Gene identification and survival prediction with L_p Cox regression and novel similarity measure

In this paper, Cox's proportional hazards model with L_p penalty method is developed for simultaneous gene selection and survival prediction. L_p penalty shrinks coefficients and produces some coefficients that are exactly zero, and therefore can be used to identify survival related downstream genes. We also define a novel similarity measure to hunt the regulatory genes that their gene expression changes may be low but they are highly correlated with the selected genes. Experimental results with gene expression data demonstrate that the proposed procedures can be used for identifying important gene clusters that are related to time to death due to cancer and for building parsimonious model for predicting the survival of future patients.