediction results of the framework are evaluated from four aspects:accuracy,accuracy,recall rate and F-measure.The experimental results show that the context-aware customer churn prediction framework can be effectively applied to predict customer churn trends,so as to obtain potential customer data with high churn probability,and then these data can be transmitted to the company’s customer service department in time,so as to improve customer churn rate and customer loyalty through accurate service.

An Optimized Random Forest Model and Its Generalization Ability in Landslide Susceptibility Mapping:Application in Two Areas of Three Gorges Reservoir,China

Numerous researches have been published on the application of landslide susceptibility assessment models;however,they were only applied in the same areas as the models were originated,the effect of applying the models to other areas than the origin of the models has not been explored.This study is purposed to develop an optimized random forest(RF)model with best ratios of positive-to-negative cells and 10-fold cross-validation for landslide susceptibility mapping(LSM),and then explore its generalization ability not only in the area where the model is originated but also in area other than the origin of the model.Two typical counties(Fengjie County and Wushan County)in the Three Gorges Reservoir area,China,which have the same terrain and geological conditions,were selected as an example.To begin with,landslide inventory was prepared based on field investigations,satellite images,and historical records,and 1522 landslides were then identified in Fengjie County.22 landslide-conditioning factors under the influence of topography,geology,environmental conditions,and human activities were prepared.Then,combined with 10-fold cross-validation,three typical ratios of positive-to-negative cells,i.e.,1:1,1:5,and 1:10,were adopted for comparative analyses.An optimized RF model(Fengjie-based model)with the best ratios of positive-to-negative cells and 10-fold cross-validation was constructed.Finally,the Fengjie-based model was applied to Fengjie County and Wushan County,and the confusion matrix and area under the receiver operating characteristic(ROC)curve value(AUC)were used to estimate the accuracy.The Fengjie-based model delivered high stability and predictive capability in Fengjie County,indicating a great generalization ability of the model to the area where the model is originated.The LSM in Wushan County generated by the Fengjie-based model had a reasonable reference value,indicating the Fengjiebased model had a great generalization ability in area other than the origin of the model.The Fengjiebased model in this study could be applied in other similar areas/countries with the same terrain and geological conditions,and a LSM may be generated without collecting landslide information for modeling,so as to reduce workload and improve efficiency in practice.

Machine learning-based classification of rock discontinuity trace:SMOTE oversampling integrated with GBT ensemble learning

This paper presents a hybrid ensemble classifier combined synthetic minority oversampling technique(SMOTE),random search(RS)hyper-parameters optimization algorithm and gradient boosting tree(GBT)to achieve efficient and accurate rock trace identification.A thirteen-dimensional database consisting of basic,vector,and discontinuity features is established from image samples.All data points are classified as either‘‘trace”or‘‘non-trace”to divide the ultimate results into candidate trace samples.It is found that the SMOTE technology can effectively improve classification performance by recommending an optimized imbalance ratio of 1:5 to 1:4.Then,sixteen classifiers generated from four basic machine learning(ML)models are applied for performance comparison.The results reveal that the proposed RS-SMOTE-GBT classifier outperforms the other fifteen hybrid ML algorithms for both trace and nontrace classifications.Finally,discussions on feature importance,generalization ability and classification error are conducted for the proposed classifier.The experimental results indicate that more critical features affecting the trace classification are primarily from the discontinuity features.Besides,cleaning up the sedimentary pumice and reducing the area of fractured rock contribute to improving the overall classification performance.The proposed method provides a new alternative approach for the identification of 3D rock trace.

Determination of Heterotic Groups and Heterosis Analysis of Yield Performance in indica Rice

