Multiclass discriminant analysis via adaptive weighted scheme
NEUROCOMPUTING
Authors: Zhao, Haifeng; Zhang, Bowen; Wang, Zheng; Nie, Feiping
Abstract
Under homoscedastic Gaussian assumption, it is demonstrated that conventional LDA is formulated by maximizing the weighted arithmetic mean of the Kullback-Leibler (KL) divergences between different classes. However, the calculation of projection directions is dominated by class pairs with large KL divergence, which causes class pairs with small KL divergence overlapping in the low-dimensional subspace when the dimensionality of subspace is strictly lower than c - 1, wherein c is the class number. Therefore, the classification accuracy is degraded significantly when dealing with classification task. In this paper, we propose a novel method, namely Multiclass Discriminant Analysis via Adaptive Weighted Scheme (MDAAWS), to alleviate the problem mentioned above by assigning weights to all between-class pairs adaptively. Since the proposed problem is a challenging task, an iterative algorithm is exploited to solve it and the corresponding theoretical analysis is presented as well. Extensive experiments conducted on various data sets demonstrate the effectiveness of MDAAWS when compared with some state-of-the-art supervised dimensionality reduction methods. (C) 2019 Elsevier B.V. All rights reserved.
Donor Klotho KL-VS Polymorphism Predicts Allograft Glomerulosclerosis and Early Post-Transplant Kidney Function
TRANSPLANTATION PROCEEDINGS
Authors: Pazik, Joanna; Rembek, Karolina; Sadowska-Jakubowicz, Anna; Sitarek, Elzbieta; Kosieradzki, Maciej; Durlik, Magdalena
Abstract
Background. The Klotho protein, encoded by the KL (Klotho) gene, exerts antiaging and antifibrotic effects. The KL-VS genotype diminishes Klotho expression and correlates with cardiovascular death, heart failure, and chronic kidney disease progression. The aim of this study was to analyze the contribution of donor Klotho rs9536314 and rs9527025 polymorphisms (KL-VS genotype) to renal allograft morphology and function in the early post-transplant period. Methods. Clinical data and biopsy reports of 170 deceased donor transplantations were retrieved from standard medical files. Donor DNA was genotyped for rs9527025 and rs9536314 SNPs using custom TaqMan assays. Results. As rs9527025 remained in full linkage with rs9536314, we report results for the latter. The analyses were performed for G dominant model (GG+GT vs TT). We found an association between reported SNP alleles, morphologic changes in the peritransplant biopsy, and kidney function 3 months after engraftment. A chronic glomerulopathy score of >0 was found in 12.2% of GG+GT cases and in 3.2% of TT cases (P = .023). For G allele carriers, the third month's median estimated glomerular filtration rate value was 35.0 (range, 20.4-76.6 mL/min), while for TT haplotype, the value was 46.3 (range, 15.5-96.8 mL/min), P = .001. At the third post-transplant month, proteinuria incidence was higher for organs with G allele than with TT haplotype (24.4% vs 9.5%; P = .030; odds ratio 3.09; 95% confidence interval 1.22-7.69). Conclusion. Deceased donor KL-VS polymorphism, altering protein dimerization and coreceptor function, predicts early renal transplant glomerular lesions and function. Further analyses for mentioned effect durability are necessary. Ethics Statement. This study complies with the Helsinki Congress and the Istanbul Declaration regarding donor source. Donors were not prisoners, and were not paid or coerced.