Browsing by Subject "Monte-Carlo-Simulation"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Publication
    Biometrical tools for heterosis research
    (2010) Schützenmeister, André; Piepho, Hans-Peter
    Molecular biological technologies are frequently applied for heterosis research. Large datasets are generated, which are usually analyzed with linear models or linear mixed models. Both types of model make a number of assumptions, and it is important to ensure that the underlying theory applies for datasets at hand. Simultaneous violation of the normality and homoscedasticity assumptions in the linear model setup can produce highly misleading results of associated t- and F-tests. Linear mixed models assume multivariate normality of random effects and errors. These distributional assumptions enable (restricted) maximum likelihood based procedures for estimating variance components. Violations of these assumptions lead to results, which are unreliable and, thus, are potentially misleading. A simulation-based approach for the residual analysis of linear models is introduced, which is extended to linear mixed models. Based on simulation results, the concept of simultaneous tolerance bounds is developed, which facilitates assessing various diagnostic plots. This is exemplified by applying the approach to the residual analysis of different datasets, comparing results to those of other authors. It is shown that the approach is also beneficial, when applied to formal significance tests, which may be used for assessing model assumptions as well. This is supported by the results of a simulation study, where various alternative, non-normal distributions were used for generating data of various experimental designs of varying complexity. For linear mixed models, where studentized residuals are not pivotal quantities, as is the case for linear models, a simulation study is employed for assessing whether the nominal error rate under the null hypothesis complies with the expected nominal error rate. Furthermore, a novel step within the preprocessing pipeline of two-color cDNA microarray data is introduced. The additional step comprises spatial smoothing of microarray background intensities. It is investigated whether anisotropic correlation models need to be employed or isotropic models are sufficient. A self-versus-self dataset with superimposed sets of simulated, differentially expressed genes is used to demonstrate several beneficial features of background smoothing. In combination with background correction algorithms, which avoid negative intensities and which have already been shown to be superior, this additional step increases the power in finding differentially expressed genes, lowers the number of false positive results, and increases the accuracy of estimated fold changes.
  • Thumbnail Image
    Publication
    Optimum schemes for hybrid maize breeding with doubled haploids
    (2011) Wegenast, Thilo; Melchinger, Albrecht E.
    In hybrid maize breeding, the doubled haploid technique is increasingly replacing conventional recurrent selfing for the development of new lines. In addition, novel statistical methods have become available as a result of enhanced computing facilities. This has opened up many avenues to develop more efficient breeding schemes and selection strategies for maximizing progress from selection. The overall aim of the present study was to compare the selection progress by employing different breeding schemes and selection strategies. Two breeding schemes were considered, each involving selection in two stages: (i) developing DH lines from S0 plants and evaluating their testcrosses in stage one and testcrosses of the promising DH lines in stage two (DHTC) and (ii) early testing for testcross performance of S1 families before production of DH lines from superior S1 families and then evaluating their testcrosses in the second stage (S1TC-DHTC). For both breeding schemes, we examined different selection strategies, in which variance components and budgets varied, the cross and family structure was considered or ignored, and best linear unbiased prediction (BLUP) of testcross performance was employed. The specific objectives were to (1) maximize through optimum allocation of test resources the progress from selection, using the selection gain (ΔG) or the probability to select superior genotypes (P(q)) as well as their standard deviations as criteria, (2) investigate the effect of parental selection, varying variance components and budgets on the optimum allocation of test resources for maximizing the progress from selection, (3) assess the optimum filial generation (S0 or S1) for DH production, (4) compare various selection strategies - sequential selection considering or ignoring the cross and family structure - for maximizing progress from selection, (5) examine the effect of producing a larger number of candidates within promising crosses and S1 families on the progress from selection, and (6) determine the effect of BLUP, where information from genetically related candidates is integrated in the selection criteria, on the progress from selection. For both breeding schemes, the best strategy was to select among all S1 families and/or DH lines ignoring the cross structure. Further, in breeding scheme S1TC-DHTC, the progress from selection increased with variable sizes of crosses and S1 families, i.e., larger numbers of DH lines devoted to superior crosses and S1 families. Parental cross selection strongly influenced the optimum allocation of test resources and, consequently, the selection gain ΔG in both breeding schemes. With an increasing correlation between the mean testcross performance of the parental lines and the mean testcross performance of their progenies, the superiority in progress from selection compared to randomly chosen parents increased markedly, whereas the optimum number of parental crosses decreased in favor of an increased number of test candidates within crosses. With BLUP, information from genetically related test candidates resulted in more precise estimates of their genotypic values and the progress from selection slightly increased for both optimization criteria ΔG and P(q), compared with conventional phenotypic selection. Analytical solutions to enable fast calculations of the optimum allocation of test resources were developed. This analytical approach superseded matrix inversions required for the solution of the mixed model equations. In breeding scheme S1TC-DHTC, the optimum allocation of test resources involved (1) 10 or more test locations at both stages, (2) 10 or fewer parental crosses each with 100 to 300 S1 families at the first stage, and (3) 500 or more DH lines within a low number of parental crosses and S1 families at the second stage. In breeding scheme DHTC, the optimum number of test candidates at the first stage was 5 to 10 times larger, whereas the number of test locations at the first stage and the number of DH lines at the second stage was strongly reduced compared with S1TC-DHTC. The possibility to reduce the number of parental crosses by selection among parental lines is of utmost importance for the optimization of the allocation of test resources and maximization of the progress from selection. Further, the optimum allocation of test resources is crucial to maximize the progress from selection under given economic and quantitative-genetic parameters. By using marker information and BLUP-based genomic selection, more efficient selection strategies could be developed for hybrid maize breeding.
