The analysis of pathways has become a powerful method for discovering information about the development of malignant diseases from healthy tissue. This progression is often driven by complex interactions between different layers of omics data, including genetic mutations, epigenetic changes, and transcriptional alterations. Thus, studying multi-omics data is of utmost importance in discovering correlations between tumor characteristics and clinical outcomes. In this work, we propose an improved multi-omics pathway analysis method based on a sequential matrix factorization technique. Our approach builds upon the padma framework, which utilizes Multiple Factor Analysis to identify aberrant individuals. The primary modification in our method lies in the way we compute genes weights, which we derive through sequential Non-Negative Matrix Factorization. By employing this technique, we extract patient-specific patterns and relationships from each omics layer sequentially, using the information gained from one layer to inform the next. We will validate the performance of the method using both simulation tests and a real-world study utilizing the TCGA breast cancer dataset. In the simulation studies, our proposed method consistently demonstrates improved performance compared to Padma, specifically achieving improvements ranging from 2.43% to 5.51% with the largest gains observed in smaller sample sizes.