Iterative bicluster-based Bayesian principal component analysis and least squares for missing-value imputation in microarray and RNA-sequencing data

Microarray and RNA-sequencing (RNA-seq) techniques each produce gene expression data that can be expressed as a matrix that often contains missing values. Thus, a process of missing-value imputation that uses coherence information of the dataset is necessary. Existing imputation methods, such as iterative bicluster-based least squares (bi-iLS), use biclustering to estimate the missing values because genes are only similar under correlative experimental conditions. Also, they use the row average to obtain a temporary complete matrix, but the use of the row average is considered to be a flaw. The row average cannot reflect the real structure of the dataset because the row average only uses the information of an individual row. Therefore, we propose the use of Bayesian principal component analysis (BPCA) to obtain the temporary complete matrix instead of using the row average in bi-iLS. This alteration produces new missing values imputation method called iterative bicluster-based Bayesian principal component analysis and least squares (bi-BPCA-iLS). Several experiments have been conducted on two-dimension independent gene expression datasets, which are microarray (e.g., cell-cycle expression dataset of yeast saccharomyces cerevisiae) and RNA-seq (gene expression data from schizosaccharomyces pombe) datasets. In the case of the microarray dataset, our proposed bi-BPCA-iLS method showed a significant overall improvement in the normalized root mean square error (NRMSE) values of 10.6% from the local least squares (LLS) and 0.6% from the bi-iLS. In the case of the RNA-seq dataset, our proposed bi-BPCA-iLS method showed an overall improvement in the NRMSE values of 8.2% from the LLS and 3.1% from the bi-iLS. The additional computational time of bi-BPCA-iLS is not significant compared to bi-iLS.

Integrative genomic analysis of a novel small nucleolar RNAs prognostic signature in patients with acute myelocytic leukemia

This study mainly used The Cancer Genome Atlas (TCGA) RNA sequencing dataset to screen prognostic snoRNAs of acute myeloid leukemia (AML), and used for the construction of prognostic snoRNAs signature for AML. A total of 130 AML patients with RNA sequencing dataset were used for prognostic snoRNAs screenning. SnoRNAs co-expressed genes and differentially expressed genes (DEGs) were used for functional annotation, as well as gene set enrichment analysis (GSEA). Connectivity Map (CMap) also used for potential targeted drugs screening. Through genome-wide screening, we identified 30 snoRNAs that were significantly associated with the prognosis of AML. Then we used the step function to screen a prognostic signature composed of 14 snoRNAs (SNORD72, SNORD38, U3, SNORA73B, SNORD79, SNORA73, SNORD12B, SNORA74, SNORD116-12, SNORA65, SNORA14, snoU13, SNORA75, SNORA31), which can significantly divide AML patients into high- and low-risk groups. Through GSEA, snoRNAs co-expressed genes and DEGs functional enrichment analysis, we screened a large number of potential functional mechanisms of this prognostic signature in AML, such as phosphatidylinositol 3-kinase-Akt, Wnt, epithelial to mesenchymal transition, T cell receptors, NF-kappa B, mTOR and other classic cancer-related signaling pathways. In the subsequent targeted drug screening using CMap, we also identified six drugs that can be used for AML targeted therapy, they were alimemazine, MG-262, fluoxetine, quipazine, naltrexone and oxybenzone. In conclusion, our current study was constructed an AML prognostic signature based on the 14 prognostic snoRNAs, which may serve as a novel prognostic biomarker for AML.

Identification of a novel snoRNA expression signature associated with overall survival in patients with lung adenocarcinoma: A comprehensive analysis based on RNA sequencing dataset

