GENT2: an updated gene expression database for normal and tumor tissues.

Park SJ, Yoon BH, Kim SK, Kim SY. · 2019

Read GENT2: an updated gene expression database for normal and tumor tissues. by Park SJ, Yoon BH, Kim SK, Kim SY. (2019) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc6624177.txt.

Introduction

Park and co-authors present GENT2, a gene expression database for normal and tumor tissues. Their article appears in BMC Medical Genomics. The first GENT helped researchers read expression patterns across tissues. It drew on public data sets from one microarray family. Growing data stores called for a full update.

GENT2 now holds more than 68,000 samples, double the first version. Coverage spans 72 tissues, up from 57. New tools profile expression by tumor subtype. Survival views show how a gene's level tracks patient outcome. A meta-analysis blends survival data across studies for firmer estimates. The rebuild uses Lucene indexing for fast search. The database is free online for any researcher.

Catalog record

CREATORPark SJ, Yoon BH, Kim SK, Kim SY.
YEAR2019
COLLECTIONPAPERS
FORMATTXT
FILE SIZE25 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC6624177

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

GENT2: an updated gene expression database for normal and tumor tissues By Seung-Jin Park, Byoung-Ha Yoon, Seon-Kyu Kim, Seon-Young Kim Published in: BMC Medical Genomics DOI: 10.1186/s12920-019-0514-7 ABSTRACT Abstract Background Gene Expression database of Normal and Tumor tissues 2 (GENT2) is an updated version of GENT, which has provided a user-friendly search platform for gene expression patterns across different normal and tumor tissues compiled from public gene expression data sets. Results We refactored GENT2 with recent technologies such as Apache Lucene indexing for fast search and Google Web Toolkit (GWT) framework for a user-friendly web interface. Now, GENT2 contains more than 68,000 samples and has several new useful functions. First, GENT2 now provides gene expression across 72 different tissues compared to 57 in GENT. Second, with increasing importance of tumor subtypes, GENT2 provides an option to study the differential expression and its prognostic significance based on tumor subtypes. Third, whenever available, GENT2 provides prognostic information of a gene of interest. Fourth, GENT2 provides a meta-analysis of survival information to provide users more reliable prognostic value of a gene of interest. Conclusions In conclusion, with these significant improvements, GENT2 will continue to be a useful tool to a wide range of researchers. GENT2 is freely available at http://gent2.appex.kr. Electronic supplementary material The online version of this article (10.1186/s12920-019-0514-7) contains supplementary material, which is available to authorized users. Keywords: Tissue and cell line wide gene expression profiling, Cancer subtype profiling, Survival analysis, Meta-survival analysis, Large scale microarray web-database BACKGROUND The explosion of publicly available omics datasets including gene expression microarray and next-generation sequencing has provided valuable resources for researchers. However, for most researchers with few bioinformatics skills, public datasets in a raw data format are of little value. To help researchers to utilize public gene expression datasets, we previously developed a web-based gene expression database named Gene Expression database of Normal and Tumor tissues (GENT) [1]. GENT provides gene expression profiles across diverse human cancer and normal tissues with more than 34,000 samples generated by the Affymetrix U133A or U133Plus2 microarray platform with consistent data processing. To our satisfaction, since its launch at 2011, GENT has been widely used by researchers across the world, with over 1600 visitors per month on average, and the method has been cited more than 100 times in research work [2–6]. In our opinion, a few strengths have contributed to the success of GENT. First, it is very simple to use. A user inputs a gene symbol of interest and gets back results within a few seconds. Second, the large amounts of samples (~ 34,000) generated using the same platform with consistent data processing makes the database reliable by removing or reducing many biases (i.e., different batches, laboratories, ethnicities, etc.) caused by a small sample size. Third, GENT provides intuitive graphics as well as raw gene expression values to users, so users can utilize the information in convenient ways. Enormous amounts of publicly available gene expression datasets have accumulated since GENT was first published in 2011, and the need