How to save space for storing genome reference and tracks？

[sutfei]

Hi,

We want to integrate jbrowse into our website for visualization. We are planning to put the website on cloud sever we rented however, we have 25 species to display which take too much space. Is there any way we could compress the assemblies and tracks we have to save space?

Thanks a lot！

[cmdcolin]

JBrowse 2 can use twobit or bgzip fasta for reference sequence which are both pretty optimally compressed sequences. For gene annotations gff3tabix is bgzip which is our recommended format (covered in quick start). Generally the biggest data files are bam and cram files where one e.g. human wgs bam can be like 50-100Gb (gigabytes) while the bgzip fasta for human is maybe around 800Mb (megabytes)

[cmdcolin]

currently we recommend loading bed tabix or bigBed. with bed tabix, you can run sort -k1,1 -k2,2n yourfile.bed > yourfile.sorted.bed; bgzip yourfile.sorted.bed; tabix yourfile.sorted.bed.gz; jbrowse add-track yourfile.sorted.bed.gz

we can probably make an plaintext bed support but the bed tabix will be compressed and good for large or small bed files anyways

[sutfei]

I have some reference seq as fna file. When I tried to run jbrowse add-assembly GCF_000001735.4_TAIR10.1_genomic.fna.gz --load copy --name ArabidopsisThaliana --force. It shows "Error: Could not determine sequence type". So fna.gz file doesn't work?

[cmdcolin]

It does not recognize .fna by default, we can probably add that, but you can use --type bgzipFasta to give it the hint

jbrowse add-assembly volvox.fna.gz --type bgzipFasta 

[sutfei]

It does not recognize .fna by default, we can probably add that, but you can use --type bgzipFasta to give it the hint

jbrowse add-assembly volvox.fna.gz --type bgzipFasta 

It works! Many thank!

[cmdcolin]

glad this works. added a note to the "Could not determine sequence type" error so that in the future it recommends adding --type