Browsing by Author "Hu, Chenyue W."

Now showing 1 - 8 of 8
  • Thumbnail Image
    Item
    Decoupling Lineage-Associated Genes in Acute Myeloid Leukemia Reveals Inflammatory and Metabolic Signatures Associated With Outcomes
    (Frontiers, 2021) Abbas, Hussein A.; Mohanty, Vakul; Wang, Ruiping; Huang, Yuefan; Liang, Shaoheng; Wang, Feng; Zhang, Jianhua; Qiu, Yihua; Hu, Chenyue W.; Qutub, Amina A.; Dail, Monique; Bolen, Christopher R.; Daver, Naval; Konopleva, Marina; Futreal, Andrew; Chen, Ken; Wang, Linghua; Kornblau, Steven M.; Bioengineering
    Acute myeloid leukemia (AML) is a heterogeneous disease with variable responses to therapy. Cytogenetic and genomic features are used to classify AML patients into prognostic and treatment groups. However, these molecular characteristics harbor significant patient-to-patient variability and do not fully account for AML heterogeneity. RNA-based classifications have also been applied in AML as an alternative approach, but transcriptomic grouping is strongly associated with AML morphologic lineages. We used a training cohort of newly diagnosed AML patients and conducted unsupervised RNA-based classification after excluding lineage-associated genes. We identified three AML patient groups that have distinct biological pathways associated with outcomes. Enrichment of inflammatory pathways and downregulation of HOX pathways were associated with improved outcomes, and this was validated in 2 independent cohorts. We also identified a group of AML patients who harbored high metabolic and mTOR pathway activity, and this was associated with worse clinical outcomes. Using a comprehensive reverse phase protein array, we identified higher mTOR protein expression in the highly metabolic group. We also identified a positive correlation between degree of resistance to venetoclax and mTOR activation in myeloid and lymphoid cell lines. Our approach of integrating RNA, protein, and genomic data uncovered lineage-independent AML patient groups that share biologic mechanisms and can inform outcomes independent of commonly used clinical and demographic variables; these groups could be used to guide therapeutic strategies.
  • Thumbnail Image
    Item
    Inferring causal molecular networks: empirical assessment through a community-based effort
    (Springer Nature, 2016) Hill, Steven M.; Heiser, Laura M.; Cokelaer, Thomas; Unger, Michael; Nesser, Nicole K.; Carlin, Daniel E.; Zhang, Yang; Sokolov, Artem; Paull, Evan O.; Wong, Chris K.; Graim, Kiley; Bivol, Adrian; Wang, Haizhou; Zhu, Fan; Afsari, Bahman; Danilova, Ludmila V.; Favorov, Alexander V.; Lee, Wai Shing; Taylor, Dane; Hu, Chenyue W.; Long, Byron L.; Noren, David P.; Bisberg, Alexander J.; HPN-DREAM Consortium; Mills, Gordon B.; Gray, Joe W.; Kellen, Michael; Norman, Thea; Friend, Stephen; Qutub, Amina A.; Fertig, Elana J.; Guan, Yuanfang; Song, Mingzhou; Stuart, Joshua M.; Spellman, Paul T.; Koeppl, Heinz; Stolovitzky, Gustavo; Saez-Rodriguez, Julio; Mukherjee, Sach; Bioengineering
    It remains unclear whether causal, rather than merely correlational, relationships in molecular networks can be inferred in complex biological settings. Here we describe the HPN-DREAM network inference challenge, which focused on learning causal influences in signaling networks. We used phosphoprotein data from cancer cell lines as well asﾠin silicoﾠdata from a nonlinear dynamical model. Using the phosphoprotein data, we scored more than 2,000 networks submitted by challenge participants. The networks spanned 32 biological contexts and were scored in terms of causal validity with respect to unseen interventional data. A number of approaches were effective, and incorporating known biology was generally advantageous. Additional sub-challenges considered time-course prediction and visualization. Our results suggest that learning causal relationships may be feasible in complex settings such as disease states. Furthermore, our scoring approach provides a practical way to empirically assess inferred molecular networks in a causal sense.