To compare the heterosis levels among various groups of parental lines used extensively in China, identify foundational heterotic groups in parental pools and understand the relationship between genetic distance and heterosis performance, 16 parental lines with extensive genetic variation were selected from various sub-groups, and 39 hybrid combinations were generated and evaluated in Fujian and Hainan Provinces of China. The main results were as follows： （1） The 16 parental lines can be grouped into 7 sub-groups consisting of 1 maintainer sub-group and 6 restorer sub-groups; （2） Mean grain yield of the restorer lines was higher than that of the maintainer lines, and mean yield of parental lines was higher than that of the hybrid combinations; （3） The two best heterotic patterns were II-32A × G5 and II-32A × G6, moreover, the order of restorer sub-groups according to grain yield, from the highest to lowest, was G7, G6, G5, G4, G3 and G2; High specific combining ability values were observed for combinations of II-32A × G5, II-32A × G6 and Tianfeng A × G7; （4） Hybrid combinations derived from II-32A crossed with 13 restorer lines had higher yield trait values （mid-parent heterosis, better-parent heterosis, standard heterosis over check and specific combining ability） than any other combinations; （5） Genetic distance was positively correlated with panicle number, grain length and length-to-width ratio （P 〈 0.05） and negatively correlated with grain width, grain yield, seed-setting rate, as well as mid-parent heterosis, standard heterosis over check, and specific combining ability for grain yield （P 〈 0.01）. These heterotic groups and patterns and their argonomic traits will provide useful information for future hybrid rice breeding programs.

Genomic prediction of the performance of hybrids and the combining abilities for line by tester trials in maize

The two most important activities in maize breeding are the development of inbred lines with high values of general combining ability(GCA)and specific combining ability(SCA),and the identification of hybrids with high yield potentials.Genomic selection(GS)is a promising genomic tool to perform selection on the untested breeding material based on the genomic estimated breeding values estimated from the genomic prediction(GP).In this study,GP analyses were carried out to estimate the performance of hybrids,GCA,and SCA for grain yield(GY)in three maize line-by-tester trials,where all the material was phenotyped in 10 to 11 multiple-location trials and genotyped with a mid-density molecular marker platform.Results showed that the prediction abilities for the performance of hybrids ranged from 0.59 to0.81 across all trials in the model including the additive effect of lines and testers.In the model including both additive and non-additive effects,the prediction abilities for the performance of hybrids were improved and ranged from 0.64 to 0.86 across all trials.The prediction abilities of the GCA for GY were low,ranging between-0.14 and 0.13 across all trials in the model including only inbred lines;the prediction abilities of the GCA for GY were improved and ranged from 0.49 to 0.55 across all trials in the model including both inbred lines and testers,while the prediction abilities of the SCA for GY were negative across all trials.The prediction abilities for GY between testers varied from-0.66 to 0.82;the performance of hybrids between testers is difficult to predict.GS offers the opportunity to predict the performance of new hybrids and the GCA of new inbred lines based on the molecular marker information,the total breeding cost could be reduced dramatically by phenotyping fewer multiple-location trials.

Genetic variation in progenies of 23 relic trees of Metasequoia glyptostroboides in their original habitat

Seeds of 23 relic Metasequoia glyptostroboides trees were randomly collected from their original, natural opulations found in Lichuan City, Hubei Province, China. The seed characteristics and growth of their three-year-old rogenies were investigated. Results show that the average intra-family coefficient of variation （CV） in height growth of ：he 23 three-year-old families was 14.53%, slightly higher than that of the inter-family CV （14.02%）. Correspondingly, hese two parameters for root-collar diameter growth were 15.44% and 13.15%. The 23 relic mother trees were ignificantly differed in their general combining abilities （GCA） and relative effect values of GCA （RGCA） of height growth and root-collar diameter growth over the three years. Both height and root-collar diameter growth had high amily heritability in this seedling stage. After eliminating two inferior families, i.e., 15 and 46, 2.21% of genetic gain in he ght growth was obtained and 1.50% in root-collar diameter growth. It was also significantly different in crown size and number of branches among the 23 three-year-old seedling families. From a cluster analysis we concluded that eight families, i.e., 23, 28, 29, 30, 31, 33, 35 and 44 displayed superior GCA and RGCA in both height and root-collar diameter growth. Therefore, these relic mother trees displayed their potential as parents in seed orchards. With the exception of family 23, the other seven superior families were all collected at Guihua Village, Zhonglu Town, Lichuan City, which has been identified as an optimum district of M. glyptostroboides for seed collection.

