A Conditional Density Estimation Partition Model Using Logistic Gaussian Processes

Conditional density estimation (density regression) estimates the distribution of a response variable y conditional on covariates x. Utilizing a partition model framework, a conditional density estimation method is proposed using logistic Gaussian processes. The partition is created using a Voronoi tessellation and is learned from the data using a reversible jump Markov chain Monte Carlo algorithm. The Markov chain Monte Carlo algorithm is made possible through a Laplace approximation on the latent variables of the logistic Gaussian process model. This approximation marginalizes the parameters in each partition element, allowing an efficient search of the posterior distribution of the tessellation. The method has desirable consistency properties. In simulation and applications, the model successfully estimates the partition structure and conditional distribution of y.

The Appetite-Regulating Hormones Leptin, Adiponectin and Ghrelin and the Development of Prostate Cancer: A Systematic Review

Background: Obesity may be associated with prostate cancer (PC), although the underlying mechanisms are not completely characterized. Levels of the appetite-regulating hormones leptin, adiponectin, and ghrelin, deregulate during obesity and a number of epidemiologic and molecular studies have indicated that they may play a role in PC incidence and progression. We conducted a systematic search and review of the published epidemiologic evidence. Aim: To collect and compare evidence analyzing the serum levels of appetite-regulating hormones in patients with PC, with a view to outline their place in the development of this disease. Methods: Adhering to PRISMA guidelines, we conducted a systematic search of peer-reviewed evidence using the databases MEDLINE/PubMed, Embase, Web of Science, and the Cochrane Library. Predefined eligibility criteria was used to screen the articles, including the study design of “case-control” or “nested case-control”, and measurement of patients' serum hormone levels. Two independent readers screened 609 articles by title and abstract, and 68 by full text. We obtained 34 studies in the final selection that was analyzed by systematic review. Results: Twenty-four studies analyzed leptin levels, 17 analyzed adiponectin levels and 2 analyzed ghrelin levels, either in PC patients versus cancer-free controls or in patients with an aggressive form of the disease compared with patients with a less-aggressive form (classified by Gleason grading or TNM staging). In general, results from case-control studies concluded that leptin levels of PC patients were significantly increased compared with cancer-free controls and low-stage or low-grade controls. Conversely, results from nested case-control studies reported no association between leptin levels and PC outcome, indicating that leptin levels increase only at the time of diagnosis. Frequently, reduced adiponectin levels were predictive of the incidence of PC and the occurrence of the aggressive form in both case-control and nested case-control studies, suggesting that adiponectin levels correlated with PC outcome prior to and during time of diagnosis. Therefore, reduced adiponectin may be an early predictor and potentially contributing factor in PC. The 2 studies that analyzed ghrelin levels reported contradictory results, revealing a lack of data on serum ghrelin in PC. In the studies that we reviewed, there was considerable heterogeneity. For example, some authors provided risk estimates for PC after adjusting for patients´ anthropometric measurements, and authors used inconsistent classifications of high- and low-risk PC. Conclusion: Abnormal levels of leptin and adiponectin may occur during PC development, and adiponectin levels may be deregulated prior to the onset of the disease. Dysregulation of the appetite-regulating hormones may partly underlie the association between obesity and PC, and they may represent a modifiable risk factor or treatment target in the future.

Modified amino acids and peptides as substrates for the intestinal peptide transporter PepT1

The binding affinities of a number of amino-acid and peptide derivatives by the mammalian intestinal peptide transporter PepT1 were investigated, using the Xenopus laevis expression system. A series of blocked amino acids, namely N-acetyl-Phe (Ac-Phe), phe-amide (Phe-NH2), N-acetyl-Phe-amide (Ac-Phe-NH2) and the parent compound Phe, was compared for efficacy in inhibiting the uptake of the peptide [3H]-D-Phe-L-Gln. In an equivalent set of experiments, the blocked peptides Ac-Phe-Tyr, Phe-Tyr-NH2 and Ac-Phe-Tyr-NH2 were compared with the parent compound Phe-Tyr. Comparing amino acids and derivatives, only Ac-Phe was an effective inhibitor of peptide uptake (Ki = 1.81+/- 0.37 mM). Ac-Phe-NH2 had a very weak interaction with PepT1 (Ki = 16.8+/-5.64 mM); neither Phe nor Phe-NH2 interacted with PepT1 with measurable affinity. With the dipeptide and derivatives, unsurprisingly the highest affinity interaction was with Phe-Tyr (Ki = 0.10+/-0.04 mM). The blocked C-terminal peptide Phe-Tyr-NH2 also interacted with PepT1 with a relatively high affinity (Ki = 0.94+/-0.38 mM). Both Ac-Phe-Tyr and Ac-Phe-Tyr-NH2 interacted weakly with PepT1 (Ki = 8.41+/-0.11 and 9.97+/-4.01 mM, respectively). The results suggest that the N-terminus is the primary binding site for both dipeptides and tripeptides. Additional experiments with four stereoisomers of Ala-Ala-Ala support this conclusion, and lead us to propose that a histidine residue is involved in binding the C-terminus of dipeptides. In addition, a substrate binding model for PepT1 is proposed.