The use of amplified fragment length polymorphism (AFLP) and morphological data to determine heterotic groups in sunflower (Helianthus annuus)

[摘要] English: Breeders would like to predict the outcome of crosses, before producing andtesting lines derived from them in field trials. One way to ensure this is byfinding a correlation between the genetic distances of inbreds and the amountof heterosis obtained by such a hybrid. The aim of this study was todetermine the genetic distances between 12 sunflower inbred lines with theuse of the AFLP technique and to correlate these results with the amount ofheterosis obtained in F1-hybrids.Twelve inbred lines, consisting of six females (lines) and six males (testers),was planted in a glasshouse at the University of the Free State (UFS) inBloemfontein, South Africa. Two experiments were done on these parentallines.Young leaves were collected from each line. DNA was extracted from theleaves and AFLP analysis was performed on the DNA. Six different primercombinations were used, namely: Mse-Cn + Eco-ACA; Mse-CAG + Eco-ACA; Mse-CTC + Eco-ACA; Mse-Cn + Eco-AAC; Mse-CAG + Eco-AAC andMse-CTC + Eco-AAC. The objective was to determine the genetic distancesof the 12 lines with the use of the AFLP technique and different primercombinations. These results would then be used to identify heterotic groupsin a hybrid breeding program.The genetic similarities were lower overall for CMS (A) x restorer (R) crossesthan for AxA or RxR. This was confirmed by Hongtrakul et a/ (1997). Msecn+ Eco-ACA, Mse-CAG + Eco-ACA and Mse-CTC + Eco-AAC identifiedthe highest amount of dissimilarity between female lines 1A and 4A. Thehighest amount of dissimilarity between male lines 14R and 16R wereidentified by Mse-CTC + Eeo-ACA and Mse-CTC + Ecc-AAC. According to Hongtrakul et al (1997), the cluster analysis separated lines intotwo groups, one for A-lines (females) and another for R-lines (males). Thiswas also found in this study. These groupings illustrate the breeding historyand basic heterotic pattern of sunflower.In the second experiment, all six testers were individually crossed with eachline to produce F1 hybrid seed. Thirty six crosses were made and sufficientseed was generated, except from the cross between parental lines 4A and16R that resulted in a sterile hybrid with no seed. Therefore, 4Ax16R wasreplaced with a standard, HV3037. The 36 F1 hybrids were plantedaccording to a randomized complete block design with three replications. Theplant height, flowering date, head diameter, 1000-seed weight, yield and oilpercentage of each hybrid was determined with the Line x Tester analysis.The aim was to determine the combining ability of the inbred lines, todetermine if there are genetic correlations between the differentcharacteristics and to determine the expression of heterosis for the differentcharacteristics.Hybrid 6Ax13R had the latest flowering date and the highest yield. The F1hybrid that had the highest head diameter and highest oil percentage was1Ax12R. Some crosses performed equally or better than their best parentsindicating the presence of heterotic effects. If yield is the most importantselection criteria, the hybrid 6Ax13R would perform the best in a breeding. program as it ranked the highest.Line 1A could be used to improve head diameter, 1000-seed weight, yieldand oil percentage, as it had the highest or second highest GCA effects forthese characteristics. The tester 13R could also be used to increaseplantheight and yield. To increase flowering date and 1000-seedweight, onecan use 16R.F1 hybrid 3Ax15R was the only hybrid that had positive SCA effects for all thecharacteristics measured. The hybrid 4Ax14R was the best specific combiner for 1000-seed weight and yield, while 1Ax11R had the highest positive effectsfor oil percentage.According to the GCA:SCA ratio the SCA was greater, indicating non-additivegene action. This was found in the studies of Merinkovic (1993) whoconcluded that non-additive gene effects controlled yield. Putt (1966) alsofound that non-additive gene effects controlled the inheritance of floweringdate, head diameter and 1000-seed weight.Correlations of interest were that when selecting for increased plant heightone would increase the head diameter, 1000-seed weight and yield, but itwould however, result in a decrease in the oil percentage. By increasing theflowering date, one would increase the oil percentage, but reduce the headdiameter. Doddamani et al (1997) also found that head diameter, 1000-seedweight and plant height had a significant positive correlation with yield. Theyalso found that flowering date had a negative correlation with yield. Theresults of this study thus confirm their results.Flowering date had the highest broad-sense heritability, followed by 1000-seed weight and plant height. Oil percentage had the highest narrow-senseheritability, followed by 1000-seed weight and flowering date.The hybrid with the highest MP and HP heterosis for yield was 6Ax13R.Hybrid 1Ax12R had the highest MP heterosis while 5Ax12R the highest HPheterosis for oil percentage. The three hybrids that expressed the highestheterosis overall, were 1Ax12R, 1Ax13R and 6Ax13R.Seetharam et al (1977) observed a significant positive heterosis for floweringdate, plant height, head diameter, oil percentage and yield. According toSchuster (1964), heterosis for yield for the hybrids was up to 70% better thanthat of the parents. Half the hybrids showed heterosis for plant height (47%better) and heterosis for head diameter was 60%. Popov and Lazarov (1963)as well as Shuravina (1972) found that only a few hybrids exceeded the parents for oil percentage (heterosis of 4.8%). Above is all similar to theresults found in this study.Correlations between genetic distance, heterosis, and hybrid performance foryield in sunflower were estimated. Genetic distances from AFLP fingerprintswere correlated with the amount of heterosis found in F1 hybrids.Mse-CTT + Eco-ACA had the highest correlation with the amount of heterosisin the F1 generation. It can therefore be recommended that this primercombination can be used to identify heterosis for flowering date, headdiameter, yield and oil percentage in hybrids.Therefore, it is possible to correlate the genetic distances found with AFLPdata with the amount of heterosis that can be expected in F1 hybrids. Thismakes it possible to screen thousands of inbred lines and shorten the hybridbreeding program. The number of crosses, trails and amount of labor willdecrease and will result in a lower farm price for hybrid seeds.

[发布机构] University of the Free State