Integrating Constitutive Gene Expression and Chemoactivity: Mining the NCI60 Anticancer Screen

Studies into the genetic origins of tumor cell chemoactivity pose significant challenges to bioinformatic mining efforts. Connections between measures of gene expression and chemoactivity have the potential to identify clinical biomarkers of compound response, cellular pathways important to efficacy and potential toxicities; all vital to anticancer drug development. An investigation has been conducted that jointly explores tumor-cell constitutive NCI60 gene expression profiles and small-molecule NCI60 growth inhibition chemoactivity profiles, viewed from novel applications of self-organizing maps (SOMs) and pathway-centric analyses of gene expressions, to identify subsets of over- and under-expressed pathway genes that discriminate chemo-sensitive and chemo-insensitive tumor cell types. Linear Discriminant Analysis (LDA) is used to quantify the accuracy of discriminating genes to predict tumor cell chemoactivity. LDA results find 15% higher prediction accuracies, using ∼30% fewer genes, for pathway-derived discriminating genes when compared to genes derived using conventional gene expression-chemoactivity correlations. The proposed pathway-centric data mining procedure was used to derive discriminating genes for ten well-known compounds. Discriminating genes were further evaluated using gene set enrichment analysis (GSEA) to reveal a cellular genetic landscape, comprised of small numbers of key over and under expressed on- and off-target pathway genes, as important for a compound’s tumor cell chemoactivity. Literature-based validations are provided as support for chemo-important pathways derived from this procedure. Qualitatively similar results are found when using gene expression measurements derived from different microarray platforms. The data used in this analysis is available at http://pubchem.ncbi.nlm.nih.gov/and http://www.ncbi.nlm.nih.gov/projects/geo ({"type":"entrez-geo","attrs":{"text":"GPL96","term_id":"96"}}GPL96, {"type":"entrez-geo","attrs":{"text":"GSE32474","term_id":"32474"}}GSE32474).     

X-box binding protein 1 (XBP1): a potential role in chemotherapy response,clinical pathologic features,non-inflamed tumour microenvironment for breast cancer

X-box binding protein 1 (XBP1) is mainly expressed in breast cancer (BC) in human cancers. Its tumorigenesis and favourable prognosis are contradictory, and its essential role in chemotherapeutic response and immunosuppression is unknown in BC. The study firstly identified XBP1 who received neoadjuvant chemotherapy (NAC) from {"type":"entrez-geo","attrs":{"text":"GSE25055","term_id":"25055"}}GSE25055 and {"type":"entrez-geo","attrs":{"text":"GSE24460","term_id":"24460"}}GSE24460. Associations between XBP1 expression and clinicopathological characteristics was investigated using Oncomine, TCGA, UALCAN and bc-GenExMiner. The prognostic value of XBP1 was assessed using the Kaplan–Meier Plotter, bc-GenExMiner, {"type":"entrez-geo","attrs":{"text":"GSE25055","term_id":"25055"}}GSE25055, and {"type":"entrez-geo","attrs":{"text":"GSE25056","term_id":"25056"}}GSE25056. Furthermore, we systematically correlated XBP1 and immunological characteristics in the BC tumour microenvironment (TME) using TISIDB, TIMER, {"type":"entrez-geo","attrs":{"text":"GSE25055","term_id":"25055"}}GSE25055, {"type":"entrez-geo","attrs":{"text":"GSE25056","term_id":"25056"}}GSE25056 and TCGA dataset. Finally, an essential role of XBP1 in chemotherapy response was evaluated based on {"type":"entrez-geo","attrs":{"text":"GSE25055","term_id":"25055"}}GSE25055, {"type":"entrez-geo","attrs":{"text":"GSE25065","term_id":"25065"}}GSE25065, {"type":"entrez-geo","attrs":{"text":"GSE24460","term_id":"24460"}}GSE24460, {"type":"entrez-geo","attrs":{"text":"GSE5846","term_id":"5846"}}GSE5846, ROC Plotter and CELL databases. Furthermore, XBP1 mRNA expression levels were obviously highest in BC among human cancers and were significantly related to a good prognosis. In addition, XBP1 mRNA and protein levels were higher in the luminal subtype than in normal tissues and basal-like subtype, which might be attributed to membrane transport-related processes. Apart from BC, negative immunological correlations of XBP1 were not observed in other malignancies. XBP1 might shape the non-inflamed TME in BC. Finally, XBP1 expression was higher in chemo-resistive than chemo-sensitive cases, it had a predictive value and could independently predict chemotherapy response in BC patients receiving NAC. Our study suggests that the essential role of XBP1 in clinical pathologic features, non-inflamed TME, chemotherapy response in BC.     

