RESUMO
In the era of precision medicine, survival outcome data with high-throughput predictors are routinely collected. Models with an exceedingly large number of covariates are either infeasible to fit or likely to incur low predictability because of overfitting. Variable screening is key in identifying and removing irrelevant attributes. Recent years have seen a surge in screening methods, but most of them rely on some particular modeling assumptions. Motivated by a study on detecting gene signatures for multiple myeloma patients' survival, we propose a model-free L q -norm learning procedure, which includes the well-known Cramér-von Mises and Kolmogorov criteria as two special cases. The work provides an integrative framework for detecting predictors with various levels of impact, such as short- or long-term impact, on censored outcome data. The framework naturally leads to a scheme which combines results from different q to reduce false negatives, an aspect often overlooked by the current literature. We show that our method possesses sure screening properties. The utility of the proposal is confirmed with simulation studies and an analysis of the multiple myeloma study.
RESUMO
The effect of rapid carcass chilling on breast meat quality was evaluated using commercial (COMM) and random-bred (RB) turkeys. Immediately after slaughter, 48 turkeys from COMM or RB line were randomly subjected to one of four chilling methods: 1) water-immersion chilling (WIC) of the carcasses at 0°C ice slurry, 2) WIC after temperature abuse (TA) of the carcasses at 40°C for 30 min (TA-WIC), 3) hot-boning, quarter sectioning, and crust-freeze-air-chilling (HB-(1)/4CFAC) of breast fillets at -12°C, and 4) HB-(1)/4CFAC of fillets after TA of carcasses (TA-HB-(1)/4CFAC). The TA increased carcass and fillet temperatures by â¼1.3 and â¼4.1°C, respectively, regardless of turkey line, whereas HB-(1)/4CFAC of fillets required 28 and 33% of carcass chilling time for COMM and RB, respectively. During chilling, COMM breast pH rapidly reduced from 6.04 to 5.82, resulting in a significantly lower pH than RB after chilling (P < 0.05), whereas COMM R-value sharply increased from 1.17 to 1.43, causing no difference from RB (P > 0.05). Significantly higher L* value and cooking yield (P < 0.05) were seen in the samples of TA and WIC than those of no TA and HB-(1)/4CFAC, respectively, with no difference observed between COMM and RB fillets (P > 0.05). Higher values of hardness, gumminess, and chewiness were found for RB, no TA, and HB-(1)/4CFAC gels than COMM, TA, and WIC, respectively. These results generally indicated that protein quality and textural properties of turkey fillets were improved, regardless of strains or temperature abuse, using HB-(1)/4CFAC technology.
Assuntos
Manipulação de Alimentos/métodos , Qualidade dos Alimentos , Carne/análise , Músculos Peitorais/fisiologia , Perus/fisiologia , Animais , Temperatura Baixa , Culinária , Géis/metabolismo , Temperatura Alta , Carne/normas , Distribuição Aleatória , Perus/genéticaRESUMO
We report an axial-asymmetric high-Q Fabry-Perot cavity supporting nondegenerate Hermite-Gaussian modes of the same mode order. Axial asymmetry of mirror surface was introduced by mechanically grinding off one side of a cylindrical mirror substrate without degrading the original mirror quality. The bases of the resulting Hermite-Gaussian modes were aligned with respect to the direction of grinding, making it possible to prescribe the mirror principal axes.
RESUMO
We report the first experimental observation of the cavity-QED microlaser spectrum, specifically the unconventional frequency pulling brought by a strong atom-cavity coupling at off resonance. The pulling is enhanced quadratically by the atom-cavity coupling to result in a sensitive response to the number of pumping atoms (2.1 kHz per atom maximally). Periodic variation of the pulling due to the coherent Rabi oscillation is also observed as the number of pumping atoms is increased across multiple thresholds.