What are the unique features of your database?

Our circExp database focuses on circRNA expression profiles in human cancer. You may explore global expression profiles of cancer-related circRNAs. We implemented a computational pipeline to annotate different probes or NGS reads of circRNAs so that you may compare their expressions in different datasets. s user-friendly web interface provides expression heatmaps and Excel-like data filtering to extract circRNAs of interest. Finally, you may download all processed data for further integrative analyses.

How is the data collected?

In detail, we found 33 microarray-based datasets, 24 from the Agilent-069978 Arraystar Human CircRNA microarray version 1 and 9 from version 2. We downloaded the array-based information from the GEO database, and when we mapped those circRNAs back to the circRNA database IDs based on their genomic locations, we found that the total probe numbers between the Arraystar Human CircRNA microarray versions 1 and 2 differed greatly. So, we designed our data processing procedure to prefer the Arraystar Human CircRNA microarray version 2, rather than those annotations from version 1. We used circBase as a reference database to annotate circRNAs consistently and then unified all probes from the Arraystar Human CircRNA microarray versions 1 and 2 into circBase IDs. An additional 14 datasets generated on Illumina high-throughput, next-generation sequencing platforms each consisted of short reads, so we mainly used the genome to map to those annotated circRNAs from circBase.

To explore human circRNA expression changes in cancer, we conducted extensive keyword-based searches of the public Gene Expression Omnibus (GEO) from NCBI. By simply using "circRNA" or "circular RNA", we found 1,103 entries, 197 of which were human datasets. We further narrowed the list to those datasets with a minimum of 4 samples. Finally, we manually read the dataset descriptions to identify cancer-related studies, subsequently collecting 48 GEO datasets.

How to search the circExp database?

We constructed three text-based query interfaces that enable the user to find circRNA-associated information, experimental design, and parental protein-coding genes information. For example, searching for "exosome" will return all exosome-related datasets that match the experimental description in our database.

How to browse the circExp database?

CircExp allows users to browse circRNA expression data based on the curated cancer types and platforms (microarray-based or RNAseq-based expression profiles).

How to access the annotation in circExp database?

A typical dataset contains 2 information categories: a circRNA annotation page and an expression profile page. On the circRNA annotation page (right chart A), users can explore all annotated information, including circRNA IDs in the circBase database, genomic location, strand, and the parental genes and their associated information. Users may load a large probe table by using the jQuery-based DataTables plug-in, which can sort, page, and filter to plain HTML tables with a few Excel-like clicks. Subsequently, the resulting information can be exported in Excel, comma-separated values (CSV), and PDF formats for further data manipulation, and the handy clipboard-based copying and printing functions ensure easy data saving on local computers.

On the expression profile page (left chat B), probe- and sample-based heatmaps change dynamically as the user scrolls down the long list of probes, thus providing a quick overview of the different expression patterns among various samples. Additionally, dataset information and links to the GEO source datasets are provided. Differential expression analysis results are provided on the download page.

How often will the database be updated?

As we all know, the cancer genomics field moves very fast. We will search the NCBI GEO database and extract new human circRNA expression profiles every six months. In the future, our collection will not only include cancer, but also more diseases.

Is the data free?

All circExp research outputs are stored under Creative Commons Licenses. Although our data is publicly available, we still retain ownership of our research objects under a CC-BY license. These data are distributed in the hope that they will be useful, but without any warranty of MERCHANTABILITY for a particular purpose.

Can we suggest new datasets?

We will gratefully accept user-provided new datasets related to circRNA expression, and not just in human cancers. Please email us at the address in our publication.

How do I cite this resource?

Please cite as: Min Zhao, Yining Liu, Hong Qu. circExp, the online based platform for human circRNA expression in cancers.

How to conduct meta-analysis based on your resource?

We suggest you start browsing the data to find multiple datasets relevant to your topic. For example, we have five datasets related to colorectal cancer. Those are a good starting point for a comparative analysis to identify consistent up/down-regulated circRNAs across multiple datasets.