JDINAC: joint density-based non-parametric differential interaction network analysis and classification using high-dimensional sparse omics data

Abstract

Motivation: A complex disease is usually driven by a number of genes interwoven into networks, rather than a single gene product. Network comparison or differential network analysis has become an important means of revealing the underlying mechanism of pathogenesis and identifying clinical biomarkers for disease classification. Most studies, however, are limited to network correlations that mainly capture the linear relationship among genes, or rely on the assumption of a parametric probability distribution of gene measurements. They are restrictive in real application.

Results: We propose a new Joint density based non-parametric Differential Interaction Network Analysis and Classification (JDINAC) method to identify differential interaction patterns of network activation between two groups. At the same time, JDINAC uses the network biomarkers to build a classification model. The novelty of JDINAC lies in its potential to capture non-linear relations between molecular interactions using high-dimensional sparse data as well as to adjust confounding factors, without the need of the assumption of a parametric probability distribution of gene measurements. Simulation studies demonstrate that JDINAC provides more accurate differential network estimation and lower classification error than that achieved by other state-of-the-art methods. We apply JDINAC to a Breast Invasive Carcinoma dataset, which includes 114 patients who have both tumor and matched normal samples. The hub genes and differential interaction patterns identified were consistent with existing experimental studies. Furthermore, JDINAC discriminated the tumor and normal sample with high accuracy by virtue of the identified biomarkers. JDINAC provides a general framework for feature selection and classification using high-dimensional sparse omics data.

Availability and implementation: R scripts available at https://github.com/jijiadong/JDINAC.

Contact: lxie@iscb.org.

Supplementary information: Supplementary data are available at Bioinformatics online.

Fig. 1.

Workflow of JDINAC

Fig. 2.

The scenarios of simulation studies. The blue square and red triangle represents the scatter plots for the two variables in class 0 and class 1 respectively, (a) scenario 1, the two variables is negatively correlated in class 0 and positively correlated in class 1, (b) scenario 2, the two variables are correlated in one group and are independent in the other, (c) scenario 3, the two variables are equally correlated but with different density in the two groups, (d) scenario 4, the two variables are independent in one group and have non-linear relationship in the other group

Fig. 3.

PRC for JDINAC for differential network analysis under scenario1 (a), scenario 2 (b), scenario 3 (c), scenario 4 (d). The differential dependency weights $w_{ij}$ were used as the differential adjacency matrix, ${({\widehat{\delta }}_{ij})}_{p\times p}=I(w_{ij}\ge t)$, $t=1,\dots ,20$

Fig. 4.

ROC curves of 5 methods for the classification under scenario 1 (a), scenario 2 (b), scenario 3 (c) and scenario 4 (d). The asterisk indicates the location where the cutoff of prediction was set to 0.5

Fig. 5.

The differential network of cancer pathway between BRCA tumor samples and controls. An edge presented in the differential network means the dependency of corresponding pair genes is different between two condition-specific groups. The red nodes stand for hub genes. (a) Differential network estimated by JDINAC; The orange edges indicate the top 10 differential dependency pairs. (b) Differential network estimated by DiffCorr; (c) Differential network estimated by DEDN; (d) Differential network estimated by cPLR

Fig. 6.

Summary of the number of edges in the differential networks for the four methods