Exploring the Novel Computational Drug Target and Associated Key Pathways of Oral Cancer

Oral cancer (OC) is a serious health concern that has a high fatality rate. The oral cavity has seven kinds of OC, including the lip, tongue, and floor of the mouth, as well as the buccal, hard palate, alveolar, retromolar trigone, and soft palate. The goal of this study is to look into new biomarkers and important pathways that might be used as diagnostic biomarkers and therapeutic candidates in OC. The publicly available repository the Gene Expression Omnibus (GEO) was to the source for the collection of OC-related datasets. {"type":"entrez-geo","attrs":{"text":"GSE74530","term_id":"74530"}}GSE74530, {"type":"entrez-geo","attrs":{"text":"GSE23558","term_id":"23558"}}GSE23558, and {"type":"entrez-geo","attrs":{"text":"GSE3524","term_id":"3524"}}GSE3524 microarray datasets were collected for analysis. Minimum cut-off criteria of |log fold-change (FC)| > 1 and adjusted p < 0.05 were applied to calculate the upregulated and downregulated differential expression genes (DEGs) from the three datasets. After that only common DEGs in all three datasets were collected to apply further analysis. Gene ontology (GO) and pathway analysis were implemented to explore the functional behaviors of DEGs. Then protein–protein interaction (PPI) networks were built to identify the most active genes, and a clustering algorithm was also implemented to identify complex parts of PPI. TF-miRNA networks were also constructed to study OC-associated DEGs in-depth. Finally, top gene performers from PPI networks were used to apply drug signature analysis. After applying filtration and cut-off criteria, 2508, 3377, and 670 DEGs were found for {"type":"entrez-geo","attrs":{"text":"GSE74530","term_id":"74530"}}GSE74530, {"type":"entrez-geo","attrs":{"text":"GSE23558","term_id":"23558"}}GSE23558, and {"type":"entrez-geo","attrs":{"text":"GSE3524","term_id":"3524"}}GSE3524 respectively, and 166 common DEGs were found in every dataset. The GO annotation remarks that most of the DEGs were associated with the terms of type I interferon signaling pathway. The pathways of KEGG reported that the common DEGs are related to the cell cycle and influenza A. The PPI network holds 88 nodes and 492 edges, and CDC6 had the highest number of connections. Four clusters were identified from the PPI. Drug signatures doxorubicin and resveratrol showed high significance according to the hub genes. We anticipate that our bioinformatics research will aid in the definition of OC pathophysiology and the development of new therapies for OC.     

Classifying Integrated Signature Molecules in Macrophages of Rheumatoid Arthritis,Osteoarthritis, and Periodontal Disease: An Omics-Based Study

Rheumatoid arthritis (RA), osteoarthritis (OA), and periodontal disease (PD) are chronic inflammatory diseases that are globally prevalent, and pose a public health concern. The search for a potential mechanism linking PD to RA and OA continues, as it could play a significant role in disease prevention and treatment. Recent studies have linked RA, OA, and PD to Porphyromonas gingivalis (PG), a periodontal bacterium, through a similar dysregulation in an inflammatory mechanism. This study aimed to identify potential gene signatures that could assist in early diagnosis as well as gain insight into the molecular mechanisms of these diseases. The expression data sets with the series IDs {"type":"entrez-geo","attrs":{"text":"GSE97779","term_id":"97779"}}GSE97779, {"type":"entrez-geo","attrs":{"text":"GSE123492","term_id":"123492"}}GSE123492, and {"type":"entrez-geo","attrs":{"text":"GSE24897","term_id":"24897"}}GSE24897 for macrophages of RA, OA synovium, and PG stimulated macrophages (PG-SM), respectively, were retrieved and screened for differentially expressed genes (DEGs). The 72 common DEGs among RA, OA, and PG-SM were further subjected to gene–gene correlation analysis. A GeneMANIA interaction network of the 47 highly correlated DEGs comprises 53 nodes and 271 edges. Network centrality analysis identified 15 hub genes, 6 of which are DEGs (API5, ATE1, CCNG1, EHD1, RIN2, and STK39). Additionally, two significantly up-regulated non-hub genes (IER3 and RGS16) showed interactions with hub genes. Functional enrichment analysis of the genes showed that “apoptotic regulation” and “inflammasomes” were among the major pathways. These eight genes can serve as important signatures/targets, and provide new insights into the molecular mechanism of PG-induced RA, OA, and PD.     

