Cluster analysis and selection of optimal lines based on the fruit quality traits of F1 progenies of wine grape

Abstract:【Objective】Grape, especially wine grape breeding through crossing is a time consuming pro- cess. Selection of superior lines from F1 progenies through fruit character evaluation in early stage is im- portant for breeders. The main purpose of this paper was to explore the method of selecting superior lines from F1 generation of wine grape based on fruit quality evaluation and cluster analysis.【Methods】The mature fruits of 69 F1 generations from two crossing combinations of‘2-1-3’בAreni’and‘2-1- 3’בRuby Cabernet’, three parents‘( 2-1-3’‘Areni’and‘Ruby Cabernet’) and two control varieties‘( Cabernet Sauvignon 685’and‘Merlot 181’) were used as materials. The vines were all grown in the vineyards of Pomology Institute, Shanxi Academy of Agricultural Science in Taigu, Shanxi Province. During the periods of August to October in 4 years (2013—2016), the fruit quality indexes of berry weight, the ratio of peel to pulp, soluble solids (SS), titratable acid (TA) content, pH value, juice yield  of fresh fruit were measured according to the conventional method. The samples were frozen using liq- uid nitrogen and stored in an ultra-low temperature freezer. The contents of phenolic compounds in peel, pulp and seeds of berries were analyzed for 4 years according to colorimetric method. Then princi- pal component and cluster analysis was performed based on the 4 years’data of 22 fruit quality traits to select superior lines.【Results】There were differences in the 22 fruit quality traits among 74 materials. In the crossing combinations of‘2-1-3’בAreni’and‘2-1-3’בRuby Cabernet’, the fruit characteris- tics of F1 progeny were quantitative traits with continuous distribution. There was no individual in the F1progenies, of which the value of the total phenol content in the peel was somewhere in between the two control varieties‘Cabernet Sauvignon 685’and‘Merlot 181’, while there were 1 to 36 individuals in the F1 progenies, of which the values of the other 21 traits were somewhere in between the two control varieties, and the value of tannin content in the peel of 31-5-2-2 strain was exceptionally in between the two control varieties. Principal component analysis performed on 22 quality indexes of 74 materials showed that the contribution rate of the first 4 main components were 32.26%, 26.25%, 15.20%, and 11.58%, respectively, and the cumulative contribution ratio reached 85.28%. The 22 fruit quality index- es of 74 materials could be simplified to these 4 principal components for comprehensive evaluation of the fruit quality. The four principal components were the composition factor 1 of the phenolic substanc- es in the peel, the composition factor 1 of the phenolic substances in the seed, the composition factor 1 of the fruit quality, and the composition factor 2 of the peel quality. According to the principal compo- nent value, the fruit quality of 31-5-2-1 was closest to that of‘Cabernet Sauvignon 685’. According to the 22 fruit quality indicators, cluster analysis of 74 materials showed that the 22 fruit quality indicators of 44-6-7-1 and 44-6-3-6 was closest to that of‘Cabernet Sauvignon 685’and‘Merlot 181’, respec- tively. According to 4 principal components, cluster analysis of 74 materials showed the 31-5-2-1 and 44-6-3-4 was closest to that of‘Cabernet Sauvignon 685’and‘Merlot 181’, respectively. Based on the results of the principal component analysis and the cluster analysis, 31-5-2-1 and 44-6-7-1 were selected as two optimal lines. The average berry weight of the two superior lines were slightly larger than those of the control varieties, the ratio of the peel to pulp was in between the control varieties; the titrat- able acid content of 44-6-7-1 was higher than those of the control varieties, while the juice yield of 31-5- 2-1 was the lowest. In 44-6-7-1, the contents of the total phenols, procyanidins and total anthocyanins in pericarp, procyanidins in pulp were higher than those of the control varieties, while the contents of total phenols and tannins in pulp and seeds were in between the control varieties. In 31-5-2-1, the contents of total phenols and tannins in peel, total flavonoids in pulp were significantly higher than those in the control cultivars, while the contents of proanthocyanidins in peel, total anthocyanins, tannins in pulp and seeds were significantly lower than those in the control cultivars.【Conclusion】The comprehensive application of principal component analysis and cluster analysis based on quality properties could simplify the evaluation indexes of fruit quality of hybrid offsprings of wine grape. Based on this, compre- hensive evaluation of hybrid progenies of wine grape and selection of superior hybrid lines could be used as reference method for selecting new wine grape lines in the future.