摘要
采用计算机模拟的方法研究了NCⅠ、NCⅡ和完全双列杂交(DIA)设计在随交和自交群体及8种样本容量下估计遗传力和显性比的可靠性,并用变异系数和总体变量的容许区间两个指标对一次试验结果估计遗传参数的可靠性作了研究。结果表明:随着家系数增加,三种交配设计估计遗传力和显性比的可靠性都相应提高,但家系容量的增加只能在一定程度上提高显性比估计的可靠性。为了最大限度提高估计遗传参数的可靠性,应尽可能增加家系数,而家系容量只要为5到10就足够了;三种交配设计在任何情况下估计遗传力的可靠性均高于估计显性比的可靠性,特别对NCI和DIA设计,估计显性比的可靠性很低,除非家系数特别大。从一次试验结果估计遗传力和显性比的可靠性,即使在试验样本容量很大的情况下也较差,尤其是估计显性比的可靠性很差。因此应用一次试验结果估计遗传参数,必须重视它们的抽样误差大小。此外还对如何提高遗传参数估计可靠性的途径进行了讨论。
Computer simulation was used to study the reliability of heritability (h_N^2) anddomihance ratio ((H/D)^(1/2)) estimates obtained by NCⅠ, NCⅡ and DIA designs fromrandom mating populations and inbreeding populations with 8 sample sizesrespectively, and the reliability of h_N^2 and (H/D)^(1/2) estimated from the resultsof one-time-experiments which were often used in plant or animal breeding, was studiedusing coefficient of variation and tolerant intervals calculated from above simulating re-sults as the indices. The results showed that the reiability of h_N^2 and (H/D)^(1/2) estimated by all 3mating designs became higher as the number of families increased, but in crease of thefamily size could only lead to improvement of the reliability to a certain extent. Thissuggested, therefore, that in order to obtain the highest reliability, it is necessary thatthe number of families should be as much as possible whereas the family size of5--10 individuals would be satisfactory enough. The reliability of h_N^2 estimates ob-tained from all populations by all mating designs was invariably higher than that of(H/D)^(1/2)p estimates. Especially, the reliability of (H/D)^(1/2) estimates by NCI and DIAdesigns seemed to be very low unless the number of family was very large. The results also indicated that h_N^2 and (H/D)^(1/2) estimates calculated from the re-sults of one-time-experiments fluctuated very much and their reliability was rather low,even when the sample size was large. Consequently, it is necessary therefore to payattention to the sampling error when the genetic parameters were estimated from theresults of one-time-experiments. Methods for the improvement of the reliability of genetic parameters estimated,especially those from the results of one-time-experiments were also discussed.
出处
《作物学报》
CAS
CSCD
北大核心
1993年第6期531-538,共8页
Acta Agronomica Sinica
关键词
交配设计
遗传力
计算机
作物育种
Mating design
Heritability
Dominance ratio
Reliability of genetic parameter extimates
Tolerant intervals
Sample size
