- Williams CG, Savolainen O. Inbreeding depression in conifers: Implications for breeding strategy. For. Sci. [Internet]. 1996 [cited 2016 Dec 9];42:102–17. Available from: http://sci-hub.cc/http://www.ingentaconnect.com/content/saf/fs/1996/00000042/00000001/art00014
使用自交作为针叶树的育种工具是有争议的;本主题通过基于广泛的植物和动物的遗传模型，理论和实验结果的回顾来解决。一些支持性证据可从针叶树研究获得。对于大多数针叶树，自交不会是减少有害等位基因的小亚群或精英品系中的近交繁殖抑制的最佳方法;同胞或随机交配是针叶树驯化的早期几代的更好的选择。可能的例外是具有少数致死等位基因的针叶树种。已经研究了很少有机体具有比针叶树更致死的当量，因此最初需要更慢的近交比自交，以防止对低后代存活和成年繁殖力的大损失。近交育种群体还将需要大量的复制品系和每个重复的后代，因为预期每条品系的灭绝概率高。像玉米一样，很少有价值的品系会在最初几代自交。如果近亲繁殖抑制是基于有害突变，则假定其对有害等位基因的严格选择（清除）而下降。在初始清洗阶段之后，自交将是有效的。自交的优点包括完美的分类交配，增加系之间的选择效率和增加系内的均匀性。针叶树近亲繁殖抑制的理论预测已超过实验。操作育种计划不会提供关于近亲繁殖抑制变化的所需数据，但是育种策略的近亲繁殖假设必须通过实验进行测试。我们主张使用实验近交种群研究直接使用近亲繁殖抑制作为育种方法。它提供了第一手的结果，并为长期人口管理决策提供了信心。最大的价值是揭示不可预见的问题，防止不可逆转的错误。作为一个例子，我们概述了快速循环的实验近交种群为松属taeda L的计划，其结合早期选择，快速筛查成年繁殖力和传统遗传测试。近亲繁殖抑郁研究是长期人口管理成功的关键。它已经变得更强大与集成的经典遗传学 - 分子方法，加速育种技术和计算机模拟模型。
使用自交而不是温和形式的近亲繁殖或甚至随机交配在精英育种品系中的问题是相关和及时的。现代高级代针对育种策略依赖于细分的育种群体（例如，Lowe和van Buijtenen 1981，Barnes 1984，Cotterill等人1988，Dean 1990，Namkoong和Kang 1990，McKeand和Bridgwater 1993，Williams和Hamrick 1996）。近交是对育种群体的控制（参见Lindgren和Gregorius 1976的综述），但没有直接用作育种方法。在较小程度上，许多程序中商业规模的控制授粉种子的新选择具有间接刺激对近交系的兴趣。这种生产果园设计使得可以使用少至两个不相关的选择，并仍然产生异型种子用于再造林（参见Williams和Askew 1993中的综述）。在所有情况下，种植的种子被假定为异型杂交，这里的近交的主题仅在育种群体内是相关的。自交作为林木的育种工具，首先是在30年前，使用玉米的经验作为证明（Righter 1960）。 Righter建议开发针叶树自交系和种植园林杂交种子生产，尽管他还指出延迟开花和繁殖能力丧失是潜在的障碍。他的想法没有实施，因为相关的育种计划正处于初期阶段，当时少数的近亲繁殖研究报告了自交种子死亡率高。全世界的早期针叶树项目被管理为近亲繁殖回避，并依赖于开放式生产种子园。然而，优选单一的大的繁殖群体，尽管如今更常使用群体细分策略（参见White 1993，Williams和Hamrick 1995中的综述）。玉米（Zea mays L.）杂交生产的成功是针叶树自交的一个很好的理由。玉米杂种生产开始于Jones（1917）首次在玉米自交系之间表现出杂种优势。到了20世纪20年代，美国农业部利用已经被驯化数千年的这种异花授粉的种子开始了强化玉米自交程序。这个计划需要非常大的繁殖群体，因为繁殖能力和产量低导致只有一小部分可接受的选择（Lindstrom 1939）。据估计，在1939年之前，实际上筛选了超过100,000条系，其中97％在1至3代自交后丢失。在保留的系中，对于一种或多种性状（包括产量）都是有缺陷的（Lindstrom，939，Hallauer和Miranda 1981，第321页）。玉米生育力的丧失是主要障碍，即使种子生产有强烈的方向选择。美国玉米方案中的自我普遍存在，因为与同胞交配相比，它需要更少的时间来开发真正的育种品系（Stringfield 1974，也参见图1）。这不是最有效的手段，玉米的外来或野生亲缘的改良依赖于同胞 - 在自交前交配。否则，如果最初使用自我训练，适当度会大大降低（Goodman 1992）。玉米成功不直接转化为针叶树。对于玉米，粮食产量既是适合性状，又是具有直接经济价值的性状，因此粮食产量的选择是选择继续繁殖。对于针叶树，生殖能力和生存也是适应性特征，但它们没有直接的经济价值。具有经济价值的针叶树性状包括木材质量和生长。对于针叶树，使用育种将意味着重点转移到选择适当的性状，如繁殖能力和生存，可能牺牲直接经济价值的性状。
Use of selfing as a breeding tool for conifers is controversial; this topic is addressed with a review of genetic models, theory, and experimental results based on a wide range of plants and animals. Some supporting evidence is available from conifer studies. For most conifers, selfing will not be the best method for reducing inbreeding depression in small subpopulations or elite lines of deleterious alleles; sib- or random-mating is a better option in the early generations of conifer domestication. Possible exceptions are conifer species that have few lethal alleles. Few organisms have been studied which have more lethal equivalents than conifers, so slower rates of inbreeding than selfing are needed initially to prevent large losses to low offspring survival and adult fecundity. Inbred breeding populations will also require large numbers of replicate lines and progeny per replicate because the probability of extinction for each line is expected to be high. Like maize, few valuable lines will result from selfing in the initial generations. If inbreeding depression is based on deleterious mutations then it is hypothesized to decline with stringent selection against deleterious alleles (purging). After the initial purging phase, selfing would be efficient. Advantages of selfing include perfect assortative mating, increased selection efficacy among lines and increased uniformity within lines. Theoretical predictions for inbreeding depression in conifers have outpaced experimentation. Operational breeding programs will not provide needed data on changes in inbreeding depression, but the inbreeding assumptions for breeding strategies must be tested experimentally. We advocate using experimental inbred populations to study direct use of inbreeding depression as a breeding method. It provides first-hand results and lends confidence to long-term population management decisions. The greatest value will be to reveal unforeseen problems, preventing irreversible mistakes. As an example, we outline a plan for a rapidly cycled experimental inbred population for Pinus taeda L, which combines early selection, rapid screening for adult fecundity, and traditional genetic testing. Inbreeding depression research is central to the success of long-term population management. It has become more powerful with integrated classical genetics-molecular approaches, accelerated breeding techniques, and computer simulation models.
