Thanks for this video! V useful :) Please would you mind letting me know if you paste both upregulated and downregulated genes OR do you paste the unregulated and downregulated genes separately (like in ToppGene)?
This was an extremely useful video. Thank you sooo much. Most of the time you have to teach all these things for yourself as a PhD candidate. One also usually do not have people to ask how to exactly do or use these things.
This is GOLD! I am a biologist who does not know anything about bioinformatics, i've done a differential expression analysis of genes using RStudio and then i was like, what now, what to do with these DEGs? And then i found this amazing tutorial! THANK YOU!
Very helpful - I have dabbled with DAVID from occasionally but this is the first time I really understood how to move beyond the initial set of results it gives.
It was a great tutorial. It would be nice if you could expand on the meaning of each part and their function in another video. For example, maybe explain how the acquired data can be used in research (or at least provide resources for beginners to follow and learn more). Also, if there are technical or specific types of plots that can visualize the results taken from DAVID, it would be very nice if you can cover them as well.
If you are using chrome or edge: hover your mouse over the information in the new window, right click to get a menu, and "save as" is one of the options - this saves the txt file. Or: click and drag to select all the info, then copy-paste into word using the "use destination files" formatting option, save as txt, then import into excel and it will give it to you in columns.
Since, after many years of literal silence from the developers, DAVID was only recently updated almost 2 years ago in 2021, how does it compare with the other alternatives that have sprung up since then including GSEA from the broad institute and FGSEA as well as the ClusterProfiler R package?
Thank you so much for such a helpful video. I am working on proteomics analysis and I have done analysis between treatment and control group samples using Mass Spec and I have analysed the data using Metaboanalyst. Now, I have lists of significantly upregulated and downregulated proteins. Now, I want to do GO and KEGG pathway analyses using DAVID. Can you please guide me on this? I have DEPs and not DEGs. Thank you in advance.
Hi, when you input a list into DAVID, you need to select what format your identifiers are in from the list in the box. There are several protein identifiers eg protein GI accession, UNIPROT ID, REFSEQ protein, so one of these probably matches what you have. if not, you could use Ensembl biomart to convert your IDs - paste your list into the relevant box in the "filter" section and select what your protein IDs are, then in the "attributes" section, choose what additional identifiers you want to be added to each row in your list www.ensembl.org/biomart/martview/0aa6277b4a0a97691488e57e813aca5f
Thank you for this useful tutorial please can you tell me how to prepare the gene list? I have a list of primers sequence, can i detect which gene isomer it is ?
You can download the DAVID results, and that gives you a file of the data that appears on the screen. It depends what you want to plot, but personally I like to do a horizontal bar chart, where each bar is a GO term. The x axis is -log10p or -log10benjamini etc. You can use excel to calculate this. That way, the lower the p or FDR value, the longer the bar. You can then use the label points function in excel to add the number of your genes in each GO category if you want. You will need to choose your cutoff for how many GO terms to show. You can use FDR of
Would you please cover how genes and protein interaction can be constructed and potential markers can be selected from GSE( Differential expressed genes) data using Both GSEA and cytoscape apps,
After analyzing cell lines for cancer patients we extract the GO (Gene Ontology), we discover that some genes in our data are related to each other that for example, share GO or share a biological pathway from KEGG, How that be useful in Cancer Research, I mean, what will be the next step? How that will help us to discover which set of genes when they mutated caused cancer? Thanks in advance
You need to filter the gene sets first. I often use p < 0.05 and FDR < 0.25, but it depends how many you want to work with. A useful way to plot them is to make a horizontal bar chart, one bar for each gene set, the x axis is NES, log10p or log10FDR, y axis cross at zero, and you make the value negative for genesets that are downregulated and positive for those that are upregulated.
Thanks for sharing this amazing tutorial. Could you please tell me how to create a sublist from the functional annotation clustering? Based on DAVID's guideline, I checked the boxes that I interested in, then clicked on the "Create Sublist" button, but nothing showed up. Do you know where David save the sublists into?
