Volume 6 Supplement 6

Beyond the Genome 2012

Open Access

In silico drug screening and potential target identification for hepatocellular carcinoma using support vector machine

  • Wu-Lung R Yang1,
  • Yu-En Lee2,
  • Ming-Huang Chen3,
  • Yu-Wen Liu3,
  • Pei-Ying Lee3,
  • Kun-Mao Chao1, 4, 5 and
  • Chi-Ying F Huang3, 6
BMC Proceedings20126(Suppl 6):P16

DOI: 10.1186/1753-6561-6-S6-P16

Published: 1 October 2012

Hepatocellular carcinoma (HCC) is a severe liver malignancy with few drug treatment options. Drug screening using FDA-approved drugs will provide a fast track in clinical trials if drugs are found effective against HCC. The Connectivity Map (cmap), a large repository of chemical-induced gene expression profiles, provides the opportunity of analyzing drug property with the expression. A support vector machine (SVM) was utilized to classify the effectiveness of drugs against HCC using gene expression profiles in cmap. The classification results will help us to identify significant chemical-sensitivity genes, and to predict the effectiveness of remaining chemicals in cmap, with a prioritized listing for biological verification. The cell viability of four HCC cell lines treated with 146 chemicals was conducted. The SVM successfully classified the effectiveness of chemicals with an average area under the receiver operating curve of 0.9. Chemical sensitivity genes which are possible HCC therapeutic targets, such as MT1E, MYC and GADD45B, were identified with opposite signs of gene differential changes compared with reported HCC patient samples. Several known HCC inhibitors, such as geldanamycin, alvespimycin (histone deacetylase inhibitors) and doxorubicin (chemotherapy drug), were predicted to be effective. Seven out of 23 predicted drugs were cardiac glycosides, suggesting a close link of these drugs to the inhibition of HCC. The study demonstrates a strategy of in silico drug screening using a large repository of microarrays based on initial in vitro drug screening results. The biological verification result can serve as a feedback into the process for the development of a more accurate chemical sensitivity model.

Authors’ Affiliations

Department of Computer Science and Information Engineering, National Taiwan University
Institute of Biotechnology in Medicine, National Yang-Ming University
Institute of Clinical Medicine, National Yang-Ming University
Division of Hematology and Oncology, Department of Medicine, Taipei Veterans General Hospital
Graduate Institute of Biomedical Electronic and Bioinformatics, National Taiwan University
Institute of Biopharmaceutical Sciences, National Yang-Ming University


© Yang et al; licensee BioMed Central Ltd. 2012

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.