Il Do HA, Jianxin PAN, Seungyoung OH, and Youngjo LEE Variable selection methods using a penalized likelihood have been widely studied in various statistical models. However, in semiparametric frailty models, these methods have been relatively less studied because the marginal likelihood function involves analytically intractable integrals, particularly when modeling multicomponent or correlated frailties. In this article, we propose a simple but unified procedure via a penalized h-likelihood (HL) for variable selection of fixed effects in a general class of semiparametric frailty models, in which random effects may be shared, nested, or correlated. We consider three penalty functions (least absolute shrinkage and selection operator [LASSO], smoothly clipped absolute deviation [SCAD], and HL) in our variable selection procedure. We show that the proposed method can be easily implemented via a slight modification to existing HL estimation approaches. Simulation studies also show that the procedure using the SCAD or HL penalty performs well. The usefulness of the new method is illustrated using three practical datasets too. Supplementary materials for the article are available online.

No abstract

Currently, boars selected for commercial use as AI sires are evaluated on grow-finish performance and carcass characteristics. If AI sires were also evaluated and selected on semen production, it may be possible to reduce the number of boars required to service sows, thereby improving the productivity and profitability of the boar stud. The objective of this study was to estimate genetic correlations between production and semen traits in the boar: average daily gain (ADG), backfat thickness (BF) and muscle depth (MD) as production traits, and total sperm cells (TSC), total concentration (TC), volume collected (SV), number of extended doses (ND), and acceptance rate of ejaculates (AR) as semen traits. Semen collection records and performance data for 843 boars and two generations of pedigree data were provided by Smithfield Premium Genetics. Backfat thickness and MD were measured by real-time ultrasound. Genetic parameters were estimated from five four-trait and one five-trait animal models using MTDFREML. Average heritability estimates were 0.39 for ADG, 0.32 for BF, 0.15 for MD, and repeatability estimates were 0.38 for SV, 0.37 for TSC, 0.09 for TC, 0.39 for ND, and 0.16 for AR. Semen traits showed a strong negative genetic correlation with MD and positive genetic correlation with BF. Genetic correlations between semen traits and ADG were low. Therefore, current AI boar selection practices may be having a detrimental effect on semen production.

Given sparse multi-dimensional data (e.g., (user, movie, time; rating) for movie recommendations), how can we discover latent concepts/relations and predict missing values? Tucker factorization has been widely used to solve such problems with multi-dimensional data, which are modeled as tensors. However, most Tucker factorization algorithms regard and estimate missing entries as zeros, which triggers a highly inaccurate decomposition. Moreover, few methods focusing on an accuracy exhibit limited scalability since they require huge memory and heavy computational costs while updating factor matrices.In this paper, we propose P-TUCKER, a scalable Tucker factorization method for sparse tensors. P-TUCKER performs alternating least squares with a row-wise update rule in a fully parallel way, which significantly reduces memory requirements for updating factor matrices. Furthermore, we offer two variants of P-TUCKER: a caching algorithm P-TUCKER-CACHE and an approximation algorithm P-TUCKER-APPROX, both of which accelerate the update process. Experimental results show that P-TUCKER exhibits 1.7-14.1× speed-up and 1.4-4.8× less error compared to the state-of-the-art. In addition, P-TUCKER scales near linearly with the number of observable entries in a tensor and number of threads. Thanks to P-TUCKER, we successfully discover hidden concepts and relations in a large-scale real-world tensor, while existing methods cannot reveal latent features due to their limited scalability or low accuracy.

The objective of this study was to estimate genetic parameters between first and later parities as different traits in reproductive traits of pigs using multiple trait animal model procedures. Data related to reproductive traits from a total of 2,371 individuals maintained at a farm were taken from the pedigree file. Sires and dams were consisted of Duroc, Landrace, and Yorkshire breeds, respectively. The first and later parity records were considered as different traits. Traits included in analyses were total pigs born (TB1), number of pigs born alive (NBA1), number of pigs weaned (NW1), and litter weaning weight (LWT1) in the first parity, and total pigs born (TB2), number of pigs born alive (NBA2), number of pigs weaned (NW2), litter weaning weight (LWT2) and interval between farrowing events (FTF) in later parities. Heritability estimates of TB1, NBA1, NW1 and LWT1 in the first parity were 0.27, 0.25, 0.16 and 0.20, respectively. For TB2, NBA2, NW2, LWT2 and FTF in later parities, heritabilities were estimated as 0.15, 0.15, 0.08, 0.11 and 0.07, respectively. Genetic correlations between sow reproductive traits in the first parity and in the second and later parity were estimated to be 0. 89, 0.77, 0.58 and 0.66, respectively, between TB1 and TB2, NBA1 and NBA2, NW1 and NW2, and LWT1 and LWT2. While phenotypic correlations between TB1 and TB2, NBA1 and NBA2, NW1 and NW2, and LWT1 and LWT2 were estimated as 0.18, 0.15, 0.06 and 0.10, respectively. Genetic correlations between reproductive traits of first and later parities were not high indicating that reproductive traits for sows should be analyzed while considering the parities as different traits.