Go back to the page with "annotation summary results" on it, and look on the left in the list manager. Below the original list of genes you uploaded, you will see the name of your sublist. You can use this sublist in further analyses, or use the "show gene list" option to see the genes in that sublist (although I had to remove the original list before it would show me the genes in the sublist). Good luck!
Can I get a much clearer picture by using a threshold? Could you use a list of the top 100 upregulated genes (or a p value)? Do I need to include all of the upregulated genes without setting any threshold? Are there any limits on the number of genes required? Can I get a meaningful result if I have 100 genes on my list?
You need to set a cutoff eg adjusted p value less than 0.05. If you have thousands of genes that meet that threshhold, then making it more strigent will focus your analysis on those that a changed more significantly or to a larger extent.
Thanks for the great tutorial. I'd be interested to know your thoughts on which background to use. Is there any benefit in using a background of all genes of the species or a background of all genes expressed (not necessarily differentially) in the tissues under investigation?
Hi Hilary, if the list of genes you are analysing is already based on differential expression between two types of sample, you should use the background of all genes so that it can identify as many of your genes as possible. However, if you only have data from a single condition and want to compare against the background of normal gene expression in that tissue, maybe the second option is better. However, I haven't tested that, and I am sceptical about using someone else's control data......
Thank you for the tutorial. Please I am really interested in how to draw the bar charts to show the BP, CC and MF enrichments as well as the dot graph/chart to show the KEGG pathway as shown in published papers. Do you have an idea how to achieve that after obtaining those values, thank you.
You can download the DAVID results, and that gives you a file of the data that appears on the screen. It depends what you want to plot, but personally I like to do a horizontal bar chart, where each bar is a GO term. The x axis is -log10p or -log10benjamini. You can use excel to calculate this. That way, the lower the p or FDR value, the longer the bar. You can then use the label points function in excel to add the number of your genes in each GO category if you want. You will need to choose your cutoff for how many GO terms to show. You can use FDR of
@@GenomicsGurus Thank you so much for the reply. I was actually trying to attach pictures of GO and KEGG images to better explain what I mean, but the attachment is not possible here. Please examples include the KEGG image shown in Su et al 2020. Transcriptome-wide map of m 6 A circRNAs identified in a rat model of hypoxia mediated pulmonary hypertension. BMC Genomics.2020 Jan 13;21(1):39. doi: 10.1186/s12864-020-6462-y. And possibly the GO image shown in Zhang et al 2020.Circular RNA expression profile and m6A modification analysis in poorly differentiated adenocarcinoma of the stomach. EPIGENOMICSVOL. 12, NO. 12. doi.org/10.2217/epi-2019-0153 I would be glad if I will be able to get high quality GO and KEGG figures to use for my publication too. Thank you.
Thank you for the tutorial. I think it would be so useful for my analysis. Unfortunately, I tried to input my Uniprot sequences but they were not recognised by David. Is this because the proteins are not in the David database? I submitted 13 rice proteins (with Uniprot ID) and only 1 was recognised. I tried TOPPgene too, and just couldn't get anything recognised (Even with RefSeq and EntreZ ID for my genes/proteins). I am wondering if there is something I can do to fix this as it seems very useful. Thanks in advance!
For the gene ontology, you would need to use a tool that has plant GO terms - a websearch finds several papers on plant GO databases, but not so many tools that allow you to interrogate them. You could give this one a go: bioinformatics.sdstate.edu/go/ or search to see what else you can find.
great video! I still cant figure out how to use this for my RNA seq analysis though. The terms that pop up doesn't make sense with my samples. I don't know how to use this to shortlist pathways and genes to test
Maybe GSEA isn't the best tool for your experiment? Drop me an email (katherine.west at glasgow.ac.uk) explaining what your experiment is and what you want to find out, and I'll have a think about it...
