Satoshi Mitsuyama

A common Ile 823 Met variant of ATP-binding cassette transporter A1 gene (ABCA1) alters high density lipoprotein cholesterol level in Japanese population

Recently, variants in ATP-binding cassette transporter A1 (ABCA1) were demonstrated to be associated with increased level of high density lipoprotein cholesterol (HDL-C) and decreased risk of coronary artery disease (CAD) in Caucasians. However, this is not universally applicable due to the ethnic or environmental differences. In this context, to clarify the effect of ABCA1 in Japanese, we evaluated the phenotypic effects of I/M 823 and R/K 219 variants on the plasma level of HDL-C in 410 patients recruited in our hospital. Subjects with M 823 allele had significantly higher level of HDL-C than those without M823 allele (49.0+/-15.1 vs. 44.9+/-11.5 mg/dl, respectively, P<0.05). This statistical significance did not change even after multiple regression analysis. In contrast, there was no difference in HDL-C level among the genotypes in R/K 219 polymorphism. Further, in our study population an inverse relationship was shown to exist between HDL-C level and incidence of CAD. However, no positive association was observed between those variants and susceptibility to CAD. In this study, we provide evidence that I/M 823 variant, not R/K 219 variant, in ABCA1 is one of the determinants of HDL-C level, suggesting the importance of this gene on lipid metabolism in Japanese.

Automatic classification of urinary sediment images by using a hierarchical modular neural network

We have developed an automated image-classification method for the examination of urinary sediment. Urine contains many kinds of particles of various colors and sizes. To classify these particles automatically, we developed a hierarchical modular neural network (HMNN) to enable accurate classification of urinary-sediment images. Simulations results showed that a neural network with a modular structure can classify artificially generated patterns more accurately than a single neural network (SNN). By using a HMNN, any kind of particle contained in urine can be automatically classified. We compared the classification accuracy when using the HMNN to that with a SNN and found that the classification accuracy for some classes of particles when using the HMNN was 25% to 30% higher than when using the SNN. With the HMNN, the examination accuracy was sufficient to allow automation of the examination process.