Background: Cancer has been a leading cause of death in the United States with significant health care costs. Accurate prediction of cancers at an early stage and understanding the genomic mechanisms that drive cancer development are vital to the improvement of treatment outcomes and survival rates, thus resulting in significant social and economic impacts. Attempts have been made to classify cancer types with machine learning techniques during the past two decades and deep learning approaches more recently. Results: In this paper, we established four models with graph convolutional neural network (GCNN) that use unstructured gene expressions as inputs to classify different tumor and non-tumor samples into their designated 33 cancer types or as normal. Four GCNN models based on a co-expression graph, co-expression+singleton graph, protein-protein interaction (PPI) graph, and PPI+singleton graph have been designed and implemented. They were trained and tested on combined 10,340 cancer samples and 731 normal tissue samples from The Cancer Genome Atlas (TCGA) dataset. The established GCNN models achieved excellent prediction accuracies (89.9–94.7%) among 34 classes (33 cancer types and a normal group). In silico gene-perturbation experiments were performed on four models based on co-expression graph, co-expression+singleton, PPI graph, and PPI+singleton graphs. The co-expression GCNN model was further interpreted to identify a total of 428 marker genes that drive the classification of 33 cancer types and normal. The concordance of differential expressions of these markers between the represented cancer type and others are confirmed. Successful classification of cancer types and a normal group regardless of normal tissues' origin suggested that the identified markers are cancer-specific rather than tissue-specific. Conclusion: Novel GCNN models have been established to predict cancer types or normal tissue based on gene expression profiles. We demonstrated the results from the TCGA dataset that these models can produce accurate classification (above 94%), using cancer-specific markers genes. The models and the source codes are publicly available and can be readily adapted to the diagnosis of cancer and other diseases by the data-driven modeling research community.

Cardiac ageing involves the progressive development of cardiac fibrosis and diastolic dysfunction coordinated by MMP-9. Here, we report a cardiac ageing signature that encompasses macrophage pro-inflammatory signalling in the left ventricle (LV) and distinguishes biological from chronological ageing. Young (6–9 months), middle-aged (12–15 months), old (18–24 months), and senescent (26–34 months) mice of both C57BL/6J wild type (WT) and MMP-9 null were evaluated. Using an identified inflammatory pattern, we were able to define individual mice based on their biological, rather than chronological, age. Bcl6, Ccl24, and Il4 were the strongest inflammatory markers of the cardiac ageing signature. The decline in early-to-late LV filling ratio was most strongly predicted by Bcl6, Il1r1, Ccl24, Crp, and Cxcl13 patterns, whereas LV wall thickness was most predicted by Abcf1, Tollip, Scye1, and Mif patterns. With age, there was a linear increase in cardiac M1 macrophages and a decrease in cardiac M2 macrophages in WT mice; of which, both were prevented by MMP-9 deletion. In vitro, MMP-9 directly activated young macrophage polarization to an M1/M2 mid-transition state. Our results define the cardiac ageing inflammatory signature and assign MMP-9 roles in mediating the inflammaging profile by indirectly and directly modifying macrophage polarization. Our results explain early mechanisms that stimulate ageing-induced cardiac fibrosis and diastolic dysfunction.

Background: About 6 million Americans suffer from heart failure and 70% of heart failure cases are caused by myocardial infarction (MI). Following myocardial infarction, increased cytokines induce two major types of macrophages: classically activated macrophages which contribute to extracellular matrix destruction and alternatively activated macrophages which contribute to extracellular matrix construction. Though experimental results have shown the transitions between these two types of macrophages, little is known about the dynamic progression of macrophages activation. Therefore, the objective of this study is to analyze macrophage activation patterns post-MI.Results: We have collected experimental data from adult C57 mice and built a framework to represent the regulatory relationships among cytokines and macrophages. A set of differential equations were established to characterize the regulatory relationships for macrophage activation in the left ventricle post-MI based on the physical chemistry laws. We further validated the mathematical model by comparing our computational results with experimental results reported in the literature. By applying Lyaponuv stability analysis, the established mathematical model demonstrated global stability in homeostasis situation and bounded response to myocardial infarction.Conclusions: We have established and validated a mathematical model for macrophage activation post-MI. The stability analysis provided a possible strategy to intervene the balance of classically and alternatively activated macrophages in this study. The results will lay a strong foundation to understand the mechanisms of left ventricular remodelling post-MI.

