| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 518681 | Journal of Biomedical Informatics | 2012 | 7 Pages |
BackgroundDomperidone treatment for gastroparesis is associated with variable efficacy as well as the potential for side effects. DNA microarray single nucleotide polymorphism (SNP) analysis may help to elucidate the role of genetic variability on the therapeutic effectiveness and toxicity of domperidone.AimThe aim of this study was to identify SNPs that are associated with clinical efficacy and side effects of domperidone treatment for gastroparesis from DNA microarray experiments. This will help develop a strategy for rational selection of patients for domperidone therapy.MethodsDNA samples extracted from the saliva of 46 patients treated with domperidone were analyzed using Affymetrix 6.0 SNP microarrays. Then least angle regression (LARS) was used to select SNPs that are related to domperidone efficacy and side effects. Decision tree based prediction models were constructed with the most correlated features selected by LARS.ResultsUsing the most stable SNP selected by LARS a prediction model for side effects of domperidone achieved (95 ± 0)% true negative rate (TN) and (78 ± 11)% true positive rate (TP) in nested leave-one-out tests. For domperidone efficacy, the prediction based on five most stable SNPs achieved (85 ± 7)% TP and (61 ± 4)% TN. Five identified SNPs are related to ubiquitin mediated proteolysis, epithelial cell signaling, leukocyte, cell adhesion, and tight junction signaling pathways. Genetic polymorphisms in three genes that are related to cancer and hedgehog signaling were found to significantly correlate with efficacy of domperidone.ConclusionLARS was found to be a useful tool for statistical analysis of domperidone-related DNA microarray data generated from a small number of patients.
Graphical abstractFigure optionsDownload full-size imageDownload high-quality image (134 K)Download as PowerPoint slideHighlights► The genome-wide association studies require analysis of thousands of genomes. ► We used Linear Angle Regression for analysis of a small number of patients’ genomes. ► This method generated simple highly accurate genetic predictors of phenotype.
