Multi-view BLUP: a promising solution for post-omics data integrative prediction

Abstract

Phenotype prediction is a promising strategy for accelerating molecular plant breeding. Data from multiple sources (called multi-view data) can provide complementary information to characterize a biological object from various aspects. By integrating multi-view information into phenotype prediction, a multi-view best linear unbiased prediction (MVBLUP) method was proposed in this paper. By assigning different weights to measure the importance of different views of data and using a differential evolution algorithm with an early stopping mechanism to adjust the weight vector adaptively, a multi-view kinship matrix was obtained first and then incorporated into the BLUP model for phenotype prediction. To validate the efficiency of MVBLUP, we conducted numerical experiments on four multi-view datasets. Compared to the average performance of the single-view method, the prediction accuracy of the MVBLUP method has improved by a range of 0.038 to 0.201.The results demonstrate that the MVBLUP model is an effective method of integrating multi-view data.