Pine needle oil from crude extract of pine needles has been used as an anti-cancer agent in Traditional Chinese Medicine. The α-pinene is a natural compound isolated from pine needle oil which has been shown anti-cancer activity. In previous study, we found that pine needle oil exhibited significant inhibitory effect on hepatoma carcinoma BEL-7402 cells. In this study, we investigate the inhibition of α-pinene on hepatoma carcinoma BEL-7402 cells in vitro and in vivo and further explore the mechanism. The results show that liver cancer cell growth was inhibited obviously with inhibitory rate of 79.3% in vitro and 69.1% in vivo, Chk1 and Chk2 levels were upregulated, CyclinB, CDC25 and CDK1 levels were downregulated.

Insulin resistance (IR) is a common risk factor for the development of metabolic diseases, and has gradually become a hot issue for research. It was reported that excessive feeding with high fructose induced insulin resistance in both humans and rats. The aim of this study was to investigate the progression of IR and identify potential biomarkers in urine, plasma and fecal extracts of high fructose-fed rats using a (1)H NMR-based metabonomics approach. The biochemical analysis was also performed. The levels of pyruvate and lactate in the plasma of the IR model rats were reduced significantly, and the levels of citrate and α-ketoglutaric acid (α-KG) in their urine, and the levels of succinate in their feces also decreased, suggesting perturbation of energy metabolism. Decreased levels of taurine in urine and fecal extracts during the whole experiment, together with increased levels of creatine/creatinine in urine, revealed liver and kidney injuries. Decreased levels of choline-containing metabolites in urine and increased levels of betaine in urine and plasma demonstrated altered transmethylation. Changes in hippurate, acetate, propionate and n-butyrate levels suggested disturbance of the intestinal flora in the IR rats. This study indicated that (1)H NMR-based metabonomics can provide biochemical information on the progression of IR and offers a non-invasive means for the discovery of potential biomarkers.

This paper presents a new divide-and-conquer based learning approach to radial basis function (RBF) networks, in which a conventional RBF network is divided into several RBF sub-networks. Each of them individually takes an input sub-space as its input. The original network's output then becomes a linear combination of the sub-networks' outputs with the coefficients adaptively learned together with the system parameters of each sub-network. Since this approach reduces the structural complexity of a RBF network by describing a high-dimensional modelling problem via several low-dimensional ones, the network's learning speed is considerably improved as a whole with the comparable generalization capability. The empirical studies have shown its outstanding performance on forecasting two real time series as well as synthetic data. Besides, we have found that the performance of this approach generally varies with the different decompositions of the network's input and the hidden layer. We therefore further explore the decomposition rule with the results verified by the experiments.

No abstract

In this paper, a Divide-and-Cooperate Machine Learning Model (DCML) based Radial Basis Function Network (RBF) is constructed. This DCML is composed of several sub-RBF networks that take some variables as their inputs. The output of DCML is the sum of sub-RBF networks' outputs. The analysis of VC dimension of DCML shows in theory that its structural complexity is less than conventional Extended Radial Basis Function Network (ENRBF). The experimental results have verified that the DCML outperforms conventional ENRBF.