Sun, Yijun

Abstract Network-based methods utilize protein–protein interaction information to identify significantly perturbed subnetworks in cancer and to propose key molecular pathways. Numerous methods have been developed, but to date, a rigorous benchmark analysis to compare the performance of existing approaches is lacking. In this paper, we proposed a novel benchmarking framework using synthetic data and conducted a comprehensive analysis to investigate the ability of existing methods to detect target genes and subnetworks and to control false positives, and how they perform in the presence of topological biases at both gene and subnetwork levels. Our analysis revealed insights into algorithmic performance that were previously unattainable. Based on the results of the benchmark study, we presented a practical guide for users on how to select appropriate detection methods and protein–protein interaction networks for cancer pathway identification, and provided suggestions for future algorithm development.

Abstract Prostate cancer is a significant health concern and the most commonly diagnosed cancer in men worldwide. Understanding the complex process of prostate tumor evolution and progression is crucial for improved diagnosis, treatments, and patient outcomes. Previous studies have focused on unraveling the dynamics of prostate cancer evolution using phylogenetic or lineage analysis approaches. However, those approaches have limitations in capturing the complete disease process or incorporating genomic and transcriptomic variations comprehensively. In this study, we applied a novel computational approach to derive a prostate cancer progression model using multidimensional data from 497 prostate tumor samples and 52 tumor-adjacent normal samples obtained from The Cancer Genome Atlas study. The model was validated using data from an independent cohort of 545 primary tumor samples. By integrating transcriptomic and genomic data, our model provides a comprehensive view of prostate tumor progression, identifies crucial signaling pathways and genetic events, and uncovers distinct transcription signatures associated with disease progression. Our findings have significant implications for cancer research and hold promise for guiding personalized treatment strategies in prostate cancer. Significance: We developed and validated a progression model of prostate cancer using >1,000 tumor and normal tissue samples. The model provided a comprehensive view of prostate tumor evolution and progression.

Background: The identification of cancer driver genes and key molecular pathways has been the focus of large-scale cancer genome studies. Network-based methods detect significantly perturbed subnetworks as putative cancer pathways by incorporating genomics data with the topological information of PPI networks. However, commonly used PPI networks have distinct topological structures, making the results of the same method vary widely when applied to different networks. Furthermore, emerging context-specific PPI networks often have incomplete topological structures, which pose serious challenges for existing subnetwork detection algorithms. Methods: In this paper, we propose a novel method, referred to as MultiFDRnet, to address the above issues. The basic idea is to model a set of PPI networks as a multiplex network to preserve the topological structure of individual networks, while introducing dependencies among them, and, then, to detect significantly perturbed subnetworks on the modeled multiplex network using all the structural information simultaneously. Results: To illustrate the effectiveness of the proposed approach, an extensive benchmark analysis was conducted on both simulated and real cancer data. The experimental results showed that the proposed method is able to detect significantly perturbed subnetworks jointly supported by multiple PPI networks and to identify novel modular structures in context-specific PPI networks.

Abstract Summary Quantifying pairwise sequence similarities is a key step in metagenomics studies. Alignment-free methods provide a computationally efficient alternative to alignment-based methods for large-scale sequence analysis. Several neural network-based methods have recently been developed for this purpose. However, existing methods do not perform well on sequences of varying lengths and are sensitive to the presence of insertions and deletions. In this article, we describe the development of a new method, referred to as AsMac that addresses the aforementioned issues. We proposed a novel neural network structure for approximate string matching for the extraction of pertinent information from biological sequences and developed an efficient gradient computation algorithm for training the constructed neural network. We performed a large-scale benchmark study using real-world data that demonstrated the effectiveness and potential utility of the proposed method. Availability and implementation The open-source software for the proposed method and trained neural-network models for some commonly used metagenomics marker genes were developed and are freely available at www.acsu.buffalo.edu/~yijunsun/lab/AsMac.html. Supplementary information Supplementary data are available at Bioinformatics online.

