Hello everyone! I would like to know how I can work with just two BAM files (Control_sorted.bam and 24_Hrs_sorted.bam) to generate a differential gene expression table specifically for these files. I have encountered several issues while using DESeq2 and limma, and thus, I have been unable to obtain the table. Can anyone please advise me on whether it’s possible to do this without any replicates? Additionally, is there any other tool on Galaxy that I can use for this purpose without replicates? Thank you!
1 Like
Welcome @Vasudev_R_Joshi
Replicates are always required now in the original tools as they are released from Bioconductor, and that flows down to Galaxy.
Older discussion at this forum (circa 2020) when that was still being discussed. → DESeq2 or EdgeR or Limma without replication - #6 by nagul
Our guide where one public topic from the Bioconductor authors was linked. → FAQ: Extended Help for Differential Expression Analysis Tools. This guide solves a lot of common usage problems, in Galaxy or anywhere else, so maybe worth a quick read anyway?
- Differential expression tools all require sample count replicates. Rationale from two of the DEseq tool authors → https://www.seqanswers.com/forum/bioinformatics/bioinformatics-aa/26388-deseq2-without-biol-replicates
- At least two factor levels/groups/conditions with two samples each.
- All must all contain unique content for valid scientific results.
Then, discussions at the Bioconductor support forum about replicates. → Bioconductor Forum
* This is a good example from the list. → Alternatives to Voom in experiments without replicate
* If you want to explore the methods they share in the EdgeR vignette, you can do that in Galaxy in our Rstudio interactive environment.
Last advice I have for single-sample expression exploratory analysis (only!) is you could use a tool like Salmon, Kallisto or Cufflinks with just a single sample. These all produce “relative to itself” expression values on a per-sample level. The scale can be difficult to interpret though..
But bigger picture – for this part
We would like to help you here to solve the usage issue! Would you like to share back your history for some troubleshooting review? Try to leave all the data in the same history: reads through to mapping to counts to your errors. It is ok if there are multiple errors – try to keep it all along with any troubleshooting you already tried. We like details and can probably help to get you back on track!! 
You can post the share link back here in a reply, then unshare once we are done. If you are not working at a UseGalaxy server yet, maybe copy your history over to one and try some reruns (to eliminate server technical issues), then if those jobs still fail, or you can’t get inputs organized at all, please share that new combined history back. I’ll be able to see what is going on (probably!). UseGalaxy.org is a good server choice for this but UseGalaxy.eu or UseGalaxy.org.au are fine too.
Let’s start there! 
Thank you very much for your assistance. I am not experiencing any issues with DESeq2 or Limma; however, I believe the lack of replicates may be the source of the problem. I appreciate your time and attention to this matter.
1 Like
AH, ok, thanks for clarifying! I wasn’t sure if replicates were the core issue or if you were falling back to that usage because of troubles with combining multiple samples.
Then yes, you’ll need to explore different tools. There are some methods that you can do directly in R as well. The Bioconductor resources will be the definitive guides for this, and you can connect an R environment to your Galaxy history with Rstudio, or work outside of Galaxy.
Okay, I got it. Thanks a lot!