Genetic Variation And QTL Detecting Of Protein Content In Soybean

The soybean contains about 40% protein in seeds. It is not only one of main sources of the plant protein for human diet, but also an important raw material for light industry, medicine, health products and feed industry. The demand for soybean protein in home and abroad markets has increased year by year. In this case, increasing protein content has become an important breeding target in soybean breeding program. Our team has developed a series of advanced backcross lines with the elite variety Zhonghuang13, a popular vareity in Yellow-Huai-Hai Valley as the recurrent parent and 78 soybean mini core soybean accessions as the donor parents. In this study, protein content was analyzed using NIR technique and genetic variation among the 78 soybean mini core donors and relationship between protein content and main agronomic traits were evaluated. The RP, DP, F₂, BC1F₂ and BC₂F₂ derived from four crosses, of which recurrent parent and donor parent were significant different in protein content, were evaluated for its genetic variation in different generations. Detecting and comparing the QTLs in BC₂F₂ random population of Zhonhhuang×Dongshan69 was done using extreme selections on protein content by one-way ANOVA and Chi-square test. The main results were obtained as follows:1. The genetic difference of protein content among 78 soybean mini core donors and correlation between protein content and oil content and main agronomic traitsThere was significant difference in protein content among 78 soybean mini core donors, which ranged from 40.68% to 50.03%, with the mean 45.95%, coefficient of variation was 4.42; there was no significant difference in oil content in two years, but significant different in protein content between 2009 and 2010, which showed that protein content was less affected by annual effect than oil content.The protein content of 78 soybean mini core donors was negatively correlated with oil content, pod No. per plant and seed No. per plant (r=-0.8247, r=-0.1556, r=-0.2054; the protein content was positively correlated with height of the 1st pod.2. Genetic variation of protein content in F₂, BC1F₂ and BC₂F₂The level of recurrent parent and donor parent in protein content would notably impact the performance of hybrid and backcross lines. The mean protein content and the ratio of transgress individuals of hybrid and backcross lines were superior in the crosses which recurrent parent and donor parent were both high in protein and the donor parent was significant higher than recurrent parent. The genetic variation of populations would decrease with backcrossing, in this study variation coefficient of protein content in F₂,BC1F₂ and BC₂F₂ descended in turn. The mean protein content and variation coefficient of BC₂F₂ were both closed to recurrent parent. In the same hybrid type, significant difference between recurrent parent and donor parents brought about great variation coefficient and wide variation range of its hybrid and backcross lines. The protein content of F₂ populations obedient to normal distribution, the BC1F₂ populations of two crosses which donor parents were significantly higher than recurrent parent obedient to skewed distribution, but it recovered normal distribution in BC₂F₂.3. Genetic composition and marker segregation distortion of BC₂F₂ random population and extreme selected populations Chi-square test was conducted in BC₂F₂ random population.56 makers of segregation distortion was detected in total of 90 makers, these makers distributed at 17 linkage groups apart from A2,E and L. 48 makers of segregation distortion was genotypic deviated to recurrent homozygote, and another 8 makers of segregation distortion was genotypic deviated to donor homozygote.4. The analysis of QTL mapping in BC₂F₂ random population and extreme selected populations by one-way ANOVA and Chi-square test(1) QTLs detected by one-way ANOVAThere were 10 QTLs identified in BC₂F₂ random population by one-way ANOVA(P≤0.05), Satt282 and Satt712, located at linkage group D1b and J.The QTLs linkaged to Satt282 and Satt712 was respectively accounted for 13.55 and 13.85% of phenotypic variation of the protein content and showed negative overdominant effects and positive overdominant effects.There were 9 QTLs identified in bilateral selected population, in which 5 QTLs were also detected in BC₂F₂ random population by one-way ANOVA. The QTLs linkaged to Satt282 and Satt712 accounted for 21.40% and 28.82% of phenotypic variation; 5 QTLs were identified in positive selected population; 5 QTLs were identified in negative selected population respectively(2) QTLs detected by Chi-square testQTLs detecting by Chi-square test by actual gene frequencies was more effective than Chi-square test by expected gene frequencies of BC₂F₂ random population due to eliminating the effects of segregation distortion.There were 24 QTLs identified in Bilateral selected population by Chi-square test by gene frequencies in BC₂F₂ random population, in which there were 4 valid loci; 17 QTLs were identified in positive selected population, in which there were 2 valid loci; 23 QTLs were identified in positive selected population, in which there were 5 valid loci.(3) The efficiency of QTLs detecting of different directional selected populationsThe comparative QTLs detection by one-way ANOVA and Chi-square test showed that one-way ANOVA was an effective method for bilateral selected populations, Chi-square test was more effective in QTLs detection for directional selected populations; negative selected population was superior to positive selected population in QTLs detecting whether by one-way ANOVA or Chi-square test.