Residual network improves the prediction accuracy of genomic selection.

Genetic improvement of complex traits in animal and plant breeding depends on the efficient and accurate estimation of breeding values. Deep learning methods have been shown to be not superior over traditional genomic selection (GS) methods, partially due to the degradation problem (i.e. with the increase of the model depth, the performance of the deeper model deteriorates). Since the deep learning method residual network (ResNet) is designed to solve gradient degradation, we examined its performance and factors related to its prediction accuracy in GS. Here we compared the prediction accuracy of conventional genomic best linear unbiased prediction, Bayesian methods (BayesA, BayesB, BayesC, and Bayesian Lasso), and two deep learning methods, convolutional neural network and ResNet, on three datasets (wheat, simulated and real pig data). ResNet outperformed other methods in both Pearson's correlation coefficient (PCC) and mean squared error (MSE) on the wheat and simulated data. For the pig backfat depth trait, ResNet still had the lowest MSE, whereas Bayesian Lasso had the highest PCC. We further clustered the pig data into four groups and, on one separated group, ResNet had the highest prediction accuracy (both PCC and MSE). Transfer learning was adopted and capable of enhancing the performance of both convolutional neural network and ResNet. Taken together, our findings indicate that ResNet could improve GS prediction accuracy, affected potentially by factors such as the genetic architecture of complex traits, data volume, and heterogeneity.

GWLD: an R package for genome-wide linkage disequilibrium analysis.

Linkage disequilibrium (LD) analysis is fundamental to the investigation of the genetic architecture of complex traits (e.g. human disease, animal and plant breeding) and population structure and evolution dynamics. However, until now, studies primarily focus on LD status between genetic variants located on the same chromosome. Moreover, genome (re)sequencing produces unprecedented numbers of genetic variants, and fast LD computation becomes a challenge. Here, we have developed GWLD, a parallelized and generalized tool designed for the rapid genome-wide calculation of LD values, including conventional D/D', r2, and (reduced) mutual information (MI and RMI) measures. LD between genetic variants within and across chromosomes can be rapidly computed and visualized in either an R package or a standalone C++ software package. To evaluate the accuracy and speed of LD calculation, we conducted comparisons using 4 real datasets. Interchromosomal LD patterns observed potentially reflect levels of selection intensity across different species. Both versions of GWLD, the R package (https://github.com/Rong-Zh/GWLD/GWLD-R) and the standalone C++ software (https://github.com/Rong-Zh/GWLD/GWLD-C++), are freely available on GitHub.

Comparison and Optimization: Research on the Structure of the PET Bottle Bottom Based on the Finite Element Method.

The polyethylene terephthalate (PET) beverage bottle is one of the most common beverage packages in the world, but the bottom of the PET bottle tends to crack due to excessive stress. In this paper, through numerical simulation and finite element analysis, the mechanical properties of four typical geometric models of bottle bottom are studied, and it is determined that "claw flap bottle bottom (CF-bottom)" has the best structure. Then, the shapes of four bottle bottom structures are fine-tuned by using the automatic optimization method. Under the premise of the same material quality, the surface maximum principal stress, the overall maximum principal stress, and the total elastic strain energy of the bottle bottom are reduced by 46.39-71.81%, 38.16-71.50%, and 38.56-61.38%, respectively, while the deformation displacement is also reduced by 0.63 mm-3.43 mm. In contrast to other papers, this paper dispenses with the manual adjustment of various variables, instead adopting automatic shape optimization to obtain a more accurate model. The percentage of maximum principal stress reduction is remarkable, which provides a feasible theoretical guidance for the structural optimization of PET bottle bottom in the production process.