Identification of genetic markers with synergistic survival effect in cancer

Synergistic markers based on inter-covariate association estimate treatment option with lower propensity to covariates

Background

Propensity constitutes a common problem in identifying clinical outcome prediction model whose covariates include the treatment option, which is assumed to be randomly assigned but indeed dependent of other covariates in the training data. The genuine effect of treatment option cannot be elucidated under the influence of propensity. Existing approaches, such as matched-pairs study design, still cannot solve the problem for imbalanced or small datasets.

Methods

This work proposed an anti-propensity estimate of treatment option, which is generated by support vector classifier based on two synergistic markers that represent the lower and upper limits of inter-covariate association level. The algorithm for generating the synergistic markers was illustrated and the performance was evaluated on a public dataset of gene expression levels, which were obtained from surgically excised tumor samples in non-small cell lung cancer (NSCLC) patients where treatment option, i.e., adjuvant therapy or not, was known.

Results

Six covariates represented by the expression levels of ZNF217, ERCC3, PMS1, PIK3CB, BARD1, and MAPK1, were selected to generate two synergistic markers and classifier for estimating the adjuvant therapy option with substantially attenuated propensity. The estimation accuracy attained an area under the receiver-operating characteristics curve, 0.78, in the test set.

Conclusions

The proposed synergistic markers demonstrated a parsimonious and anti-propensity estimation of treatment option, which is ready for the further evaluation and application in the clinical outcome prediction model.

Integrating Multiple Data Sources for Combinatorial Marker Discovery: A Study in Tumorigenesis

Identification of combinatorial markers from multiple data sources is a challenging task in bioinformatics. Here, we propose a novel computational framework for identifying significant combinatorial markers ( s) using both gene expression and methylation data. The gene expression and methylation data are integrated into a single continuous data as well as a (post-discretized) boolean data based on their intrinsic (i.e., inverse) relationship. A novel combined score of methylation and expression data (viz., ) is introduced which is computed on the integrated continuous data for identifying initial non-redundant set of genes. Thereafter, (maximal) frequent closed homogeneous genesets are identified using a well-known biclustering algorithm applied on the integrated boolean data of the determined non-redundant set of genes. A novel sample-based weighted support ( ) is then proposed that is consecutively calculated on the integrated boolean data of the determined non-redundant set of genes in order to identify the non-redundant significant genesets. The top few resulting genesets are identified as potential s. Since our proposed method generates a smaller number of significant non-redundant genesets than those by other popular methods, the method is much faster than the others. Application of the proposed technique on an expression and a methylation data for Uterine tumor or Prostate Carcinoma produces a set of significant combination of markers. We expect that such a combination of markers will produce lower false positives than individual markers.