Background—: We have previously shown that cardiac sarcopenia occurs with age in C57/BL6J mice. However, underlying mechanisms and plasma biomarkers of cardiac aging have not been identified. Accordingly, the objective of this study was to identify and evaluate plasma biomarkers that reflect cardiac aging phenotypes. Methods and Results—: Plasma from adult (7.5±0.5 months old, n=27) and senescent (31.7±0.5 months old, n=25) C57/BL6J mice was collected, and levels of 69 markers were measured by multi-analyte profiling. Of these, 26 analytes were significantly increased and 3 were significantly decreased in the senescent group compared with the adult group. The majority of analytes that increased in the senescent group were inflammatory markers associated with macrophage functions, including matrix metalloproteinase-9 (MMP-9) and monocyte chemotactic protein-1 (MCP-1/CCL-2). Immunoblotting (n=12/group) showed higher MMP-9 and MCP-1 levels in the left ventricle (LV) of senescent mice ( P <0.05), and their expression levels in the LV correlated with plasma levels (ρ=0.50 for MMP-9 and ρ =0.62 for MCP1, P <0.05). Further, increased plasma MCP-1 and MMP-9 levels correlated with the increase in end-diastolic dimensions that occurs with senescence. Immunohistochemistry (n=3/group) for Mac-3, a macrophage marker, showed increased macrophage densities in the senescent LV, and dual-labeling immunohistochemistry of Mac-3 and MMP-9 revealed robust colocalization of MMP-9 to the macrophages in the senescent LV sections, indicating that the macrophage is a major contributor of MMP-9 in the senescent LV. Conclusions—: Our results suggest that MCP-1 and MMP-9 are potential plasma markers for cardiac aging and that augmented MCP-1 and MMP-9 levels and macrophage content in the LV could provide an underlying inflammatory mechanism of cardiac aging.

Progressive remodeling of the left ventricle (LV) following myocardial infarction (MI) can lead to congestive heart failure, but the underlying initiation factors remain poorly defined. The objective of this study, accordingly, was to determine the key factors and elucidate the regulatory mechanisms of LV remodeling using integrated computational and experimental approaches. By examining the extracellular matrix (ECM) gene expression and plasma analyte levels in C57/BL6J mice LV post-MI and ECM gene responses to transforming growth factor (TGF-β₁) in cultured cardiac fibroblasts, we found that key factors in LV remodeling included macrophages, fibroblasts, transforming growth factor-β₁, matrix metalloproteinase-9 (MMP-9), and specific collagen subtypes. We established a mathematical model to study LV remodeling post-MI by quantifying the dynamic balance between ECM construction and destruction. The mathematical model incorporated the key factors and demonstrated that TGF-β₁ stimuli and MMP-9 interventions with different strengths and intervention times lead to different LV remodeling outcomes. The predictions of the mathematical model fell within the range of experimental measurements for these interventions, providing validation for the model. In conclusion, our results demonstrated that the balance between ECM synthesis and degradation, controlled by interactions of specific key factors, determines the LV remodeling outcomes. Our mathematical model, based on the balance between ECM construction and destruction, provides a useful tool for studying the regulatory mechanisms and for predicting LV remodeling outcomes.

Since its selection as the method of the year in 2013, single-cell technologies have become mature enough to provide answers to complex research questions. With the growth of single-cell profiling technologies, there has also been a significant increase in data collected from single-cell profilings, resulting in computational challenges to process these massive and complicated datasets. To address these challenges, deep learning (DL) is positioned as a competitive alternative for single-cell analyses besides the traditional machine learning approaches. Here, we survey a total of 25 DL algorithms and their applicability for a specific step in the single cell RNA-seq processing pipeline. Specifically, we establish a unified mathematical representation of variational autoencoder, autoencoder, generative adversarial network and supervised DL models, compare the training strategies and loss functions for these models, and relate the loss functions of these models to specific objectives of the data processing step. Such a presentation will allow readers to choose suitable algorithms for their particular objective at each step in the pipeline. We envision that this survey will serve as an important information portal for learning the application of DL for scRNA-seq analysis and inspire innovative uses of DL to address a broader range of new challenges in emerging multi-omics and spatial single-cell sequencing.