Coordinate Descent-Based Sparse Nonnegative Matrix Factorization for Robust Cancer-Class Discovery and Microarray Data Analysis

Determining the number of clusters in high-dimensional real-life datasets and interpreting the final outcome are among the challenging problems in data science. Discovering the number of classes in cancer and microarray data plays a vital role in the treatment and diagnosis of cancers and other related diseases. Nonnegative matrix factorization (NMF) plays a paramount role as an efficient data exploratory tool for extracting basis features inherent in massive data. Some algorithms which are based on incorporating sparsity constraints in the nonconvex NMF optimization problem are applied in the past for analyzing microarray datasets. However, to the best of our knowledge, none of these algorithms use block coordinate descent method which is known for providing closed form solutions. In this paper, we apply an algorithm developed based on columnwise partitioning and rank-one matrix approximation. We test this algorithm on two well-known cancer datasets: leukemia and multiple myeloma. The numerical results indicate that the proposed algorithm performs significantly better than related state-of-the-art methods. In particular, it is shown that this method is capable of robust clustering and discovering larger cancer classes in which the cluster splits are stable.

Transcriptome comparisons detect new genes associated with apoptosis of cattle and buffaloes preantral follicles

Background To develop new breeding technology to improve the breeding ability of bovine, it is the development trend to find the main reason for the occurrence of atresia in these organisms. Transcriptomes of small (100–120 μm) and large (200–220 μm) preantral follicles from cattle and buffalo ovaries were evaluated in vivo and in vitro to understand the transcriptional modulation in preantral follicles that leads to the phenomenon of atresia. Methods The preantral follicles were checked as dead, damage, or live follicles in vivo and in vitro by using trypan blue then bisbenzimide and propidium iodine. Transcriptomes of small (100–120 μm) and large (200–220 μm) preantral follicles of cattle and buffalo were evaluated in vivo and in vitro by microarray and RT-PCR. Healthy preantral follicles were selected based on staining results, and then RNA was extracted from them. Results The viability percentage of preantral follicles in cattle was higher (26.7% and 20%) than buffalo (10%) in vivo and in vitro, respectively. According to the microarray data analysis for cattle preantral follicles, only eleven genes were detected corresponding to five upregulated and six downregulated in large size (200–220 μm) compared to small (100–120 μm) size preantral follicles, while in buffalo, 171 genes were detected (92 upregulated and 79 downregulated) in large size compared to small preantral follicle size. The results of RT-PCR of the selected genes (FASTKD1, BAG2, RHOB, AGTR2, MEF2C, BCL10, G2E3, TM2D1, IGF-I, IGFBP3, PRDX3, and TRIAP1) validated the microarray results. In conclusion, the data of gene expression showed significant differences between small and large sizes in both buffalo and cattle preantral follicles. Conclusion Apoptotic genes were upregulated in the large preantral follicle compared with the small preantral follicles. Moreover, the expression level of these apoptotic genes was significantly upregulated in buffalo than in the cattle. Most of these genes were significantly upregulated in the large buffalo preantral follicle compared with the small size. However, anti-apoptotic genes were upregulated in large cattle preantral follicle and downregulated in large buffalo preantral follicle.

Hsa-miR-3120/miR-214 Mirror miRNAs Alteration in Colorectal Cancer

Colorectal cancer is one of the most common cancers, and various studies have shown that many genes, including miRNAs, play important roles in the development of this cancer. Here, we investigated the molecular and cellular effect of miR-3120 and its mirror miRNA, miR-214, in colorectal cancer using bioinformatics and experimental techniques. Microarray data analysis showed that miR-3120/miR-214 expression are deregulated in colorectal cancer tumors and RT-qPCR analysis of these miRNAs showed a negative expression correlation in different colorectal cancer-originated cell lines. Also, RT-qPCR result indicated that miR-3120 and miR-214 affects each other expression. Overexpression of miR-3120 in HCT-116 cell line followed by qRT-PCR showed increased expression of SMAD3, SMAD4 and AKT2 genes, whereas overexpression of miR-214 inversed this effect. In addition, the expression of specific target genes of each microRNA showed a pattern of co-expression with its microRNAs. Also, investigating the effect of miR-3120/miR-214 on the cell cycle, showed their promoting effect on the progression of the cell cycle. Overall, these data suggest that miR-3120/miR-214 may be involved in regulating the molecular pathways of colorectal cancer, and part of this regulation could be related to the interaction of these genes with each other.

