A Computational Method of Defining Potential Biomarkers based on Differential Sub-Networks

Abstract

Analyzing omics data from a network-based perspective can facilitate biomarker discovery. To improve disease diagnosis and identify prospective information indicating the onset of complex disease, a computational method for identifying potential biomarkers based on differential sub-networks (PB-DSN) is developed. In PB-DSN, Pearson correlation coefficient (PCC) is used to measure the relationship between feature ratios and to infer potential networks. A differential sub-network is extracted to identify crucial information for discriminating different groups and indicating the emergence of complex diseases. Subsequently, PB-DSN defines potential biomarkers based on the topological analysis of these differential sub-networks. In this study, PB-DSN is applied to handle a static genomics dataset of small, round blue cell tumors and a time-series metabolomics dataset of hepatocellular carcinoma. PB-DSN is compared with support vector machine-recursive feature elimination, multivariate empirical Bayes statistics, analyzing time-series data based on dynamic networks, molecular networks based on PCC, PinnacleZ, graph-based iterative group analysis, KeyPathwayMiner and BioNet. The better performance of PB-DSN not only demonstrates its effectiveness for the identification of discriminative features that facilitate disease classification, but also shows its potential for the identification of warning signals.

Figure 1

Star sub-networks based on ratio 1. (a–d) The star sub-networks consisting of the edges linked with ratio 1 in SG _EWS, SG _EWS-RMS, SG _EWS-BL and SG _EWS-NB, respectively. Ratios in the four sub-networks are same. The numbers of the connections with ratio 1 in SG _EWS, SG _EWS-RMS, SG _EWS-BL and SG _EWS-NB are 370, 10, 32 and 50.

Figure 2

Statistical analysis of the top 5 ratios. (a–e) The statistical analysis (the mean ± the S.E.) of ratio i (1 ≤ i ≤ 5), respectively.

Figure 3

Star sub-networks based on N,N-dimethylglycine/threonic acid and box plot. (a–g) The star sub-networks consisting of the edges linked with N,N-dimethylglycine/threonic acid during disease development. Ratios in the seven sub-networks are same. The numbers of the connections with N,N-dimethylglycine/threonic acid in SG _i (1 ≤ i ≤ 7) are 4, 5, 4, 18, 36, 4 and 4. (h) The box plot of N,N-dimethylglycine/threonic acid.

Figure 4

Statistical analysis of the 3 metabolite ratios. (a–c) The metabolic trajectories (the mean ± the S.E.) of N,N-dimethylglycine/threonic acid, N,N-dimethylglycine/mucic acid and betaine/mucic acid in the training set. (d,e) The ROC curves of these 3 metabolite ratios in the training and test sets.