  • Thumbnail Image
    Item
    Mycoplasma contamination of leukemic cell lines alters protein expression determined by reverse phase protein arrays
    (Springer, 2018) Hoff, Fieke W.; Hu, Chenyue W.; Qutub, Amina A.; Qiu, Yihua; Graver, Elizabeth; Hoang, Giang; Chauhan, Manasi; de Bont, Eveline S.J.M.; Kornblau, Steven M.; Bioengineering
    Mycoplasma contamination is a major problem in cell culturing, potentially altering the results of cell line-based experiments in largely uncharacterized ways. To define the consequences of mycoplasma infection at the level of protein expression we utilized the reverse phase protein array technology to analyze the expression of 235 proteins in mycoplasma infected, uninfected post treatment, and never-infected leukemic cell lines. Overall, protein profiles of cultured cells remained relatively stable after mycoplasma infection. However, paired comparisons for individual proteins identified that 18.7% of the proteins significantly changed between the infected and the never-infected cell line samples, and that 14.0% of the proteins significantly altered between the infected and the post treatment samples. Six percent of the proteins were affected in the post treatment samples compared to the never-infected samples, and 7.2% compared to treated cells that had never had mycoplasma infection before. Proteins that were significantly altered in the infected cells were enriched for apoptotic signaling processes and auto-phosphorylation, suggesting an increased cellular stress and a decreased growth rate. In conclusion, this study shows that mycoplasma infection of leukemic cell lines alters the proteins expression levels, potentially confounding experimental results. This reinforces the need for regular testing of mycoplasma.
  • Thumbnail Image
    Item
    Progeny Clustering: A Method to Identify Biological Phenotypes
    (Nature Publishing Group, 2015) Hu, Chenyue W.; Kornblau, Steven M.; Slater, John H.; Qutub, Amina A.; Bioengineering
    Estimating the optimal number of clusters is a major challenge in applying cluster analysis to any type of dataset, especially to biomedical datasets, which are high-dimensional and complex. Here, we introduce an improved method, Progeny Clustering, which is stability-based and exceptionally efficient in computing, to find the ideal number of clusters. The algorithm employs a novel Progeny Sampling method to reconstruct cluster identity, a co-occurrence probability matrix to assess the clustering stability, and a set of reference datasets to overcome inherent biases in the algorithm and data space. Our method was shown successful and robust when applied to two synthetic datasets (datasets of two-dimensions and ten-dimensions containing eight dimensions of pure noise), two standard biological datasets (the Iris dataset and Rat CNS dataset) and two biological datasets (a cell phenotype dataset and an acute myeloid leukemia (AML) reverse phase protein array (RPPA) dataset). Progeny Clustering outperformed some popular clustering evaluation methods in the ten-dimensional synthetic dataset as well as in the cell phenotype dataset, and it was the only method that successfully discovered clinically meaningful patient groupings in the AML RPPA dataset.
  • Thumbnail Image
    Item
    progenyClust: an R package for Progeny Clustering
    (The R Foundation, 2016) Hu, Chenyue W.; Qutub, Amina A.
    Identifying the optimal number of clusters is a common problem faced by data scientists in various research fields and industry applications. Though many clustering evaluation techniques have been developed to solve this problem, the recently developed algorithm Progeny Clustering is a much faster alternative and one that is relevant to biomedical applications. In this paper, we introduce an R package progenyClust that implements and extends the original Progeny Clustering algorithm for evaluating clustering stability and identifying the optimal cluster number. We illustrate its applicability using two examples: a simulated test dataset for proof-of-concept, and a cell imaging dataset for demonstrating its application potential in biomedical research. The progenyClust package is versatile in that it offers great flexibility for picking methods and tuning parameters. In addition, the default parameter setting as well as the plot and summary methods offered in the package make the application of Progeny Clustering straightforward and coherent.
  • Thumbnail Image
    Item
    Proteomic Profiling of Acute Promyelocytic Leukemia Identifies Two Protein Signatures Associated with Relapse
    (Wiley, 2019) Hoff, Fieke W.; Hu, Chenyue W.; Qutub, Amina A.; Qiu, Yihua; Hornbaker, Marisa J.; Bueso‐Ramos, Carlos; Abbas, Hussein A.; Post, Sean M.; Bont, Eveline S.J.M. de; Kornblau, Steven M.
