Does the website save my expression data?
No. Nor do we store your results. The only usage data we collect is through Google Analytics, which collects the time and duration of access, geolocation down to city level, and operating system information and so on. Google Analytics also seems to be trying to guess user demographics. We do not have access to user’s IP addresses.
“ID not recognized.”
The iDEP website cannot recognize your gene IDs. Please try to use common gene IDs like Entrez gene ID, Ensembl gene ID, gene symbols, and so on. iDEP’s gene ID conversion is based on the information available at Ensembl. Alternatively, try to convert your data into such IDs.
Occasionally, this can happen if users upload their data file before the app fully initializes. Users need to wait until the app is ready after it shows “Loading R packages … … Done. Ready to load data files.”
My species is not covered in the list, can I still use the site?
Yes. You can still do exploratory data analysis and differential gene expression. To do enrichment analysis and pathway analysis, you can choose “NEW SPECIES” in the species drop down and then upload a gene set file in the GMT format. Note that gene IDs should match those used in the data file.
Should I use DESeq2, limma-voom, or limma-trend?
The differences we observed by switching among these methods are small. So it really does not matter that much. Your choices of pre-process parameters probably have a bigger effect. Also, make sure you are setting up the correct model.
The website does not give me differentially expressed genes, and it said: “Failed to parse sample names to define groups.”
You either do not have replicates or the replicates are not clearly marked as “_Rep1”, “_Rep2” at the end of sample names. Please change column names of your data file.
Help! I got an error message.
We would appreciate it if you could send us an email regarding that error with as much information as possible, include the first few lines of your data file. That will help us improve the software and serve our users better.
“No significant pathway found.”
Using the current choices of options, iDEP does not find significant pathways. You can try to lower the FDR cutoff or change the gene set database.
I cannot download files from iDEP! (Web Browsers)
This could be caused by your browser or extensions and plug-ins. Try using a different browser. We know that the Data Saver extension on Chrome causes problems for file download and should be disabled when using iDEP. About 67% of our users use Google Chrome. This is followed by Safari (15%), Firefox (11%), Internet Explorer (3%). We noticed that Safari has a much shorter session timeout period. We recommend Google Chrome.
My results stay the same when I switch databases while using ReactomePA.
When you choose ReactomePA, the Bioconductor package uses its own gene set database based on Reactome.
Can I install a local version of iDEP?
Researchers at non-profit organizations can install a local copy. The source code is available at GitHub. Researchers at NCBI has contributed some scripts for install iDEP and ShinyGO as Singularity containers. A copy of the entire supporting database is available at Zenodo. But we update the annotation frequently.
How do I cite iDEP?
Just include the URL is not enough. Please do cite our paper: Ge, Son, and Yao: iDEP: an integrated web application for differential expression and pathway analysis of RNA-Seq data, BMC Bioinformatics 19:534, 2018.
I just want to identify DEGs and pathways, why do I need hierarchical clustering and K-means clustering, PCA and so many other functionalities in iDEP?
iDEP encourages researchers to visualize and interact with their data. As I explained in another paper, the power of exploratory data analysis (EDA) cannot be overstated.
Get to know your data. Play with your data. Then you can ask the question you wanted to ask.
The iDEP server is slow!
One thing you can try is to close the browser window. Not just the tab, the entire browser window and start a new Google Chrome window. Or any other browser. This is because all computational tasks from one browsing session are assigned to a CPU core. If you start a new window, you can move to another virtual server.
I didn’t find significant genes in DEG analysis.
We can adjust the fold-change and FDR cutoffs. The default for fold-change is 2, which could be lowered. The FDR cutoff could be increased to as high as 0.25 or more, as it is the proportion of DE genes that are false.
Some stimuli induce profound changes in the expression profiles (thousands of genes changed with big foldchanges), while others only introduce subtle changes. It is possible that little change occurred, or there is small changes that could not be detected with our small sample sizes.