The question of using selfing rather than milder forms of inbreeding or even random mating in elite breeding lines is relevant and timely. Advanced-generation conifer breeding strategies now rely on subdivided breeding populations(e.g., Lowe and van Buijtenen 1981,Barnes 1984, Cotterill et al. 1988, Dean 1990, Namkoong and Kang 1990, McKeand and Bridgwater 1993, Williams and Hamrick 1996). Inbreeding is controlled of breeding populations (see review by Lindgren and Gregorius 1976) but has not been used directly as a breeding method. To a lesser extent, the new option of control-pollinated seed on a commercial scale in many programs has indirectly stimu-lated interest in inbred lines. This production orchard design makes it possible to use as few as two unrelated selections and still produce outcrossed seed for reforestation (see review in Williams and Askew 1993). In all cases, seeds for plantation are assumed to be outcrossed, the topic of inbreeding here is relevant only within breeding populations. Selfing as a breeding tool for forest trees was first advo- cated 3 decades ago, using the experience with maize as a justification (Righter 1960). Righter suggested development of selfed lines for conifers and production of hybrid seed for plantation forestry, although he also noted that delayed flowering and loss of reproductive capacity were potential impediments. His ideas were not implemented because coni- fer breeding programs were in their infancy, and the few inbreeding studies at that time reported high rates of selfed seed mortality. Early conifer programs worldwide were man- aged for inbreeding avoidance and relied upon open-polli- nated production seed orchards. A single, large breeding population was preferred then, although today population subdivision strategies are more commonly used (see review in White 1993, Williams and Hamrick 1995). The success of maize (Zea mays L.) hybrid production is a poor justification for selfing in conifers. Maize hybrid production began when Jones (1917) first demonstrated hybrid vigor between selfed lines of maize. By the 1920s, intensive maize selfing programs were begun by the U.S. Department of Agriculture using this cross-pollinated spe- cies that had been under domestication for several thousands of years. This program required very large breeding popula- tions because loss of reproduc[ive capacity and poor yields resulted in only a fraction of acceptable selections (Lindstrom 1939). It is estimated that more than 100,000 lines were actually screened prior to 1939 and that 97% of these were lost after 1 to 3 generations of selfing. Of the lines retained, all were defective for one or more traits including yield (Lindstrom ! 939, Hallauer and Miranda 1981, p. 321). Loss of fertility in maize was the chief barrier, even though there was strong directional selection for seed production. Selfing prevails in U.S. maize programs because it requires less time to develop a true-breeding line than sib-mating (Stringfield 1974, see also Figure 1) It is not the most efficient means, improvement of exotic or wild relatives of maize rely on sib- mating prior to selfing. Otherwise fitness is reduced substan- tially if selfing is used initially (Goodman 1992). The maize success does not translate directly to conifers. For maize, grain yield is both a fitness trait and a trait of direct economic value and thus selection for grain yield is selection for continued reproduction. For conifers, reproductive ca- pacity and survival are also fitness traits, but they do not have direct economic value. Conifer traits of economic value include wood quality and growth. For conifers, use of in- breeding will mean a temporary shift in emphasis to selection for fitness traits such as reproductive capacity and survival, perhaps at the expense of traits of direct economic value