@reaven Sorry - your comment referred to the video on DAVID, so my answer wasn't useful! Try Toppgene instead (see my tutorial) - it uses a more recent database and there are so many categories there is usually something more related to your sample.
Are we supposed to separate the differentially expressed genes from RNA seq into upregulated and downregulated lists before doing DAVID GO analysis? i.e. 2 runs one for upreg one for downreg
Hi ma'am, i have a question. Is there any experiment according to molecular mechanisms by which increases in levels of microRNA-21 in ovarian cancer may drive drug resistance?
That is not something we can help with. I am sure there are good review articles on miR-21 you can find to get you started. Best wishes with your studies
Thank you for sharing such a useful video :) I have tried to define transcription factors from my gene list. I uploaded my gene list to DAVID and clicked on the GO, I have checked many transcription factors related terms on both BP and MF parts, I could not be sure which one should be selected to identify transcription factors. Can you guide me? Thank you in advance :)
Transcription factors can be found in both BP and the MF sections. I would say, though, that the GO terms are not 100% accurate - I think they often miss things, so it might be worth trying to find another source for a list of TFs you can compare your list to. You might like to try GO using Toppgene instead of DAVID (see my other video). I prefer the output from that, and one of the additional output sections might be useful in flagging up transcription factors.
DAVID is useful as a first step for many users. The pathways and clustering tools are helpful. Many researchers have used DAVID in the past, so it remains useful, even if only for comparison. However, DAVID is not up to date, as you say, and is not particularly sensitive. The outputs are susceptible to confirmation bias too. We will review a range of gene ontology and pathway analysis tools once our tutorial series is complete, and will offer recommendations. In the meantime, experienced users should take a look at the guidance in this article, which we are supportive of: www.nature.com/articles/s41596-018-0103-9.pdf 😀
Genomics Gurus Thank you. Would you be interested in collaborating to do an IPA Software tutorial? Is there a contact email for you/your team? I’m an advance user of it, and a postdoctoral fellow at MD Anderson. I know that not very many people have access to this license, but it is quite heavy on users and can compliment your freeware tutorials.
Hello. Great tutorial thank you! I am having a hard time visualising certain components of this software such as the diagram with the green and black boxes for the functional clustering and also have difficulty downloading the gene list to an Excel format (it keeps showing it as a txt file on the browser when I click save as). Is there specific browser version you need to use for visualising or downloading these gene list? Thank you!
Hi Felix. Using chrome: when you get the txt file showing in your browser, right click to bring up a dialogue, then select "save as". This saves as a txt file. Open excel, then open the txt file and it will import it into an excel file. What is the problem with visualising the functional clustering diagram? Do you get the page but it's not displaying the diagram? I think this is an issue with bandwidth/download time. It might help to try a different browser. Or, right click to save the html page to your computer (not a cloud drive), then open that as a standalone page - that might avoid problems with bandwidth. let me know if you solve it!
@@GenomicsGurus Hi there, thank you for the very quick reply. The first problem is now solved. I can now export the txt file as an Excel file which is definitely really helpful! Thank you for the very clear directions. The problem with visualising the functional clustering diagram was exactly as you described. When I click on the little green icon, I was redirected to the page but the diagram is not displaying. I tried waiting in case it is a bandwidth issue but it still would not load even after waiting a while. I also tried multiple browsers: Chrome, Edge, IE, Firefox, Safari and could not for the life of me get this to work. I also tried what you suggested and tried opening as a standalone page and it still would not load. I really dont think its bandwidth/download time related as I am accessing this on my home computer connected to pretty fast and stable internet connection. Any other suggestion? A bit lost and disappointed if I could not get the beautiful functional clustering diagram as shown on your video!
A bit of experimentation shows I also get this problem when there are too many genes. It's a javascript diagram, and I guess it's too big for javascript to cope with. I'm not a programmer so can't work out how to solve it, I'm afraid :(
Dear Genomic Gurus, Thank for the great video. I am wondering whether we can perform overrepresentation analysis using DAVID tool in addition to Enrichment? thanks