was recognized to upgrade the GENT system both quantitatively and qualitatively. Here, GENT2 is provided as an updated version of GENT with more samples and newer functionalities. Major improvements of GENT2 over GENT are the following. First, GENT2 contains more than 68,000 samples compared to 34,000 samples in GENT. Second, GENT2 provides gene expression information on 72 paired tissues compared to 57 tissues of GENT. Third, GENT2 provides results of statistical tests (t-test, log2 fold changes, etc.) as well as boxplot summaries for user convenience. Fourth, considering the importance of tumor subtypes, GENT2 provides results of various cancer subtype profiling analysis (e.g., stage or grade with various molecular subtypes), Fifth, it provides results of prognostic value estimation using Kaplan-Meier plots. Finally, GENT2 provides results of meta-survival analysis combining various independent studies for reliable prognostic information. With these significant improvements, GENT2 is suggested as a versatile tool to assess the biological and prognostic relevance of a specific gene in various cancers. CONSTRUCTION AND CONTENT To construct the GENT2 system, gene expression data is downloaded from the NCBI GEO database generated by two microarray platforms (Affymetrix U133A or U133Plus2). Then, the data were stratified into cancer and normal samples across 72 tissues. The GENT2 database provides the following five functions: 1) a landscape of gene expression profile across 72 normal and tumor tissues, 2) cancer subtype profiling, 3) statistical significance of gene expression difference between normal and tumor samples, 4) a prognostic value of gene expression, and 5) meta-survival analysis (Fig. 1). The following sections present the workflows of how the database was constructed. [Figure: Overview of the GENT2 database system. We collected data from the GEO public repository using only two platforms, U133Plus2 (GPL570) and U133A (GPL96). Next, we classified the data into cancer and normal tissue. The functions of the GENT2 database are listed on the right] Collection and processing of publicly available gene expression datasets We collected data from the NCBI GEO public database and processed them with the MAS5 algorithm using the affy package in Bioconductor [7]. We used the same MAS5 algorithm used in the original GENT, as it is a single-array algorithm, which allows the comparison of multiple independent datasets [1]. GENT2 implementation The GENT2 system consists of the following two main parts: indexed DB and web-based user interface. To construct a database that can handle a huge gene expression profile (over 68,000 cancer patients) efficiently, we applied the Apache Lucene index machine (ver. 7.4.0) into whole gene expression datasets. A web-based graphical user interface (GUI) was implemented with the Google Web Toolkit (GWT, ver. 2.7.0) and GWT extended (GXT, ver. 4.0.0) frameworks based on JAVA language. Data exchange between a web browser and the GENT2 server is controlled by a GWT remote procedure call. All statistical tests provided from the GENT2, such as two-sample t-test, Kaplan-Meier with log-rank test, and meta-survival analysis with Cox proportional hazard modes, were implemented using R (ver. 3.2.5) with Bioconductor plugins (ver. 3.6). The system architecture of GENT2 is illustrated in Additional file 3: Figure S1. Database description GENT2 contains more than 44,000 (U133Plus2; 887 datasets) and 23,000 (U133A; 358 datasets) samples (Table 1, Additional file 1: Table S1, Additional file 2: Table S2). A larger number of samples were added to the original GENT, and over 380 datasets were added for the U133Plus2 platform. The top three tissues of each platform are as follows: blood, brain and breast for U133Plus2 and blood, bone marrow and brain for U133A. Additionally, subtype analysis and meta-analysis data in GENT2 are displayed in Tables 2 and 3, respectively. [Table: GENT2 database description] [Table: Subtypes of each cancer type] [Table: Datasets used with meta survival analysis in GENT2] UTILITY Tissue- and cell line-wide expression profile comparing normal and tumor samples When a user puts a gene symbol in the ‘Search Gene profile’ page, GENT2 performs tissue-wide expression profiling and statistical testing quickly (less than 5 s) on two Affymetrix platforms, simultaneously. ‘Search Gene profile’ provides two boxplots of gene expression from the Affymetrix U133Plus2 and the Affymetrix U133A platform, allowing users to compare both results. In Fig. 2a, for example, we illustrate the tissue-wide gene expression profile of an oncogene, ERBB2. Compared

Read the complete text