Expression of the Genetic Suppressor Element 24.2 (GSE24.2) Decreases DNA Damage and Oxidative Stress in X-Linked Dyskeratosis Congenita Cells

The predominant X-linked form of Dyskeratosis congenita results from mutations in DKC1, which encodes dyskerin, a protein required for ribosomal RNA modification that is also a component of the telomerase complex. We have previously found that expression of an internal fragment of dyskerin ({"type":"entrez-geo","attrs":{"text":"GSE24","term_id":"24"}}GSE24.2) rescues telomerase activity in X-linked dyskeratosis congenita (X-DC) patient cells. Here we have found that an increased basal and induced DNA damage response occurred in X-DC cells in comparison with normal cells. DNA damage that is also localized in telomeres results in increased heterochromatin formation and senescence. Expression of a cDNA coding for {"type":"entrez-geo","attrs":{"text":"GSE24","term_id":"24"}}GSE24.2 rescues both global and telomeric DNA damage. Furthermore, transfection of bacterial purified or a chemically synthesized {"type":"entrez-geo","attrs":{"text":"GSE24","term_id":"24"}}GSE24.2 peptide is able to rescue basal DNA damage in X-DC cells. We have also observed an increase in oxidative stress in X-DC cells and expression of {"type":"entrez-geo","attrs":{"text":"GSE24","term_id":"24"}}GSE24.2 was able to diminish it. Altogether our data indicated that supplying {"type":"entrez-geo","attrs":{"text":"GSE24","term_id":"24"}}GSE24.2, either from a cDNA vector or as a peptide reduces the pathogenic effects of Dkc1 mutations and suggests a novel therapeutic approach.     

Identification of potential biomarkers in dengue via integrated bioinformatic analysis

Dengue fever virus (DENV) is a global health threat that is becoming increasingly critical. However, the pathogenesis of dengue has not yet been fully elucidated. In this study, we employed bioinformatics analysis to identify potential biomarkers related to dengue fever and clarify their underlying mechanisms. The results showed that there were 668, 1901, and 8283 differentially expressed genes between the dengue-infected samples and normal samples in the {"type":"entrez-geo","attrs":{"text":"GSE28405","term_id":"28405"}}GSE28405, {"type":"entrez-geo","attrs":{"text":"GSE38246","term_id":"38246"}}GSE38246, and {"type":"entrez-geo","attrs":{"text":"GSE51808","term_id":"51808"}}GSE51808 datasets, respectively. Through overlapping, a total of 69 differentially expressed genes (DEGs) were identified, of which 51 were upregulated and 18 were downregulated. We identified twelve hub genes, including MX1, IFI44L, IFI44, IFI27, ISG15, STAT1, IFI35, OAS3, OAS2, OAS1, IFI6, and USP18. Except for IFI44 and STAT1, the others were statistically significant after validation. We predicted the related microRNAs (miRNAs) of these 12 target genes through the database miRTarBase, and finally obtained one important miRNA: has-mir-146a-5p. In addition, gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment were carried out, and a protein–protein interaction (PPI) network was constructed to gain insight into the actions of DEGs. In conclusion, our study displayed the effectiveness of bioinformatics analysis methods in screening potential pathogenic genes in dengue fever and their underlying mechanisms. Further, we successfully predicted IFI44L and IFI6, as potential biomarkers with DENV infection, providing promising targets for the treatment of dengue fever to a certain extent.     

Circulating lncRNA UCA1 and lncRNA PGM5-AS1 act as potential diagnostic biomarkers for early-stage colorectal cancer

