Command-Line Quickstart

Set up the dx client for command-line access to the full range of DNAnexus Platform features.

The dx command-line client is included in the DNAnexus SDK (dx-toolkit) , which helps users get the most out of the DNAnexus platform. You can use the dx client to log in, upload data, navigate and organize your data, and launch analyses on the DNAnexus platform.

All of the projects and data that we use in this Quickstart are publicly available, so you can follow along step-by-step. See also index of dx commands.

In the examples that follow, text is sometimes shown in bold to indicate when and what you will need to enter at the bash prompt or an interactive prompt of dx.

Command-Line Help

If you have questions about what a dx command does, you can use the dx help command or add the -h/--help flag to any dx command. For example, to learn more about what the command dx ls can do, you can run dx help ls, dx ls -h, or dx ls --help.

$ dx help ls
usage: dx ls [-h] [--color {off,on,auto}] [--delimiter [DELIMITER]]
[--env-help] [--brief | --summary | --verbose] [-a] [-l] [--obj]
[--folders] [--full]
List folders and/or objects in a folder
... # output truncated for brevity

You can also use the dx help command to explore the available commands. Running dx help by itself will show you the different categories of dx commands, and you can run dx help <category> to see the related commands.

Tab Completion

In addition to the command dx help, tab completion is available for the dx tool when using the bash shell. Simply press <TAB> once to see available completions for the current command. When there are multiple completions, you can press <TAB> twice to see all possible completions.

Step 0: Install the SDK

To use the command-line interface, make sure you have installed the DNAnexus Software Development Kit (SDK).

Upgrading the SDK

To update your version of the command-line tool, you can run the command dx upgrade. For more information about upgrading your SDK, please see the Downloads page.

Step 1: Log In

The first thing you'll need to do is to log in. If you haven't created a DNAnexus account yet, visit the website and sign up. User signup is not supported on the command line.

$ dx login
Acquiring credentials from
Username: <your username>
Password: <your password>
No projects to choose from. You can create one with the command "dx new project". To pick from projects for which you only have VIEW permissions, use "dx select --level VIEW" or "dx select --public".

Your authentication token and your current project settings have now been saved in a local configuration file, and you're ready to start accessing your project.

Step 2: Look Around

Public Projects

Let's look inside some of the public projects that have already been set up.

$ dx select --public --name "Reference Genome Files*"
Available public projects:
0) Reference Genome Files: Azure Amsterdam (CONTRIBUTE)
1) Reference Genome Files: AWS China (VIEW)
2) Reference Genome Files: AWS Germany (VIEW)
3) Reference Genome Files: Azure US (West) (VIEW)
4) Reference Genome Files: AWS US (East) (VIEW)
5) Reference Genome Files: AWS Sydney (CONTRIBUTE)
Pick a numbered choice: 4
Setting current project to: Reference Genome Files: AWS US (East)

By running the dx select command and picking a project, you've now done the command-line equivalent of going to the project page for Reference Genome Files: AWS US (East) (platform login required to access this link) on the website. This is a DNAnexus-sponsored project containing popular genomes for you to use when running analyses with your own data.

For more information about the dx select command, please see the Changing Your Current Project page.

NOTE: You will never be charged for DNAnexus-sponsored data, so you can copy data from this project however many times you'd like, free of charge.

Now, you can list all of the data in the top-level directory of the project you've just selected by running the command dx ls. You can also see the contents of a folder by running the command dx ls <folder_name>.

$ dx ls
C. Elegans - Ce10/
D. melanogaster - Dm3/
H. Sapiens - GRCh37 - b37 (1000 Genomes Phase I)/
H. Sapiens - GRCh37 - hs37d5 (1000 Genomes Phase II)/
H. Sapiens - GRCh38/
H. Sapiens - hg19 (Ion Torrent)/
H. Sapiens - hg19 (UCSC)/
M. musculus - mm10/
M. musculus - mm9/
$ dx ls "C. Elegans - Ce10/"
... # output truncated for brevity

You can avoid typing out the full name of the folder by typing in dx ls C and then pressing <TAB>. The folder name will auto-complete from there.

You don't have to be in a project to inspect its contents. You can also look into another project, and a folder within the project, by giving the project name or ID, followed by a colon (:) and the folder path. Here, we list the contents of the publicly available project "Demo Data" using both its name and ID.

