Signaling Pathway Analysis Combined With the Strength Variations of Interactions Between Genes Under Different Conditions

Abstract

Signaling pathway analysis has become a routine task after differentially expressed gene (DEG) studies across disease conditions, drug treatments, or developmental stages. A signaling pathway can be represented by a graph that consists of Genes and interactions (the genetic regulation) between them. However, existing signaling pathway analysis methods ignore the strength variations of interactions in signaling pathways under different conditions. Here, we developed a novel method named SPACI (Signaling Pathway Analysis Combined with the strength variations of Interactions between genes in signaling pathways under different conditions) to improve signaling pathway analysis after DEG studies. To further evaluate the performance of SPACI, we compared SPACI with nine other methods by using a benchmark of 28 gene expression datasets in two standard measures: sensitivity and prioritization. The False positive rate (FPR) of SPACI was also compared with five methods. The results show that SPACI is the second-ranked method in terms of prioritization and the third-ranked method in terms of sensitivity. SPACI is the top method when compared in terms of the sum value of the two ranks. Also, the FPR of SPACI is modest compared with the classic methods. Furthermore, the strength variation of the interaction is demonstrated as coherent with the biological problem. The interactions with high strength variations under different conditions can help improve the discovery of the underlying biological information. The R package of SPACI can be accessed at https://github.com/ZhenshenBao/SPACI.

Keywords

• differentially expressed gene
• strength variation of interactions
• signaling pathway analysis
• SPACI