Marker-assisted Selection of ZmC_4Ppc in Rice Breeding and Yield Trait Performances of Advanced Lines

The full-length of intact Zea mays gene for phosphoenolpyruvate carboxylase gene （ZmC4Ppc） is 6 781 bp. The products of PCR for this gene were not clear with poor repeatability, resulting in that it was difficult for marker-assisted selection （MAS） both in rice and maize. For selecting the markers for MAS, sequences presented only in maize rather than in rice were identified by BLAST, and used for primer design using Primer Premier 5.0. A pair of specific primer termed MRpc （Forward： 5＇ AAGCAGGGAAGCGAGACG 3＇, Reverse： 5＇ GATTGCCGCCAGCAGTAG 3＇） was used for selection of transformed rice, and ZmC4Ppc could be highly and constitutively expressed at each tested developmental stages in the transformed rice selected by using MRpc. Thus, MRpc was used for MAS of progenies carrying ZmC4Ppc gene in rice and some restorer lines with ZmC4Ppc （e.g. FPM881） derived from ZmC4Ppc-transformed Kitaake backcrossed with a restorer line Shuhui 881 were obtained. The analyses on genetic background, PEPCase activity, net photosynthetic rate, general combining ability （GCA） and specific combining ability （SCA） of FPM881 showed that similarity of genetic background reached above 95%, the PEPCase and net photosynthetic rate were higher than those of the control, and some of the progenies carrying ZmC4Ppc gene had better GCA and SCA for grain yield per plant, number of panicles per plant, and 1000-grain weight than those of the control. This suggested that the introduction of maize ZmC4Ppc gene via MAS and its stable expression could increase grain yield of rice and would likely provide a pathway for rice varietal improvement.

Biplot Analysis of Genotype by Environment for Cooking Quality in Hybrid Rice: A Tool for Line × Tester Data

A study of combining ability for improving rice cooking quality was carried out via genotype plus genotype x environment （GGE） biplot. Four restorer lines and three male sterile lines were used to obtain F1 in a line x tester trial at the Rice Research Institute, Amol, Iran in 2009. GGE biplot analysis showed that Neda and IR56 were the best general combiners for amylose content （AC）, whereas Nemat and IR28 had the highest general combining ability （GCA） effects for gelatinization temperature （GT）, and IR58 and IR59 showed the highest GCA effects in terms of gel consistency （GC）. Meanwhile IR58 and IR59 showed large specific combining ability （SCA） effects for AC, while Neda and SA13 had high SCA effects for GT. Nemat and IR28 had large SCA effects for GC. Because intermediate levels ofAC, GT and GC are ideal, Nemat × IR59 was considered as the best possible cross. Based on these results, the GGE biplot showed good potential for identifying suitable parents, heterotic crosses and the best hybrids in line x tester data.

Uncertain Relation Suited to Overfitting of BP Neural Network

A general uncertainty relation between the change of weighted value which represents learning ability and the discrimination error of unlearning sample sets which represents generalization ability is revealed in the modeling of back propagation (BP) neural network. Tests of numerical simulation for multitype of complicated functions are carried out to determine the value distribution (1×10?5~5×10?4) of overfitting parameter in the uncertainty relation. Based on the uncertainty relation, the overfitting in the training process of given sample sets using BP neural network can be judged.

Within-and between-population variations in seed and seedling traits of Juglans mandshurica

In order to quantify within-and between-population variation in seed and seedling traits of Juglans mandshurica and reveal the relationship among genetic and environmental variations and select elite families,samples of 50 J.mandshurica families from five natural populations in Liaoning Province,northeast China,were taken to measure seed and seedling traits.The results show that all seed traits varied significantly among families within the population,but only kernel weight and kernel rate showed significant variations among populations.Average values of single seed weight,length,width,lateral diameter,and average size,and kernel weight and rate were 10.1 g,43.0 mm,29.2 mm,28.1 mm,33.4 mm,and 2.2 g and 22.5%,respectively.Significant variations were observed in seedling height and root collar diameter among families and interaction between families and blocks,but the block effects on height and root collar diameter were insignificant.Average values of height and root collar diameter were 94.0 cm and 8.7 mm,respectively.Family heritability of traits ranged from 0.6 gm(kernel weight)to 0.9 mm(seedling height).Correlation analysis showed a strong relationship among seed traits but a weak correlation between seed and seedling traits.Cluster analysis grouped the five natural populations of J.mandshurica into three significant clusters with different characteristics.The general combining ability analysis showed that most traits in one family(WD11)were higher,suggesting parental traits were excellent for selecting elite parent clones.Using the comprehensive evaluation method,five families with better seed traits and five families with better seedling traits were selected as elite materials with 10%selection rate.The genetic gains of these elite materials for seed weight,kernel weight,average seed size,kernel rate,seedling height,and root collar diameter were 13.1%,10.3%,4.1%,2.4%,29.7%,and 21.1%,respectively.