PDF(Pubmed)
Abstract：
BACKGROUND: Inbreeding and relationship coefficients are essential for conservation and breeding programs. Whether dealing with a small conserved population or a large commercial population, monitoring the inbreeding rate and designing mating plans that minimize the inbreeding rate and maximize the effective population size is important. Free, open-source, and efficient software may greatly contribute to conservation and breeding programs and help students and researchers. Efficient methods exist for calculating inbreeding coefficients. Therefore, an efficient way of calculating the numerator relationship coefficients is via the inbreeding coefficients. i.e., the relationship coefficient between parents is twice the inbreeding coefficient of their progeny. A dummy progeny is introduced where no progeny exists for a pair of individuals. Calculating inbreeding coefficients is very fast, and finding whether a pair of individuals has a progeny and picking one from multiple progenies is computationally more demanding. Therefore, the R package introduces a dummy progeny for any pair of individuals whose relationship coefficient is of interest, whether they have a progeny or not.
RESULTS: Runtime and peak memory usage were benchmarked for calculating relationship coefficients between two sets of 250 and 800 animals (200,000 dummy progenies) from a pedigree of 2,721,252 animals. The program performed efficiently (200,000 relationship coefficients, which involved calculating 2,721,252 + 200,000 inbreeding coefficients) within 3:45 (mm:ss). Providing the inbreeding coefficients (for real animals), the runtime was reduced to 1:08. Furthermore, providing the diagonal elements of D in A = TDT \' (d), the runtime was reduced to 54s. All the analyses were performed on a machine with a total memory size of 1 GB.
CONCLUSIONS: The R package FnR is free and open-source software with implications in conservation and breeding programs. It proved to be time and memory efficient for large populations and many dummy progenies. Calculation of inbreeding coefficients can be resumed for new animals in the pedigree. Thus, saving the latest inbreeding coefficient estimates is recommended. Calculation of d coefficients (from scratch) was very fast, and there was limited value in storing those for future use.
RESULTS: Runtime and peak memory usage were benchmarked for calculating relationship coefficients between two sets of 250 and 800 animals (200,000 dummy progenies) from a pedigree of 2,721,252 animals. The program performed efficiently (200,000 relationship coefficients, which involved calculating 2,721,252 + 200,000 inbreeding coefficients) within 3:45 (mm:ss). Providing the inbreeding coefficients (for real animals), the runtime was reduced to 1:08. Furthermore, providing the diagonal elements of D in A = TDT \' (d), the runtime was reduced to 54s. All the analyses were performed on a machine with a total memory size of 1 GB.
CONCLUSIONS: The R package FnR is free and open-source software with implications in conservation and breeding programs. It proved to be time and memory efficient for large populations and many dummy progenies. Calculation of inbreeding coefficients can be resumed for new animals in the pedigree. Thus, saving the latest inbreeding coefficient estimates is recommended. Calculation of d coefficients (from scratch) was very fast, and there was limited value in storing those for future use.
摘要：
背景：近亲繁殖和关系系数对于保护和育种计划至关重要。无论是与少数保守人口或大量商业人口打交道，监测近亲繁殖率并设计交配计划以最大程度地降低近亲繁殖率并最大程度地提高有效种群规模是很重要的。免费,开源,和高效的软件可以大大有助于保护和育种计划，并帮助学生和研究人员。存在计算近亲繁殖系数的有效方法。因此,计算分子关系系数的有效方法是通过近交系数。即，亲本之间的关系系数是其后代近交系数的2倍。引入虚拟后代,其中一对个体不存在后代。计算近亲繁殖系数非常快，并且发现一对个体是否具有后代并从多个后代中挑选一个在计算上要求更高。因此,R包为任何一对关系系数感兴趣的个体引入了虚拟后代，不管他们有没有后代.
结果：运行时间和峰值记忆使用是基准的，用于计算来自2,721,252只动物谱系的两组250和800只动物（200,000个虚拟后代）之间的关系系数。程序执行高效(200,000个关系系数，其中涉及在3:45(mm:ss)内计算2,721,252+200,000个近交系数)。提供近亲繁殖系数（对于真实的动物），运行时间减少到1:08。此外，在A=TDT\'(D)中提供D的对角元素，运行时间减少到54s。所有分析均在总内存大小为1GB的机器上进行。
结论：R软件包FnR是免费的开源软件，对保护和育种计划具有重要意义。事实证明，对于大量种群和许多虚拟后代而言，它具有时间和记忆效率。可以针对谱系中的新动物恢复近交系数的计算。因此，建议保存最新的近亲繁殖系数估计。d系数的计算(从头开始)非常快，存储这些以备将来使用的价值有限。
结果：运行时间和峰值记忆使用是基准的，用于计算来自2,721,252只动物谱系的两组250和800只动物（200,000个虚拟后代）之间的关系系数。程序执行高效(200,000个关系系数，其中涉及在3:45(mm:ss)内计算2,721,252+200,000个近交系数)。提供近亲繁殖系数（对于真实的动物），运行时间减少到1:08。此外，在A=TDT\'(D)中提供D的对角元素，运行时间减少到54s。所有分析均在总内存大小为1GB的机器上进行。
结论：R软件包FnR是免费的开源软件，对保护和育种计划具有重要意义。事实证明，对于大量种群和许多虚拟后代而言，它具有时间和记忆效率。可以针对谱系中的新动物恢复近交系数的计算。因此，建议保存最新的近亲繁殖系数估计。d系数的计算(从头开始)非常快，存储这些以备将来使用的价值有限。