Background: Colorectal cancer (CRC) is one of the most common and significant malignant diseases worldwide. In the present study, we evaluated two long non-coding RNAs (lncRNAs) in CRC patients as diagnostic markers for early-stage CRC.Methods: Using Gene Expression Omnibus (GEO) datasets {"type":"entrez-geo","attrs":{"text":"GSE102340","term_id":"102340"}}GSE102340, {"type":"entrez-geo","attrs":{"text":"GSE126092","term_id":"126092"}}GSE126092, {"type":"entrez-geo","attrs":{"text":"GSE109454","term_id":"109454"}}GSE109454 and {"type":"entrez-geo","attrs":{"text":"GSE115856","term_id":"115856"}}GSE115856, 14 differentially expressed lncRNAs were identified between cancer and adjacent tissues, among which, the two most differentially expressed were confirmed using quantitative real-time polymerase chain reaction (qRT-PCR) in 200 healthy controls and 188 CRC patients. A receiver operating characteristic (ROC) analysis was employed to evaluate the diagnostic accuracy for CRC.Results: From four GEO datasets, three up-regulated and eleven down-regulated lncRNAs were identified in CRC tissues, among which, lncRNA urothelial carcinoma-associated 1 (UCA1) and lncRNA phosphoglucomutase 5-antisense RNA 1 (PGM5-AS1) were the most significantly up- and down-regulated lncRNAs in CRC patient plasma, respectively. The area under the ROC curve was calculated to be 0.766, 0.754 and 0.798 for UCA1, PGM5-AS1 and the combination of these two lncRNAs, respectively. Moreover, the diagnostic potential of these two lncRNAs was even higher for the early stages of CRC. The combination of UCA1 and PGM5-AS1 enhanced the AUC to 0.832, and when the lncRNAs were used with carcinoembryonic antigen (CEA), the AUC was further improved to 0.874.Conclusion: In the present study, we identified two lncRNAs, UCA1 and PGM5-AS1, in CRC patients’ plasma, which have the potential to be used as diagnostic biomarkers of CRC.     

Doubly Optimized Calibrated Support Vector Machine (DOC-SVM): An Algorithm for Joint Optimization of Discrimination and Calibration

Historically, probabilistic models for decision support have focused on discrimination, e.g., minimizing the ranking error of predicted outcomes. Unfortunately, these models ignore another important aspect, calibration, which indicates the magnitude of correctness of model predictions. Using discrimination and calibration simultaneously can be helpful for many clinical decisions. We investigated tradeoffs between these goals, and developed a unified maximum-margin method to handle them jointly. Our approach called, Doubly Optimized Calibrated Support Vector Machine (DOC-SVM), concurrently optimizes two loss functions: the ridge regression loss and the hinge loss. Experiments using three breast cancer gene-expression datasets (i.e., {"type":"entrez-geo","attrs":{"text":"GSE2034","term_id":"2034"}}GSE2034, {"type":"entrez-geo","attrs":{"text":"GSE2990","term_id":"2990"}}GSE2990, and Chanrion''s datasets) showed that our model generated more calibrated outputs when compared to other state-of-the-art models like Support Vector Machine ( = 0.03,  = 0.13, and <0.001) and Logistic Regression ( = 0.006,  = 0.008, and <0.001). DOC-SVM also demonstrated better discrimination (i.e., higher AUCs) when compared to Support Vector Machine ( = 0.38,  = 0.29, and  = 0.047) and Logistic Regression ( = 0.38,  = 0.04, and <0.0001). DOC-SVM produced a model that was better calibrated without sacrificing discrimination, and hence may be helpful in clinical decision making.     

Up‐regulated SAMD9L modulated by TLR2 and HIF‐1α as a promising biomarker in tuberculosis

The aim of this study was to identify potential biomarkers of TB in blood and determine their function in Mtb‐infected macrophages. First of all, WGCNA was used to analyse 9451 genes with significant changes in TB patients’ whole blood. The 220 interferon‐γ‐related genes were identified, and then 30 key genes were screened using Cytoscape. Then, the AUC values of key genes were calculated to further narrow the gene range. Finally, we identified 9 genes from {"type":"entrez-geo","attrs":{"text":"GSE19444","term_id":"19444"}}GSE19444. ROC analysis showed that SAMD9L, among 9 genes, had a high diagnostic value (AUC = 0.925) and a differential diagnostic value (AUC>0.865). To further narrow down the range of DEGs, the top 10 hub‐connecting genes were screened from monocytes ({"type":"entrez-geo","attrs":{"text":"GSE19443","term_id":"19443"}}GSE19443). Finally, we obtained 4 genes (SAMD9L, GBP1, GBP5 and STAT1) by intersections of genes from monocytes and whole blood. Among them, it was found that the function of SAMD9L was unknown after data review, so this paper studied this gene. Our results showed that SAMD9L is up‐regulated and suppresses cell necrosis, and might be regulated by TLR2 and HIF‐1α during Mtb infection. In addition, miR‐181b‐5p is significantly up‐regulated in the peripheral blood plasma of tuberculosis patients, which has a high diagnostic value (AUC = 0.969).     

