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 geneperturbation 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.

Patients with metastatic castration-resistant prostate cancer (mCRPC) develop resistance to conventional therapies including docetaxel (DTX). Identifying molecular pathways underlying DTX resistance is critical for developing novel combinatorial therapies to prevent or reverse this resistance. To identify transcriptomic signatures associated with acquisition of chemoresistance we profiled gene expression in DTX-sensitive and -resistant mCRPC cells using RNA sequencing (RNA-seq). PC3 and DU145 cells were selected for DTX resistance and this phenotype was validated by immunoblotting using DTX resistance markers (e.g. clusterin, ABCB1/P-gp, and LEDGF/p75). Overlapping genes differentially regulated in the DTX-sensitive and -resistant cells were ranked by Gene Set Enrichment Analysis (GSEA) and validated to correlate transcript with protein expression. GSEA revealed that genes associated with cancer stem cells (CSC) (e.g., NES, TSPAN8, DPPP, DNAJC12, and MYC) were highly ranked and comprised 70% of the top 25 genes differentially upregulated in the DTX-resistant cells. Established markers of epithelial-to-mesenchymal transition (EMT) and CSCs were used to evaluate the stemness of adherent DTX-resistant cells (2D cultures) and tumorspheres (3D cultures). Increased formation and frequency of cells expressing CSC markers were detected in DTX-resistant cells. DU145-DR cells showed a 2-fold increase in tumorsphere formation and increased DTX resistance compared to DU145-DR 2D cultures. These results demonstrate the induction of a transcriptomic program associated with stemness in mCRPC cells selected for DTX resistance, and strengthen the emerging body of evidence implicating CSCs in this process. In addition, they provide additional candidate genes and molecular pathways for potential therapeutic targeting to overcome DTX resistance.

BackgroundMacrophages show versatile functions in innate immunity, infectious diseases, and progression of cancers and cardiovascular diseases. These versatile functions of macrophages are conducted by different macrophage phenotypes classified as classically activated macrophages and alternatively activated macrophages due to different stimuli in the complex in vivo cytokine environment. Dissecting the regulation of macrophage activations will have a significant impact on disease progression and therapeutic strategy. Mathematical modeling of macrophage activation can improve the understanding of this biological process through quantitative analysis and provide guidance to facilitate future experimental design. However, few results have been reported for a complete model of macrophage activation patterns.ResultsWe globally searched and reviewed literature for macrophage activation from PubMed databases and screened the published experimental results. Temporal in vitro macrophage cytokine expression profiles from published results were selected to establish Boolean network models for macrophage activation patterns in response to three different stimuli. A combination of modeling methods including clustering, binarization, linear programming (LP), Boolean function determination, and semi-tensor product was applied to establish Boolean networks to quantify three macrophage activation patterns. The structure of the networks was confirmed based on protein-protein-interaction databases, pathway databases, and published experimental results. Computational predictions of the network evolution were compared against real experimental results to validate the effectiveness of the Boolean network models.ConclusionThree macrophage activation core evolution maps were established based on the Boolean networks using Matlab. Cytokine signatures of macrophage activation patterns were identified, providing a possible determination of macrophage activations using extracellular cytokine measurements.

No abstract

No abstract