@article{zhang_ahn_lin_park_2006, title={Gene selection using support vector machines with non-convex penalty}, volume={22}, ISSN={["1460-2059"]}, DOI={10.1093/bioinformatics/bti736}, abstractNote={Abstract Motivation: With the development of DNA microarray technology, scientists can now measure the expression levels of thousands of genes simultaneously in one single experiment. One current difficulty in interpreting microarray data comes from their innate nature of ‘high-dimensional low sample size’. Therefore, robust and accurate gene selection methods are required to identify differentially expressed group of genes across different samples, e.g. between cancerous and normal cells. Successful gene selection will help to classify different cancer types, lead to a better understanding of genetic signatures in cancers and improve treatment strategies. Although gene selection and cancer classification are two closely related problems, most existing approaches handle them separately by selecting genes prior to classification. We provide a unified procedure for simultaneous gene selection and cancer classification, achieving high accuracy in both aspects. Results: In this paper we develop a novel type of regularization in support vector machines (SVMs) to identify important genes for cancer classification. A special nonconvex penalty, called the smoothly clipped absolute deviation penalty, is imposed on the hinge loss function in the SVM. By systematically thresholding small estimates to zeros, the new procedure eliminates redundant genes automatically and yields a compact and accurate classifier. A successive quadratic algorithm is proposed to convert the non-differentiable and non-convex optimization problem into easily solved linear equation systems. The method is applied to two real datasets and has produced very promising results. Availability: MATLAB codes are available upon request from the authors. Contact:  hzhang@stat.ncsu.edu Supplementary information:  }, number={1}, journal={BIOINFORMATICS}, author={Zhang, HH and Ahn, J and Lin, XD and Park, C}, year={2006}, month={Jan}, pages={88–95} } @article{kaur_park_lewis_haugh_2006, title={Quantitative model of Ras-phosphoinositide 3-kinase signalling cross-talk based on co-operative molecular assembly}, volume={393}, ISSN={["1470-8728"]}, DOI={10.1042/bj20051022}, abstractNote={In growth-factor-stimulated signal transduction, cell-surface receptors recruit PI3Ks (phosphoinositide 3-kinases) and Ras-specific GEFs (guanine nucleotide-exchange factors) to the plasma membrane, where they produce 3′-phosphorylated phosphoinositide lipids and Ras-GTP respectively. As a direct example of pathway networking, Ras-GTP also recruits and activates PI3Ks. To refine the mechanism of Ras–PI3K cross-talk and analyse its quantitative implications, we offer a theoretical model describing the assembly of complexes involving receptors, PI3K and Ras-GTP. While the model poses the possibility that a ternary receptor–PI3K–Ras complex forms in two steps, it also encompasses the possibility that receptor–PI3K and Ras–PI3K interactions are competitive. In support of this analysis, experiments with platelet-derived growth factor-stimulated fibroblasts revealed that Ras apparently enhances the affinity of PI3K for receptors; in the context of the model, this suggests that a ternary complex does indeed form, with the second step greatly enhanced through membrane localization and possibly allosteric effects. The apparent contribution of Ras to PI3K activation depends strongly on the quantities and binding affinities of the interacting molecules, which vary across different cell types and stimuli, and thus the model could be used to predict conditions under which PI3K signalling is sensitive to interventions targeting Ras.}, number={1}, journal={BIOCHEMICAL JOURNAL}, publisher={Portland Press Ltd.}, author={Kaur, H and Park, CS and Lewis, JM and Haugh, JM}, year={2006}, month={Jan}, pages={235–243} } @article{park_schneider_haugh_2003, title={Kinetic analysis of platelet-derived growth factor receptor/phosphoinositide 3-kinase/Akt signaling in fibroblasts}, volume={278}, ISSN={["0021-9258"]}, DOI={10.1074/jbc.M304968200}, abstractNote={Isoforms of the serine-threonine kinase Akt coordinate multiple cell survival pathways in response to stimuli such as platelet-derived growth factor (PDGF). Activation of Akt is a multistep process, which relies on the production of 3′-phosphorylated phosphoinositide (PI) lipids by PI 3-kinases. To quantitatively assess the kinetics of PDGF receptor/PI 3-kinase/Akt signaling in fibroblasts, a systematic study of this pathway was performed, and a mechanistic mathematical model that describes its operation was formulated. We find that PDGF receptor phosphorylation exhibits positive cooperativity with respect to PDGF concentration, and its kinetics are quantitatively consistent with a mechanism in which receptor dimerization is initially mediated by the association of two 1:1 PDGF/PDGF receptor complexes. Receptor phosphorylation is transient at high concentrations of PDGF, consistent with the loss of activated receptors upon endocytosis. By comparison, Akt activation responds to lower PDGF concentrations and exhibits more sustained kinetics. Further analysis and modeling suggest that the pathway is saturated at the level of PI 3-kinase activation, and that the p110α catalytic subunit of PI 3-kinase contributes most to PDGF-stimulated 3′-PI production. Thus, at high concentrations of PDGF the kinetics of 3′-PI production are limited by the turnover rate of these lipids, while the Akt response is additionally influenced by the rate of Akt deactivation.}, number={39}, journal={JOURNAL OF BIOLOGICAL CHEMISTRY}, publisher={American Society for Biochemistry & Molecular Biology (ASBMB)}, author={Park, CS and Schneider, IC and Haugh, JM}, year={2003}, month={Sep}, pages={37064–37072} }