Skip to main content

JBrowse web setup using the CLI

Prerequisites

  • Node.js 12+. Note: we recommend not using apt to install Node.js, it often installs old versions. Good alternatives include NodeSource or NVM.
  • Samtools installed e.g. sudo apt install samtools or brew install samtools, used for creating FASTA index and BAM/CRAM processing for creating tabix GFF
  • tabix installed e.g. sudo apt install tabix and brew install htslib, used for creating tabix indexes for BED/VCF/GFF files
  • (optional) genometools installed e.g. sudo apt install genometools or brew install brewsci/bio/genometools used for sorting GFF3. can use awk instead of genometools instead

Installing the JBrowse CLI

The JBrowse CLI can help perform many tasks to help you manage JBrowse 2, such as:

  • create a new instance of JBrowse 2 automatically
  • update an existing instance of JBrowse 2 with the latest released version
  • configure your JBrowse 2 instance

To globally install the JBrowse CLI, run

npm install -g @jbrowse/cli

After running this command you can then test the installation with

jbrowse --version

which will output the current version of the JBrowse CLI.

note

If you can't or don't want to globally install the JBrowse CLI, you can also use the npx command, which is included with Node.js, to run JBrowse CLI without installing it. Simply replace jbrowse with npx @jbrowse/cli in any command, e.g.

npx @jbrowse/cli --version

Using jbrowse create to download JBrowse 2

In the directory where you would like to download JBrowse 2, run

jbrowse create jbrowse2

This fetches the latest version of jbrowse-web and unzips it to a folder named "jbrowse2" from github (https://github.com/GMOD/jbrowse-components/releases), you could run this step manually if you wanted to instead.

Checking the download

The directory where you downloaded JBrowse should look something like this:

jbrowse2/
├── asset-manifest.json
├── favicon.ico
├── index.html
├── manifest.json
├── robots.txt
├── static/
├── test_data/
└── version.txt

Running JBrowse 2

JBrowse 2 requires a web server to run. It won't work if you try to directly open the index.html in your web browser.

Oftentimes, you may put the folder on a web server in the static html folder e.g. /var/www/html/jbrowse2/ once in place, you can then visit http://yourserver/jbrowse2

You could also use a simple server to check that JBrowse 2 has been downloaded properly. Run

cd jbrowse2/
npx serve .
# or
npx serve -S . # if you want to refer to symlinked data later on

which will start a web server in our JBrowse 2 directory.

Navigate to the location specified in the CLI's output (likely http://localhost:3000).

Your page should look something like this:

JBrowse 2 screen showing no configuration found
Figure: JBrowse 2 screen showing no configuration found

Click on the sample config to see JBrowse 2 running with a demo configuration. It should look like this:

JBrowse 2 screen with a sample configuration
Figure: JBrowse 2 screen with a sample configuration

Congratulations! You're running JBrowse 2.

Adding tracks

Now that JBrowse 2 is set up, you can configure it with your own genomes and tracks.

Adding a genome assembly in FASTA format

The first step to creating a jbrowse config is to load a genome assembly. This is normally in FASTA format, and we will start by creating a "FASTA index" with samtools:

samtools faidx genome.fa
jbrowse add-assembly genome.fa --load copy --out /var/www/html/jbrowse/

This will output a configuration snippet to a file named /var/www/html/jbrowse/config.json if it does not already exist, or append a new assembly to that config file if it does exist. It will also copy genome.fa and genome.fa.fai to the /var/www/html/jbrowse/ folder because we used --load copy. If you wanted to symlink instead, can use --load symlink

JBrowse 2 also supports other assembly file formats, including bgzip-compressed indexed FASTA, and 2bit files.

JBrowse 2 linear genome view setup with volvox in assembly dropdown
Figure: JBrowse 2 linear genome view setup with volvox in assembly dropdown

Adding a BAM or CRAM track

For this example we will use a BAM file to add an alignments track. As with assemblies, you can add a track using local files or remote locations of your files.

samtools index file.bam
jbrowse add-track file.bam --load copy --out /var/www/html/jbrowse
samtools index file.cram
jbrowse add-track file.cram --load copy --out /var/www/html/jbrowse

This will add a track configuration entry to /var/www/html/jbrowse/config.json and copy the files into the folder as well. If you use --load symlink, it can symlink the files instead. To see more options adding the track, such as specifying a name, run jbrowse add-track --help.

If you have JBrowse 2 running as described in the JBrowse web quickstart, you can refresh the page and an add a linear genome view of the volvox assembly. Then open track selector, and you will see the alignments track.

JBrowse 2 linear genome view with alignments track
Figure: JBrowse 2 linear genome view with alignments track

Adding a VCF track

Adding a variant track is similar to adding an alignments track. For this example, we will use a VCF file for the track. JBrowse 2 expects VCFs to be compressed with bgzip and tabix indexed.