TGF-β-induced PLEK2 promotes metastasis and chemoresistance in oesophageal squamous cell carcinoma by regulating LCN2

Oesophageal squamous cell carcinoma (ESCC) has a relatively unfavourable prognosis due to metastasis and chemoresistance. Our previous research established a comprehensive ESCC database ({"type":"entrez-geo","attrs":{"text":"GSE53625","term_id":"53625"}}GSE53625). After analysing data from TCGA database and {"type":"entrez-geo","attrs":{"text":"GSE53625","term_id":"53625"}}GSE53625, we found that PLEK2 predicted poor prognosis in ESCC. Moreover, PLEK2 expression was also related to the overall survival of ESCC patients undergoing chemotherapy. Repression of PLEK2 decreased the proliferation, migration, invasion and chemoresistance of ESCC cells in vitro and decreased tumorigenicity and distant metastasis in vivo. Mechanistically, luciferase reporter assay and chromatin immunoprecipitation assay suggested that TGF-β stimulated the process that Smad2/3 binds to the promoter sequences of PLEK2 and induced its expression. RNA-seq suggested LCN2 might a key molecular regulated by PLEK2. LCN2 overexpression in PLEK2 knockdown ESCC cells reversed the effects of decreased migration and invasion. In addition, TGF-β induced the expression of LCN2, but the effect disappeared when PLEK2 was knockdown. Moreover, AKT was phosphorylated in all regulatory processes. This study detected the major role of PLEK2 in driving metastasis and chemoresistance in ESCC by regulating LCN2, which indicates the potential use of PLEK2 as a biomarker to predict prognosis and as a therapeutic target for ESCC.Subject terms: Cancer, Molecular biology     

Modeling Oncogenic Signaling in Colon Tumors by Multidirectional Analyses of Microarray Data Directed for Maximization of Analytical Reliability

Background

Clinical progression of colorectal cancers (CRC) may occur in parallel with distinctive signaling alterations. We designed multidirectional analyses integrating microarray-based data with biostatistics and bioinformatics to elucidate the signaling and metabolic alterations underlying CRC development in the adenoma-carcinoma sequence.

Methodology/Principal Findings

Studies were performed on normal mucosa, adenoma, and carcinoma samples obtained during surgery or colonoscopy. Collections of cryostat sections prepared from the tissue samples were evaluated by a pathologist to control the relative cell type content. The measurements were done using Affymetrix GeneChip HG-U133plus2, and probe set data was generated using two normalization algorithms: MAS5.0 and GCRMA with least-variant set (LVS). The data was evaluated using pair-wise comparisons and data decomposition into singular value decomposition (SVD) modes. The method selected for the functional analysis used the Kolmogorov-Smirnov test. Expressional profiles obtained in 105 samples of whole tissue sections were used to establish oncogenic signaling alterations in progression of CRC, while those representing 40 microdissected specimens were used to select differences in KEGG pathways between epithelium and mucosa. Based on a consensus of the results obtained by two normalization algorithms, and two probe set sorting criteria, we identified 14 and 17 KEGG signaling and metabolic pathways that are significantly altered between normal and tumor samples and between benign and malignant tumors, respectively. Several of them were also selected from the raw microarray data of 2 recently published studies ({"type":"entrez-geo","attrs":{"text":"GSE4183","term_id":"4183"}}GSE4183 and {"type":"entrez-geo","attrs":{"text":"GSE8671","term_id":"8671"}}GSE8671).

Conclusion/Significance

Although the proposed strategy is computationally complex and labor–intensive, it may reduce the number of false results.     

Development of a microarray for two rice subspecies: characterization and validation of gene expression in rice tissues

Background

Rice is one of the major crop species in the world helping to sustain approximately half of the global population’s diet especially in Asia. However, due to the impact of extreme climate change and global warming, rice crop production and yields may be adversely affected resulting in a world food crisis. Researchers have been keen to understand the effects of drought, temperature and other environmental stress factors on rice plant growth and development. Gene expression microarray technology represents a key strategy for the identification of genes and their associated expression patterns in response to stress. Here, we report on the development of the rice OneArray® microarray platform which is suitable for two major rice subspecies, japonica and indica.

