Compendia Bioscience announces the addition of 23 highly curated data sets — representing over a thousand microarray experiments — to the Oncomine database. Because sample data is critical to understanding gene expression data, we work very hard to ensure that all available sample data is captured in Oncomine. As a result of these efforts nearly half of the new studies in this release contain sample data that is not available by direct download from public repositories — but is available in Oncomine.
Compendia Bioscience is proud to announce the newest release of the Oncomine database with the following features:
• New Oncomine Data: 11 tissue types, 23 studies, 1250 microarrays. more >
• All Oncomine Data: 39 cancer types, 311 studies, 21,949 microarrays. more >
• Oncomine Curation: Nearly half of the new studies contain sample data that is not available by direct download from public repositories. list of studies added in this release list of all studies available in Oncomine
New Data Permits Meta-Analysis of Esophageal Cancer Progression
New Oncomine Studies Permit Meta-Analysis of Esophageal Cancer Progression
• To identify biomarkers or target genes across multiple independent studies, comparable analyses are required.
• Oncomine now contains 3 studies that can be analyzed together via Meta-analysis to make new discoveries in Esophageal Cancer.
• The figure below shows one result: COL4A2 expression is significantly correlated with progression from Normal Esophagus (blue) to Barretts Esophagus (red) to Esophageal Carcinoma (green) in each of the three studies.
Meta-Analysis of Esophageal Cancer Progression in Oncomine
The incidence of esophageal malignancy has increased more than 3-fold over the last several decades, exceeding that of all other types of cancer1,2. Esophageal adenocarcinoma is associated with an increase in gastroesophageal reflux disease (GERD), and with Barretts Esophagus, a transformation of normal squamous epithelium to columnar epithelium. Patients with Barretts Esophagus have a 40- to 125-fold increased risk of developing adenocarcinoma; those patients who present with advanced disease face a 5-year survival rate of 25 percent1,2.
Although molecular studies have identified alterations associated with the progression of Barretts esophagus to invasive adenocarcinoma, the most reliable predictors of cancer progression remain the histological measures of high-grade dysplasia. This may be about to change, however, as a number of labs have recently completed high-throughput studies on samples that span the range of esophageal cancer progression, as indicated in the table below.
While each of these studies makes an important, individual contribution to esophageal cancer research, the studies have additional research utility when considered together. In particular, the availability of three studies with the same sample types offers the opportunity to analyze the studies jointly, using meta-analysis, to produce a list of genes that are most significant across studies.
Meta-analysis results are easily generated in Oncomine (see How to...), and visualized as heat maps or as gene lists; the latter is shown in Figure 1. The most significant result is the first gene on the list, COL4A2. Selecting the Box Plot for that gene shows that in each of the three studies expression of COL4A2 increases as the population changes from Normal Esophagus (blue), to Barretts Esophagus (red), to Esophageal Adenocarcinoma (green) (Figure 2). To drill even deeper into the study data, bar charts can be accessed that show expression levels for each individual sample (Figure 3).
COL4A2 is just one of several hundred genes that show statistically significant over-expression in the progression of normal esophagus to esophageal adenocarcinoma in at least two of these three studies. Oncomine provides a variety of filters to narrow that list to genes with specific attributes, such as known cancer genes or known therapeutic targets. Alternatively, to expand the list of potential target genes, a similar analysis could be conducted to derive the list of genes with significant under-expression across the same samples.
The clear advantage of using meta-analysis to compare results across studies is that the genes identified by this method are highly likely to represent real, reproducible results. Another advantage is that meta-analysis does not require the generation of new data to generate new knowledge, but instead relies on the strategic re-utilization of existing data. It is a simple idea, but can only be put into practice when multiple comparable datasets are available. Oncomine currently contains data from 310 studies, providing many opportunities to compare and validate results using meta-analysis.
1Williams et al., Current Oncology 13 (1): 33-43, 2006 2Kimchi et al., Cancer Research 65 (8): 3146-3154, 2005 3Hao et al., Gastroenterology 131: 925-933, 2006 4Wang et al., Oncogenomics 25: 3346-3356, 2006
Q-VALUE:
A q-value is used in Oncomine to account for false positives resulting from large numbers of t-tests. A lower q-value indicates a more significant result for a gene.
In order to visualize data across multiple independant studies in Oncomine, data is scaled by z-score median normalization.
Using Meta-Analysis in Oncomine
To repeat the above analysis, take the following steps:
Class 1: Normal Esophagus
Class 2: Barretts Esophagus
Class 3: Esophageal Adenocarcinoma
Click the Advanced Analysis button in the bottom right corner.
On the next page, click View Gene List near the bottom of the page. This will generate the image shown in Figure 1.
Click on the View button next to the first gene, COL4A2. This will generate the image shown in Figure 2.
Click on one of the last three bars in the bar chart. This will generate the image shown in Figure 3.
To filter results by known therapeutic targets, return to Step 4. Before clicking View Gene List, click the radio button next to Filter, and select Therapeutic Target from the drop-down menu.
To retrieve genes that are under-expressed in esophageal cancer progression, return to Step 4. Before clicking View Gene List, click the radio button next to Down.
Meta-Analysis of Esophageal Cancer Progression in Oncomine 