Exploring the Pathogenesis of Psoriasis Complicated With Atherosclerosis via Microarray Data Analysis

BackgroundAlthough more and more evidence has supported psoriasis is prone to atherosclerosis, the common mechanism of its occurrence is still not fully elucidated. The purpose of this study is to further explore the molecular mechanism of the occurrence of this complication.MethodsThe gene expression profiles of psoriasis (GSE30999) and atherosclerosis (GSE28829) were downloaded from the Gene Expression Omnibus (GEO) database. After identifying the common differentially expressed genes (DEGs) of psoriasis and atherosclerosis, three kinds of analyses were performed, namely functional annotation, protein‐protein interaction (PPI) network and module construction, and hub gene identification and co-expression analysis.ResultsA total of 94 common DEGs (24 downregulated genes and 70 upregulated genes) was selected for subsequent analyses. Functional analysis emphasizes the important role of chemokines and cytokines in these two diseases. In addition, lipopolysaccharide-mediated signaling pathway is closely related to both. Finally, 16 important hub genes were identified using cytoHubba, including LYN, CSF2RB, IL1RN, RAC2, CCL5, IRF8, C1QB, MMP9, PLEK, PTPRC, FYB, BCL2A1, LCP2, CD53, NCF2 and TLR2.ConclusionsOur study reveals the common pathogenesis of psoriasis and atherosclerosis. These common pathways and hub genes may provide new ideas for further mechanism research.

RAMRSGL: A Robust Adaptive Multinomial Regression Model for Multicancer Classification

In view of the challenges of the group Lasso penalty methods for multicancer microarray data analysis, e.g., dividing genes into groups in advance and biological interpretability, we propose a robust adaptive multinomial regression with sparse group Lasso penalty (RAMRSGL) model. By adopting the overlapping clustering strategy, affinity propagation clustering is employed to obtain each cancer gene subtype, which explores the group structure of each cancer subtype and merges the groups of all subtypes. In addition, the data-driven weights based on noise are added to the sparse group Lasso penalty, combining with the multinomial log-likelihood function to perform multiclassification and adaptive group gene selection simultaneously. The experimental results on acute leukemia data verify the effectiveness of the proposed method.

A two-stage normalization method for robust differential expression analysis in microarray experiments.

In this research, we introduce an approach to improve the reliability of genetic data analysis. Consistency of the results obtained from microarray data analysis strongly relies on elimination of non-biological variations during data normalization process. Instability in Housekeeping Gene (HKG) expression after performing common normalization methods might be an indication of inefficiency potentially resulting in sampling bias in differential expression analysis. This research aims to reduce the sampling bias in microarray experiments proposing a two-stage normalization algorithm. Proposed approach consists of non-linear Quantile normalization at the first stage and linear HKG based normalization at the second stage. We tested the efficiency of the two-stage normalization method using publicly available microarray datasets obtained from the experiments mainly in the field of reproductive biology. Results show that combined Robust Multiarray Average (RMA) and HKG normalization method reduces the sampling bias in experiments when variations in HKG expression is observed after RMA normalization.

A two-stage normalization method for robust differential expression analysis in microarray experiments.

In this research, we introduce an approach to improve the reliability of genetic data analysis. Consistency of the results obtained from microarray data analysis strongly relies on elimination of non-biological variations during data normalization process. Instability in Housekeeping Gene (HKG) expression after performing common normalization methods might be an indication of inefficiency potentially resulting in sampling bias in differential expression analysis. This research aims to reduce the sampling bias in microarray experiments proposing a two-stage normalization algorithm. Proposed approach consists of non-linear Quantile normalization at the first stage and linear HKG based normalization at the second stage. We tested the efficiency of the two-stage normalization method using publicly available microarray datasets obtained from the experiments mainly in the field of reproductive biology. Results show that combined Robust Multiarray Average (RMA) and HKG normalization method reduces the sampling bias in experiments when variations in HKG expression is observed after RMA normalization.

Molecular Pathways in Low and High-grade Non-muscle Invasive Bladder Cancer as Revealed by Several OMICs Data Analyses

Background: Bladder cancer is one of the most prevalent cancers, accounting for 2.1% of cancer mortalities worldwide. Bladder cancer is categorized into non-muscle invasive and muscle-invasive bladder cancers. Non-muscle invasive bladder cancer (NMIBC) is the most common and widely heterogeneous type with different outcomes. Objectives: This study was designed to categorize NMIBC tumor grade based on microarray data analysis. Methods: We performed microarray data analysis using GSE7476, GSE13507, and GSE37815 in patients diagnosed with NMIBC. Differentially expressed genes (DEGs) were identified based on low-grade and high-grade NMIBC. Protein-protein interaction (PPI) network analysis was carried out, and hub genes and underlying molecular pathways were identified. Results: We observed low-grade Hub genes, including GAS6, TGFB3, TPM1, COL5A1, COL1A2, SERPING1, ACTA2, TPM2, SDC1, and A2M involved in a variety of gene ontology (GO) biological processes, while high-grade genes were involved in cell cycle and cell division. The most relevant pathways suggested for low-grade NMIBC were extracellular matrix organization, platelet degranulation, and muscle contraction. Conclusions: The identification of gene hubs and underlying pathways in several low and high-grade NMIBC samples may offer better treatment management and prognostication based on molecular profiling.