Results

The rice OneArray® 60-mer, oligonucleotide microarray consists of a total of 21,179 probes covering 20,806 genes of japonica and 13,683 genes of indica. Through a validation study, total RNA isolated from rice shoots and roots were used for comparison of gene expression profiles via microarray examination. The results were submitted to NCBI’s Gene Expression Omnibus (GEO). Data can be found under the GEO accession number GSE50844 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE50844). A list of significantly differentially expressed genes was generated; 438 shoot-specific genes were identified among 3,138 up-regulated genes, and 463 root-specific genes were found among 3,845 down-regulated genes. GO enrichment analysis demonstrates these results are in agreement with the known physiological processes of the different organs/tissues. Furthermore, qRT-PCR validation was performed on 66 genes, and found to significantly correlate with the microarray results (R?=?0.95, p?<?0.001***).

Conclusion

The rice OneArray® 22 K microarray, the first rice microarray, covering both japonica and indica subspecies was designed and validated in a comprehensive study of gene expression in rice tissues. The rice OneArray® microarray platform revealed high specificity and sensitivity. Additional information for the rice OneArray® microarray can be found at http://www.phalanx.com.tw/index.php.     

Immune implication of an autophagy-related prognostic signature in uveal melanoma

Uveal Melanoma (UM) is a rare cancer deriving from melanocytes within the uvea. It has a high rate of metastasis, especially to the liver, and a poor prognosis thereafter. Autophagy, an intracellular programmed digestive process, has been associated with the development and progression of cancers, with controversial pro- and anti-tumour roles. Although previous studies have been conducted on autophagy-related genes (ARGs) in various cancer types, its role in UM requires a deeper understanding for improved diagnosis and development of novel therapeutics. In the present study, Zheng et al. used univariate Cox regression followed by least absolute shrinkage and selection operator (Lasso) regression to identify a robust 9-ARG signature prognostic of survival in a total of 230 patients with UM. The authors used the Cancer Genome Atlas (TCGA) UM cohort as a training cohort (n=80) to identify the signature and validated it in another four independent cohorts of 150 UM patients from the Gene Expression Omnibus (GEO) repository ({"type":"entrez-geo","attrs":{"text":"GSE22138","term_id":"22138"}}GSE22138, {"type":"entrez-geo","attrs":{"text":"GSE27831","term_id":"27831"}}GSE27831, {"type":"entrez-geo","attrs":{"text":"GSE44295","term_id":"44295"}}GSE44295 and {"type":"entrez-geo","attrs":{"text":"GSE84976","term_id":"84976"}}GSE84976). This 9-ARG signature was also significantly associated with the enrichment of cancer hallmarks, including angiogenesis, IL6-KJAK-STAT3 signalling, reactive oxygen species pathway and oxidative phosphorylation. More importantly, this signature is associated with immune-related functional pathways and immune cell infiltration. Thus, this 9-ARG signature predicts prognosis and provides deeper insights into the immune mechanisms in UM, with potential implications for future immunotherapy.     

A 16-Gene Signature Distinguishes Anaplastic Astrocytoma from Glioblastoma

Anaplastic astrocytoma (AA; Grade III) and glioblastoma (GBM; Grade IV) are diffusely infiltrating tumors and are called malignant astrocytomas. The treatment regimen and prognosis are distinctly different between anaplastic astrocytoma and glioblastoma patients. Although histopathology based current grading system is well accepted and largely reproducible, intratumoral histologic variations often lead to difficulties in classification of malignant astrocytoma samples. In order to obtain a more robust molecular classifier, we analysed RT-qPCR expression data of 175 differentially regulated genes across astrocytoma using Prediction Analysis of Microarrays (PAM) and found the most discriminatory 16-gene expression signature for the classification of anaplastic astrocytoma and glioblastoma. The 16-gene signature obtained in the training set was validated in the test set with diagnostic accuracy of 89%. Additionally, validation of the 16-gene signature in multiple independent cohorts revealed that the signature predicted anaplastic astrocytoma and glioblastoma samples with accuracy rates of 99%, 88%, and 92% in TCGA, {"type":"entrez-geo","attrs":{"text":"GSE1993","term_id":"1993"}}GSE1993 and {"type":"entrez-geo","attrs":{"text":"GSE4422","term_id":"4422"}}GSE4422 datasets, respectively. The protein-protein interaction network and pathway analysis suggested that the 16-genes of the signature identified epithelial-mesenchymal transition (EMT) pathway as the most differentially regulated pathway in glioblastoma compared to anaplastic astrocytoma. In addition to identifying 16 gene classification signature, we also demonstrated that genes involved in epithelial-mesenchymal transition may play an important role in distinguishing glioblastoma from anaplastic astrocytoma.