This paper introduces a novel method for analyzing the impact of signaling pathways in gene expression data, called Signaling Pathway Impact Analysis (SPIA). SPIA combines two types of evidence: (i) the overrepresentation of differentially expressed (DE) genes in a pathway and (ii) the actual perturbation of the pathway, measured by propagating expression changes across the pathway's topology. The method calculates two probability values, P_NDE and P_PERT, which represent the significance of DE genes and the perturbation of the pathway, respectively. These values are then combined into a global probability value, P_G, which is used to rank pathways and test the hypothesis that a pathway is significantly perturbed under a given condition. SPIA was tested on four real datasets and simulations. The results showed that SPIA has better specificity and sensitivity compared to existing pathway analysis methods such as over-representation analysis (ORA) and gene set enrichment analysis (GSEA). SPIA was able to identify significant pathways that other methods missed, particularly those where the position of DE genes in the pathway had a significant impact on the pathway's activity. The method also demonstrated independence between the two types of evidence, P_NDE and P_PERT, which allowed for a more accurate assessment of pathway significance. The study also showed that SPIA performs well under various null hypothesis scenarios, with low false positive rates. The method was implemented as an R package available at http://vortex.cs.wayne.edu/ontoexpress/. The results indicate that SPIA provides a more accurate and comprehensive analysis of pathway impact compared to existing methods, making it a valuable tool for systems biology research.This paper introduces a novel method for analyzing the impact of signaling pathways in gene expression data, called Signaling Pathway Impact Analysis (SPIA). SPIA combines two types of evidence: (i) the overrepresentation of differentially expressed (DE) genes in a pathway and (ii) the actual perturbation of the pathway, measured by propagating expression changes across the pathway's topology. The method calculates two probability values, P_NDE and P_PERT, which represent the significance of DE genes and the perturbation of the pathway, respectively. These values are then combined into a global probability value, P_G, which is used to rank pathways and test the hypothesis that a pathway is significantly perturbed under a given condition. SPIA was tested on four real datasets and simulations. The results showed that SPIA has better specificity and sensitivity compared to existing pathway analysis methods such as over-representation analysis (ORA) and gene set enrichment analysis (GSEA). SPIA was able to identify significant pathways that other methods missed, particularly those where the position of DE genes in the pathway had a significant impact on the pathway's activity. The method also demonstrated independence between the two types of evidence, P_NDE and P_PERT, which allowed for a more accurate assessment of pathway significance. The study also showed that SPIA performs well under various null hypothesis scenarios, with low false positive rates. The method was implemented as an R package available at http://vortex.cs.wayne.edu/ontoexpress/. The results indicate that SPIA provides a more accurate and comprehensive analysis of pathway impact compared to existing methods, making it a valuable tool for systems biology research.