To add the track, run

bgzip file.vcf
tabix file.vcf.gz
jbrowse add-track file.vcf.gz --load copy --out /var/www/html/jbrowse
note

Note if you get errors about your VCF file not being sorted when using tabix, you can use bcftools to sort your VCF.

bcftools sort file.vcf > file.sorted.vcf
bgzip file.sorted.vcf
tabix file.sorted.vcf.gz

You can also bgzip and index with the bcftools tool.

bcftools view volvox.vcf --output-type z > volvox.vcf.gz
rm volvox.vcf
bcftools index --tbi volvox.vcf.gz

For more info about bgzip, tabix, and bcftools, see https://www.htslib.org/.

JBrowse 2 linear genome view with variant track
Figure: JBrowse 2 linear genome view with variant track

Adding a BigWig/BigBed track

Probably one of the most simple track types to load is a BigWig/BigBed file since it does not have any external index file, it is just a single file.

jbrowse add-track file.bw --load copy --out /var/www/html/jbrowse

Adding a GFF3 file with GFF3Tabix

To load a GFF3 file, we can sort and index it with tabix, make sure you have GenomeTools (to install can use sudo apt install genometools).

gt gff3 -sortlines -tidy -retainids yourfile.gff > yourfile.sorted.gff
bgzip yourfile.sorted.gff
tabix yourfile.sorted.gff.gz
jbrowse add-track yourfile.sorted.gff.gz --load copy

As an alternative to gt gff3 -sortlines, use awk and GNU sort, as follows:

awk '$1 ~ /^#/ {print $0;next} {print $0 | "sort -t\"\t\" -k1,1 -k4,4n"}' file.gff > file.sorted.gff
bgzip file.sorted.gff
tabix file.sorted.gff.gz

The awk command is inspired by the method in the tabix documentation, but avoids subshells and properly sets the tab delimiter for GNU sort in case there are spaces in the GFF.

Adding a synteny track from a PAF file

Loading synteny data makes use of all the previous functions we've used so far in this guide.

Here, we make use of the grape and peach genome assemblies, but replace with your own data if applicable.

Use minimap2 to create a PAF file from FASTA files:

## Use minimap2 to create a PAF from your assemblies
minimap2 grape.fa peach.fa > peach_vs_grape.paf

## add each assembly to jbrowse config
## the -n flag names the assemblies explicitly
jbrowse add-assembly grape.fa --load copy -n grape --out /var/www/html/jbrowse
jbrowse add-assembly peach.fa --load copy -n peach --out /var/www/html/jbrowse

Next, we'll load the synteny "track" from the PAF file.

Order matters here for the --assemblyNames parameter:

If minimap2 is run as minimap2 grape.fa peach.fa, then you need to load as --assemblyNames peach,grape.

The order is reversed between the minimap2 and jbrowse tools.

jbrowse add-track peach_vs_grape.paf --assemblyNames peach,grape --load copy --out /var/www/html/jbrowse

Indexing feature names for searching

The final step of loading your JBrowse instance may include adding a "search index" so that you can search by genes or other features by their name or ID.

To do this we can use the jbrowse text-index command:

jbrowse text-index --out /var/www/html/jbrowse

This will index relevant track types e.g. any track with Gff3TabixAdapter (gene names and IDs) or VcfTabixAdapter (e.g. variant IDs). The command will print out a progress bar for each track that it is indexing.

This will also update your config.json so that after it completes, you can type a gene name into the "search box" in the linear genome view or other views and quickly navigate to genes by gene name.

See the text-index command docs for more info. Also see the FAQ entries for text searching

Conclusion

Now that you have JBrowse configured with an assembly and a couple of tracks, you can start customizing it further. Check out the rest of the docs for more information, especially the JBrowse CLI docs for more details on some of the steps shown here.

Miscellaneous tips

You can use --subDir to organize your data directory:

mkdir my_bams
## Copies .bam and .bai files to my_bams folder
jbrowse add-track myfile.bam --subDir my_bams --load copy --out /var/www/html/jbrowse

Upgrade JBrowse Web to the latest version

You can upgrade your JBrowse release to the latest version with:

# run this command in an existing jbrowse 2 installation
jbrowse upgrade

The above command downloads the latest jbrowse-web from github and unzips it into the current directory

Upgrade @jbrowse/cli to the latest

To upgrade the CLI tools, you can re-run the install command

npm install -g @jbrowse/cli

Output to a custom named config file, and output to subfolders

You can use filenames other than config.json, and put configs in subfolders of your jbrowse 2 installation too

jbrowse add-assembly mygenome.fa --out /path/to/my/jbrowse2/subfolder/alt_config.json --load copy

This would then be accessible at e.g. http://localhost/jbrowse2/?config=subfolder/alt_config.json