  • Thumbnail Image
    Publication
    Wirtschaftliche Analyse der Tierhaltungsbetriebe um die Metropole Moskau unter besonderer Berücksichtiung von Aufwands- und Ertragsrisiken
    (2017) Droganova, Yulia; Fuchs, Clemens
    The slow modernisation of the agricultural sector in the Russian Federation after the USSR era, the adoption and the ratification of the Basel Accords, the accession of Russia to the World Trade Organisation in 2012, and finally the crisis in the Ukraine, followed by the import ban on numerous agricultural, fishery products from the EU, USA, Canada, Australia, Norway in August 2014 are the most significant problems which found their reflection in this dissertation. This lead to an increased interest to analyse livestock farms in the Moscow region in consideration of risks in order to predict their profitable development. The goal of the current research was to identify the impending bankruptcy of the Russian livestock farms as early as possible in order to engage in efficient counter planning. The majority of the livestock farms in the Moscow region are dairy farms, which was why this type of livestock farming became the main topic of research for this thesis. The classification of dairy farms into solvent and insolvent farms is based on the application of the multivariate discriminant analysis, a bankruptcy predicting method that is widely used by many banks in Europe and the USA. The risk factor is taken into account in the empirical model of the dairy farm by setting up the stochastic Monte Carlo simulation with the most important random variables (prices, yields and interest rate) in order to quantitatively measure their influence on the economic profitability of a typical dairy farm. Following the results of the discriminant analysis, questions concerning the validation of this model were be raised. What measures were required for the dairy farms, classified as insolvent to deter bankruptcy? This question was examined using a cash flow model, summaries of relevant data and requirements for an empirical model of the dairy farms were collected through interviews of subject experts. On the basis of reference scenario/status quo scenario, three main scenarios were created: Scenario 1 Re-structuring, scenario 2 Improvement of Management and Marketing Activities, and scenario 3 Risk analysis, whereby the measures from scenarios 1 and 2 were stochastically simulated in the scenario 3 Risk analysis in order to be able to estimate the economic risks. From the data set of 31 farms, five typical model farms were selected: two correctly classified solvent, two correctly classified insolvent, and one, which showed up as a type 1 error in the discriminant analysis. A reference scenario describes the data period based on the average values of operational performance from 2008-2010, and the individualized data from the Russian statistics of 2011-2013 and forms a data basis for the scenarios 1 to 3. Scenario 1a Restructuring under Russian Insolvency Law is counterpoised to scenario 1b Restructuring under German Insolvency Law. Scenario 2a Improvement of Management and Marketing Activities without Investment and scenario 2b Improvement of Management and Marketing Activities with Investment contains measures to improve management and marketing. Labour costs were doubled, maintenance, repair costs as well as some other costs were adjusted; while the milk yields, the weight of the dairy cows, the silage yields and the yields of pastures, meadows have been estimated with a logistic function. Over a planning period of twelve years, the dairy farms classified as solvent maximised the increase of their equity capital in scenario 2b, which represents the best result compared to all other scenarios considered. Firstly, it has shown that an adequate insolvency law should support the restructuring process, secondly that training and education, consulting, motivation of employees through higher wages can lead to a better-combined performance in comparison to restructuring. In scenario 3 Risk Analysis, ten relevant random variables and their volatility were simulated and analysed within the frame conditions of the initial Scenarios 1 and 2. In addition, the target values selected were: equity after tax, equity change per hectare of agricultural area, internal equity interest and profit after tax. The presented results explain how on one hand, an adequate insolvency law can support the restructuring process and lead to reinstate solvency of the dairy farms. On the other hand, these results confirm, that the improvements in management can also lead to significant positive achievements in operational performance as opposed to restructuring. The farm, which belongs to type 1 error in the discriminant analysis, has ranked as a solvent dairy farm over the planning period of twelve years in all the scenarios considered. In this case, it can be concluded that the simulation model in the researched composition with the multivariate discriminant analysis has indirectly served to be applicable for validation purposes of the determined discriminant function.