Since multiple studies have reported that small nucleolar RNAs (snoRNAs) can be serve as prognostic biomarkers for cancers, however, the prognostic values of snoRNAs in lung adenocarcinoma (LUAD) remain unclear. Therefore, the main work of this study is to identify the prognostic snoRNAs of LUAD and conduct a comprehensive analysis. The Cancer Genome Atlas LUAD cohort whole-genome RNA-sequencing dataset is included in this study, prognostic analysis and multiple bioinformatics approaches are used for comprehensive analysis and identification of prognostic snoRNAs. There were seven LUAD prognostic snoRNAs were screened in current study. We also constructed a novel expression signature containing five LUAD prognostic snoRNAs (snoU109, SNORA5A, SNORA70, SNORD104 and U3). Survival analysis of this expression signature reveals that LUAD patients with high risk score was significantly related to an unfavourable overall survival (adjusted P = 0.01, adjusted hazard ratio = 1.476, 95% confidence interval = 1.096-1.987). Functional analysis indicated that LUAD patients with different risk score phenotypes had significant differences in cell cycle, apoptosis, integrin, transforming growth factor beta, ErbB, nuclear factor kappa B, mitogen-activated protein kinase, phosphatidylinositol-3-kinase and toll like receptor signaling pathway. Immune microenvironment analysis also indicated that there were significant differences in immune microenvironment scores among LUAD patients with different risk score. In conclusion, this study identified an novel expression signature containing five LUAD prognostic snoRNAs, which may be serve as an independent prognostic indicator for LUAD patients.

Upregulation of BIRC5 plays essential role in esophageal squamous cell carcinoma

BACKGROUND

Esophageal squamous cell carcinoma (ESCC) is one of the most common cancers in the world, the detection and prognosis of which are still unsatisfactory. Thus, it is essential to explore the factors that may identify ESCC and evaluate the prognosis of ESCC patients.

RESULTS

Both protein and mRNA expression levels of BIRC5 are upregulated in ESCC group rather than non-ESCC group (standardized mean difference > 0). BIRC5 mRNA expression is related to the age, tumor location, lymph node stage and clinical stage of ESCC patients (p < 0.05). BIRC5 expression makes it feasible to distinguish ESCC from non-ESCC (area under the curve > 0.9), and its high expression is related to poor prognosis of ESCC patients (restrictive survival time difference = -0.036, p < 0.05). BIRC5 may play an important role in ESCC by influencing the cell cycle pathway, and CDK1, MAD2L and CDC20 may be the hub genes of this pathway. The transcription factors-MAZ and TFPD1 -are likely to regulate the transcription of BIRC5, which may be one of the factors for the high expression of BIRC5 in ESCC.

CONCLUSIONS

The current study shows that upregulation of BIRC5 may have essential clinical value in ESCC, and contributes to the understanding of the pathogenesis of ESCC.

Downregulation of CDC14B in 5218 breast cancer patients: A novel prognosticator for triple-negative breast cancer

Breast cancer is the most common female malignancy worldwide and the prognosis of triple-negative breast cancer (TNBC) and advanced breast cancer patients is unsatisfying. The exploration of novel prognostic indicators and appropriate targets is crucial for improving the treatment outcomes of breast cancer patients. The cell division cycle protein 14B (CDC14B) is known for its roles in cell cycle control, but its expression status and molecular function in breast cancer is unknown. This study explores the expression patterns and clinical values of CDC14B in breast cancer tissues. For this research, the authors downloaded gene microarrays and RNA sequencing datasets to examine the expression levels of CDC14B in 5218 breast cancer tissues, comparing them to the expression levels in 1176 normal breast tissues. The relationships between CDC14B and clinicopathologic characteristics of breast cancer were also addressed. The mutation conditions of CDC14B were then clarified using cBioPortal. Finally, differentially expressed genes and co-expressed genes related to CDC14B were filtered using the Limma-Voom package. These genes were intersected to conduct functional annotations and to construct a protein-protein interaction network. It was observed that CDC14B was significantly downregulated in breast cancer tissues but not in normal breast tissues (standardized mean difference = -1.17 [-1.50--0.85], area under the curve = 0.88). In addition, CDC14B downregulation was correlated with the poor prognosis of TNBC patients (hazard ratios < 1; p < 0.05). Amplification was detected to be the most frequent alteration of CDC14B. The presence of this alteration forecasted unfavourable overall survival outcomes in breast cancer patients (p < 0.05). Dysregulated genes that co-expressed with CDC14B were pivotal in cell cycle (namely mitotic-nuclear division and DNA packaging complex) and cancer-related signaling pathways (namely the peroxisome proliferators activated receptor [PPAR] signalling pathway and the AMP-activated protein kinase [AMPK] signalling pathway). Moreover, the genes ADIPOQ and CCNE2 were identified as two promising prognostic factors in breast cancer. In summary, CDC14B was downregulated in breast cancer tissue and may be a promising hallmark in TNBC patients. The dysregulated genes co-expressed with CDC14B may play an important role in the development of breast cancer through PPAR and AMPK signalling pathways.