11 weeks ago by
Is Jbrowse the right tool?
I have some RNA-sequencing data and transcriptome assembly for a
Jbrowse is a genome browser. A genome consists of relatively fewer 'scaffolds/contigs/chromosomes' containing
multiple structures (gene-models) and onto which other genomic feature tracks (e.g. quantitative) can be mapped.
A de-novo transcriptome consists of relatively many (possibly 100,000 + ) and shorter contigs, with little additional structure or names to display. Even if there are few structures, a genome browser like Jbrowse handles that many contigs poorly, creates a relatively high burden for the database back-end, and it will give little benefit for the user, because they will have a hard time to find the right 'gene/transcript', one will mostly need tblast(x/n) to find anything sensible.
That is why many assemblies simply exist in large repositories like the Trace/SRA and sadly some published assemblies are not even provided or made available upon request only.
Maybe a way to address sharing of your data is simply to
- Upload your raw data to either SRA or ENA (this is the most crucial step, doing a re-assembly of the transcriptome has become relatively quick and straight-forward nowadays, but if a commercial tool was used this could be blocker to exactly reproduce your results).
- Submit your assembly to TSA: https://www.ncbi.nlm.nih.gov/genbank/tsaguide/
- Provide a FASTA file as supplementary data with your paper (if you can't do 2)
That way, you ensure availability of your assembly data even after the event that you need to discontinue your service in the future.
There are many ways to get a sequence database web-site up and running.
Possibly one of the easiest ways to get a running Jbrowse up on your local computer is using Docker and pulling one of the Jbrowse images.
The most up-to-date I found just now is the Bioconda Jbrowse repository https://quay.io/repository/biocontainers/jbrowse?tab=info
That should also be the fastest way to get an instance up including a web server and back-end (I think this should run more or less out of the box) and it might be even doable for beginners in the field, however you need to be willing to learn the basics of the operating system and web-servers, docker, jbrowse and their configuration.
Another option could be GMOD in the Cloud but the software in this image is very outdated according to the docs (state of ~2013).
The essential decision is then where to deploy:
- deploy on local computer
- deploy 'in the cloud'
Here is an overview of some commercial cloud/docker hosters, some provide limited 'free' resources.
I recommend to start on a local Linux machine, any (half-)recent computer will do, to get some experience and practice without a pricing to consider.