    Purpose: Acute promyelocytic leukemia (APL) is the most prognostically favorable subtype of Acute myeloid leukemia (AML). Defining the features that allow identification of APL patients likely to relapse after therapy remains challenging. Experimental Design: Proteomic profiling is performed on 20 newly diagnosed APL, 205 non‐APL AML, and 10 normal CD34+ samples using Reverse Phase Protein Arrays probed with 230 antibodies. Results: Comparison between APL and non‐APL AML samples identifies 8.3% of the proteins to be differentially expressed. Proteins higher expressed in APL are involved in the pro‐apoptotic pathways or are linked to higher proliferation. The “MetaGalaxy” approach that considers proteins in relation to other assayed proteins stratifies the APL patients into two protein signatures. All of the relapse patients (n = 4/4) are in protein signature 2 (S2). Comparison of proteins between the signatures shows significant differences in relative expression for 38 proteins. Protein expression summary plots suggest less translational activity in combination with a less proliferative character for S2 compared to signature 1. Conclusions and Clinical Relevance: This study provides a potential proteomic‐based classification of APL patients that may be useful for risk stratification and therapeutic guidance. Validation in a larger independent cohort is required.
  • Thumbnail Image
    Item
    Recapitulation and Modulation of the Cellular Architecture of a User-Chosen Cell of Interest Using Cell-Derived, Biomimetic Patterning
    (American Chemical Society, 2015) Slater, John H.; Culver, James C.; Long, Byron L.; Hu, Chenyue W.; Hu, Jingzhe; Birk, Taylor F.; Qutub, Amina A.; Dickinson, Mary E.; West, Jennifer L.; Bioengineering
    Heterogeneity of cell populations can confound population-averaged measurements and obscure important findings or foster inaccurate conclusions. The ability to generate a homogeneous cell population, at least with respect to a chosen trait, could significantly aid basic biological research and development of high-throughput assays. Accordingly, we developed a high-resolution, image-based patterning strategy to produce arrays of single-cell patterns derived from the morphology or adhesion site arrangement of user-chosen cells of interest (COIs). Cells cultured on both cell-derived patterns displayed a cellular architecture defined by their morphology, adhesive state, cytoskeletal organization, and nuclear properties that quantitatively recapitulated the COIs that defined the patterns. Furthermore, slight modifications to pattern design allowed for suppression of specific actin stress fibers and direct modulation of adhesion site dynamics. This approach to patterning provides a strategy to produce a more homogeneous cell population, decouple the influences of cytoskeletal structure, adhesion dynamics, and intracellular tension on mechanotransduction-mediated processes, and a platform for high-throughput cellular assays.
  • Thumbnail Image
    Item
    Shrinkage Clustering: a fast and size-constrained clustering algorithm for biomedical applications
    (BMC, 2018) Hu, Chenyue W.; Li, Hanyang; Qutub, Amina A.; Bioengineering
    Background: Many common clustering algorithms require a two-step process that limits their efficiency. The algorithms need to be performed repetitively and need to be implemented together with a model selection criterion. These two steps are needed in order to determine both the number of clusters present in the data and the corresponding cluster memberships. As biomedical datasets increase in size and prevalence, there is a growing need for new methods that are more convenient to implement and are more computationally efficient. In addition, it is often essential to obtain clusters of sufficient sample size to make the clustering result meaningful and interpretable for subsequent analysis. Results: We introduce Shrinkage Clustering, a novel clustering algorithm based on matrix factorization that simultaneously finds the optimal number of clusters while partitioning the data. We report its performances across multiple simulated and actual datasets, and demonstrate its strength in accuracy and speed applied to subtyping cancer and brain tissues. In addition, the algorithm offers a straightforward solution to clustering with cluster size constraints. Conclusions: Given its ease of implementation, computing efficiency and extensible structure, Shrinkage Clustering can be applied broadly to solve biomedical clustering tasks especially when dealing with large datasets.