Bladder cancer is a biologically heterogeneous disease with variable clinical presentations, outcomes and responses to therapy. Thus, the clinical utility of single biomarkers for the detection and prediction of biological behavior of bladder cancer is limited. We have previously identified and validated a bladder cancer diagnostic signature composed of 10 biomarkers, which has been incorporated into a multiplex immunoassay bladder cancer test, Oncuria™. In this study, we evaluate whether these 10 biomarkers can assist in the prediction of bladder cancer clinical outcomes. Tumor gene expression and patient survival data from bladder cancer cases from The Cancer Genome Atlas (TCGA) were analyzed. Alignment between the mRNA expression of 10 biomarkers and the TCGA 2017 subtype classification was assessed. Kaplan–Meier analysis of multiple gene expression datasets indicated that high expression of the combined 10 biomarkers correlated with a significant reduction in overall survival. The analysis of three independent, publicly available gene expression datasets confirmed that multiplex prognostic models outperformed single biomarkers. In total, 8 of the 10 biomarkers from the Oncuria™ test were significantly associated with either luminal or basal molecular subtypes, and thus, the test has the potential to assist in the prediction of clinical outcome.

Background: The Oncuria™ urine test for the detection of bladder cancer measures a multiplex protein signature. In this study, we investigated the influence of urinary cellularity, protein, and hematuria on the performance of the Oncuria™ test in an ex vivo experimental model. Materials and Methods: Pooled urine from healthy subjects was spiked with cultured benign (UROtsa) or malignant cells (T24), cellular proteins, or whole blood. The resulting samples were analyzed using the Oncuria™ test following the manufacturer’s instructions. Results: Urine samples obtained from healthy subjects were negative for bladder cancer by Oncuria™ test criteria. The majority of the manipulated conditions did not result in a false-positive test. The addition of whole blood (high concentration) did result in a false-positive result, but this was abrogated by sample centrifugation prior to analysis. The addition of cellular proteins (high concentration) resulted in a positive Oncuria™ test, and this was unaffected by pre-analysis sample centrifugation. Conclusions: The Oncuria™ multiplex test performed well in the ex vivo experimental model and shows promise for clinical application. The identification of patients who require additional clinical evaluation could reduce the need to subject patients who do not have bladder cancer to frequent, uncomfortable and expensive cystoscopic examinations, thus benefiting both patients and the healthcare system.

Abstract The interpretation of accumulating genomic data with respect to tumor evolution and cancer progression requires integrated models. We developed a computational approach that enables the construction of disease progression models using static sample data. Application to breast cancer data revealed a linear, branching evolutionary model with two distinct trajectories for malignant progression. Here, we used the progression model as a foundation to investigate the relationships between matched primary and metastasis breast tumor samples. Mapping paired data onto the model confirmed that molecular breast cancer subtypes can shift during progression and supported directional tumor evolution through luminal subtypes to increasingly malignant states. Cancer progression modeling through the analysis of available static samples represents a promising breakthrough. Further refinement of a roadmap of breast cancer progression will facilitate the development of improved cancer diagnostics, prognostics, and targeted therapeutics. Significance: Analysis of matched primary and metastatic tumor samples supports a unidirectional, linear cancer evolution process and sheds light on longstanding issues regarding the origins of molecular subtypes and their progression relationships.

Abstract Motivation Cancer subtype classification has the potential to significantly improve disease prognosis and develop individualized patient management. Existing methods are limited by their ability to handle extremely high-dimensional data and by the influence of misleading, irrelevant factors, resulting in ambiguous and overlapping subtypes. Results To address the above issues, we proposed a novel approach to disentangling and eliminating irrelevant factors by leveraging the power of deep learning. Specifically, we designed a deep-learning framework, referred to as DeepType, that performs joint supervised classification, unsupervised clustering and dimensionality reduction to learn cancer-relevant data representation with cluster structure. We applied DeepType to the METABRIC breast cancer dataset and compared its performance to state-of-the-art methods. DeepType significantly outperformed the existing methods, identifying more robust subtypes while using fewer genes. The new approach provides a framework for the derivation of more accurate and robust molecular cancer subtypes by using increasingly complex, multi-source data. Availability and implementation An open-source software package for the proposed method is freely available at http://www.acsu.buffalo.edu/~yijunsun/lab/DeepType.html. Supplementary information Supplementary data are available at Bioinformatics online.