Estimation Of Pairwise Relationship Using Molecular Marker And Software Developing For Estimating Genetic Parameters

The relationship between animals which depends on pedigree records plays a central role in many areas of quantitative and conservation genetics, especially in animal breeding. In the absence of known pedigree structure (Such as wild population or lack of pedigree), we can only use molecular marker to rebuild the relationship between animals to make effective mating plans and estimate the genetic parameters. In this paper, I propose a novel model for estimating pairwise relatedness from randomly mating population with co-dominant genetic markers, comparing the four alleles from two loci in two individuals and taking the sample size, number of loci, number of alleles per locus and the allele frequency distribution into consideration. And I also set up Monte Carlo simulation which can simulate the ideal population of parent-offspring, full sibs, half sibs and nonrelatives to evaluate the new model and compare between different models by mean and sampling variances of relatedness.Giving 10 loci, 10 alleles per locus and 10 cardinal numbers for sample size, we simulate 4 types of relative population for evaluating the precision and accuracy of the new model. The result follows: mean and variance of pairwise relatedness for FS, PO, HS and NR are 0.5039 and 0.0017, 0.5464 and 0.0043, 0.2873 and 0.0064, 0.0654 and 0.0043 respectivlely. It indicates that the new model is well behaved under 4 actual relationships. Comparing with previous estimators, the new one also has stable property and is generally much more advantageous than the others by changing the variable of sample size, number of loci and number of alleles per locus.Finally, I design novel software which has a friendly interface and high extension base on Visual Basic 6.0, and also can be applied easily and effectively. Its main function includes: estimating the pairwise relatedness between pairwise individuals using loading data by the new estimator and the others, and comparing the results between different estimators. Besides, we can also evaluate the precision and accuracy of each model and make a comparison through the simulating system.