MMAP:Mining the Maximum Accuracy of Prediction

The whole genomic selection(GS) is a kind of method for genomic breeding value (GEBV) selection, by detecting the molecular markers which covered the whole genome and getting genetic evaluation which were conducted on individuals by using genetic marker information of the whole genome to obtain a higher accuracy of breeding value estimation. Bayesian model and BLUP (Best Linear Unbiased Prediction) model can all be used as genomic selection methods, such as Bayesian B, Bayesian CPi and Bayesian Lasso etc. in the BLUP model, it contains gBLUP (genomic Best Linear Unbiased Prediction), rrBLUP (Ridge Regression Best Linear Unbiased Prediction) etc.. However, there is no way to predict the accuracy of all phenotype values for different species traits. Each method has its own limition for predicting phenotype traits of species.

Here we give the valuable knowledge database, built according to the popular genome selection method library. At first step, formed a library, and then a better accuracy was got by means of cross validation, it was obtained through predicting the large number of different phenotype data that belonged different species. There are hundreds of phenotype traits which come from different species, and they all have been predicted with an accuracy by different GS methods in the knowledge base. The new GS methods will be integrated to the method library and more phenotype predicting data of species can also be accumulated to the knowledge database.

For predicting of the new treatment phenotype traits, 1) the citing of mining technology was applied to GS methods, and 2) combined the clustering analysis of knowledge database with popular genome prediction methods predict the new phenotype data of different varieties, 3)at last, obtained the corresponding breeding values via a highest prediction accuracy coming from the minimum amount of genome prediction methods. Above all, they provide more help for breeding experts, at the same time reduce the operation time and the use of resources.