• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.153-166
• /
• 2021

A main goal of pharmacogenomics studies is to predict individual's drug responsiveness based on high dimensional genetic variables. Due to a large number of variables, feature selection is required in order to reduce the number of variables. The selected features are used to construct a predictive model using machine learning algorithms. In the present study, we applied several hybrid feature selection methods such as combinations of logistic regression, ReliefF, TurF, random forest, and LASSO to a next generation sequencing data set of 400 epilepsy patients. We then applied the selected features to machine learning methods including random forest, gradient boosting, and support vector machine as well as a stacking ensemble method. Our results showed that the stacking model with a hybrid feature selection of random forest and ReliefF performs better than with other combinations of approaches. Based on a 5-fold cross validation partition, the mean test accuracy value of the best model was 0.727 and the mean test AUC value of the best model was 0.761. It also appeared that the stacking models outperform than single machine learning predictive models when using the same selected features.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.43 no.2
• /
• pp.179-186
• /
• 2014

Metabolomics is the study of changes in the metabolic status of an organism as a consequence of drug treatment, environmental influences, nutrition, lifestyle, genetic variations, toxic exposure, disease, stress, etc, through global or comprehensive identification and quantification of every single metabolite in a biological system. Since most chronic diseases have been demonstrated to be linked to nutrition, nutritional metabolomics has great potential for improving our understanding of the relationship between disease and nutritional status, nutrient, or diet intake by exploring the metabolic effects of a specific food challenge in a more global manner, and improving individual health. In particular, metabolite profiling of biofluids, such as blood, urine, or feces, together with multivariate statistical analysis provides an effective strategy for monitoring human metabolic responses to dietary interventions and lifestyle habits. Therefore, studies of nutritional metabolomics have recently been performed to investigate nutrition-related metabolic pathways and biomarkers, along with their interactions with several diseases, based on animal-, individual-, and population-based criteria with the goal of achieving personalized health care in the future. This article introduces analytical technologies and their application to determination of nutritional phenotypes and nutrition-related diseases in nutritional metabolomics.

• Korean Journal of Biological Psychiatry
• /
• v.8 no.2
• /
• pp.226-232
• /
• 2001

The pharmacotherapy of depression has reduced morbidity and improved outcome for many depressive patients. A wide range of classical and new antidepressants are available for their treatment. However, 30-40% of all patients do not respond sufficiently to the initial treatment and present adverse effects. Pharmacogenetics studies the genetic basis of an individual's ability to respond to pharmacotherapy. Recently, some reports on serotonin transporter gene polymorphisms and their influence on the response to antidepressive therapy provide an interesting diagnostic tool in assessing the chances of response to antidepressants. We also investigated the relationship between serotonin transprter polymorphisms(5-HTTLPR) and the long-term effect of the antidepressant treatment. 128 depressive patients were enrolled into 2nd year study. The therapeutic response of each subset was not different at 8th, 16th week, but the subset with homozygote(l/l) of long variant showed a better therapeutic response to antidepressant than the heterozygote(l/s) of long and short variant, which showed a better therapeutic response than the subset with homozygote (s/s) of short variant at 1st year and 2nd year after the antidepressant treatment. This result shows that the serotonin transporter polymorphisms may be related to the long-term effect of antidepressant treatment. The potential for pharmacogenomics, the use of genetic information to guide pharmacotherapy and improve outcome by providing individualized treatment decisions, has gained increasing attention. pharmacogenomics will contribute to individualize drug choice by using genotype to predict positive clinical outcomes, adverse reactions, and levels of drug metabolism. Personalized medicine, the use of marker-assisted diagnosis and targeted therapies derived from an individual molecular profile, will impact the antidepressant therapy and this approach will replace the traditional trial-and-error practice of medicine.

• Korean Journal of Biological Psychiatry
• /
• v.8 no.2
• /
• pp.208-219
• /
• 2001

The pharmacotherapy of schizophrenia exhibits wide inter-individual variabilities in clinical efficacy and adverse effects. Recently, human genetic diversity has been known as one of the essential factors to the variation in human drug response. This suggests that drug therapy should be tailored to the genetic characteristics of the individual. Pharmacogenetics is the field of investigation that attempts to elucidate genetic basis of an individual's responses to pharmacotherapy, considering drug effects divided into two categories as pharmacokinetics and pharmacodynamics. The emerging field of pharmacogenomics, which focuses on genetic determinants of drug response at the level of the entire human genome, is important for development and prescription of safer and more effective individually tailored drugs and will aid in understanding how genetics influence drug response. In schizophrenia, pharmacogenetic studies have shown the role of genetic variants of the cytochrome P450 enzymes such as CYP2D6, CYP2C19, and CYP2A1 in the metabolism of antipsychotic drugs. At the level of drug targets, variants of the dopamine $D_2$, $D_3$ and $D_4$, and 5-$HT_{2A}$ and 5-$HT_{2C}$ receptors have been examined. The pharmacogenetic studies in schizophrenia presently shows controversial findings which may be related to the multiple involvement of genes with relatively small effects and to the lack of standardized phenotypes. For further development in the pharmacogenomics of schizophrenia, there would be required the extensive outcome measures and definitions, and the powerful new tools of genomics, proteomics and so on.

• Korean Journal of Biological Psychiatry
• /
• v.15 no.4
• /
• pp.303-309
• /
• 2008

Objectives : Some reports have suggested that 5-HT5A polymorphism allelic association was associated with depression, however, there has been no report about relationship between the 5-HT5A gene and antidepressant response. We conducted the association study of the 5-HT5A receptor gene polymorphisms (-19G/C,12A/T) and response to citalopram in Korean patients with major depressive disorder(MDD). Methods : A total of 106 patients with major depressive disorder were included in this study. The patient's symptoms were measured by 21-item Hamilton Depression Rating Scale(HAMD) at baseline, week 1, week 2, week 4 and week 8 during citalopram treatment. A Responder to citalopram was defined by 50% reduction of total HAMD scores. To analyze genetic polymorphisms, a polymerase chain reaction based method was used. Results : At week 8, responders were 62, non-responders were 44. No significant differences of genotypes or allelic association in 19G/C and 12A/T polymorphisms were observed between responsive and non-responsive patients. Conclusion : These results do not support the hypothesis that this polymorphism of the HT5A receptor gene is involved in the therapeutic response to citalopram.