In any event, thanks a bunch. This is more than enough to get me started.
@jogoodma glad we could help! would be very interested to see what you do with this. As @cmdcolin mentioned, we have been extending our analytics framework recently, trying to instrument the code more carefully to model user behavior better. It would be great to develop a good data resource around this and we could potentially partner up with you if this is something you are interested in as well.
@scottcain I just saw your comments from yesterday about using NextStrain data from UCSC. Bit concerned about the idea of "liberating" GISAID data, and similarly about the idea of using UCSC bandwidth. Is there any way to do what you want to do while bypassing these issues? e.g. something you could do is not distribute the data yourself, but include a README that describes how to get it?
@ihh I’m a little less worried about UCSC’s bandwidth after @cmdcolin pointed out this entry in the FAQ: https://genome.ucsc.edu/FAQ/FAQdownloads.html#download2
I think that entry is for tools like IGV that might be pulling data directly from UCSC, but applies equally well to JBrowse. Additionally, I really doubt they’d notice a blip of increased bandwidth utilization from JBrowse users.
fixBounds was added pretty recently i think for doing that
I’m thinking about how I can make use of variant data in VCF where the results of the Variant Effect Predictor (VEP, Ensembl) for individual transcripts are included (so if a variant occurs in a gene with 10 transcripts, I have 10 results in that line of VCF to indicate how it affects each transcript). While I have a picture in my head, I don’t think it will be realistic to implement in JBrowse. I’m envisioning a super-glyph where it would display transcripts with the variant glyph just above (like, touching) the transcript, where I could color code the variant according to, say, severity of the effect. Potential problems include so many variants that there is just a mess (in which case, filtering would come in handy). Does this feel realistic and useful?
that is interesting
i would say, megaglyph sounds cool
the display of multiple transcripts per gene is already quite space-consuming
might as well just eat up a little more space for the good benefit
i think, even, the default behavior of jbrowse to display all transcript variants is somewhat debateable at times. that results in tall tracks very often
Re: variant tracks: agreed. That’s why I’d like to see filtering with lots of knobs and switches.
more expressive ui
i proposed having bcftools filters recently https://samtools.github.io/bcftools/howtos/filtering.html which i think are quite powerful, and are flexible text expressions. slightly cryptic but powerful in the right hands
But yeah, that would be cool
but in general they are one of the most powerful tools for analysis
@cmdcolin just now, on dev branch:
TypeError: this.config.replaceRefRename is not a function at http://18.212.36.249/jbrowse/?data=data%2FSARS-CoV-2&loc=NC_045512.2&tracks=DNA
in jbrowse.conf
replaceRefRename = nc_045512v2
customRefRename = nc_045512.2 which I assume is the appropriate syntax
they are callbacks
customrefname is to add to the custom refname renaming logic
replacerefname is to fully replace the refname renaming logic