$ dx ls "Demo Data:/SRR100022/"
$ dx ls -l "project-BQbJpBj0bvygyQxgQ1800Jkk:/SRR100022/"
Project: Demo Data (project-BQbJpBj0bvygyQxgQ1800Jkk)
Folder : /SRR100022
State Last modified Size Name (ID)
... # output truncated for brevity

As shown above, you can use the -l flag in conjunction with dx ls to list more details about files, such as the time a file was last modified, its size (if applicable), and its full DNAnexus ID.

Describe DNAnexus Objects

You can use the dx describe command to learn more about files and other objects on the platform. Given a DNAnexus object ID or name, dx describe will return detailed information about the object in question. dx describe will only return results for data objects to which you have access.

Besides describing data and projects (examples for which are shown below), you can also describe apps, jobs, and users.

Example: Describing a file

Below, we describe the reference genome file for C. elegans, which lives in the "Reference Genome Files: AWS US (East)" project that we've been using.

$ dx describe "Reference Genome Files\: AWS US (East):/C. Elegans - Ce10/ce10.fasta.gz"
Result 1:
ID file-BQbY9Bj015pB7JJVX0vQ7vj5
Class file
Project project-BQpp3Y804Y0xbyG4GJPQ01xv
Folder /C. Elegans - Ce10
Name ce10.fasta.gz
State closed
Visibility visible
Types -
Properties Assembly=UCSC ce10,
s/ce10.2bit, Species=Caenorhabditis elegans, Taxonomy
Tags -
Outgoing links -
Created Tue Sep 30 18:54:35 2014
Created by bhannigan
via the job job-BQbY8y80KKgP380QVQY000qz
Last modified Thu Mar 2 12:17:27 2017
Media type application/x-gzip
archivalState "live"
Size 29.21 MB, sponsored by DNAnexus

Example: Describing a project

Below, we describe the publicly available Reference Genome Files project that we've been using.

$ dx describe "Reference Genome Files\: AWS US (East):"
Result 1:
ID project-BQpp3Y804Y0xbyG4GJPQ01xv
Class project
Name Reference Genome Files: AWS US (East)
Billed to org-dnanexus
Access level VIEW
Region aws:us-east-1
Protected true
Restricted false
Contains PHI false
Created Wed Oct 8 16:42:53 2014
Created by tnguyen
Last modified Tue Oct 23 14:15:59 2018
Data usage 0.00 GB
Sponsored data 519.77 GB
Sponsored egress 0.00 GB used of 0.00 GB total
Tags -
Properties -
downloadRestricted false
defaultInstanceType "mem2_hdd2_x2"

Step 3: Create Your Own Project

Now, we'll use the command dx new project to create a new project.

$ dx new project "My First Project"
Created new project called "My First Project"
Switch to new project now? [y/N]: y

The text project-xxxx denotes a placeholder for a unique, immutable project ID. For more information about object IDs, see the Entity IDs page.

You're now ready to start uploading your data and running your own analyses!

Step 4: Upload and Manage Your Data

If you have a sample you would like to analyze, you can use the dx upload command or the Upload Agent if you have installed it. You can also download the file small-celegans-sample.fastq, which represents the first 25000 C. elegans reads from SRR070372. We will use this file again later to run through a sample analysis.

NOTE: For uploading multiple or large files, we strongly recommend that you use the Upload Agent; it will compress your files and upload them in parallel over multiple HTTP connections and boasts other features such as resumable uploads.

The following command uploads the small-celegans-sample.fastq file into the current directory of the current project. The --wait flag tells dx upload to wait until it has finished uploading the data before returning the prompt and describing the result.

$ dx upload --wait small-celegans-sample.fastq
[===========================================================>] Uploaded (16801690 of 16801690 bytes) 100% small-celegans-sample.fastq
ID file-xxxx
Class file
Project project-xxxx
Folder /
Name small-celegans-sample.fastq
State closed
Visibility visible
Types -
Properties -
Tags -
Details {}
Outgoing links -
Created Sun Jan 1 09:00:00 2017
Created by amy
Last modified Sat Jan 1 09:00:00 2017
Media type text/plain
Size 16.02 MB

NOTE: If you run the same command but add the flag --brief, only the file ID (in the form of file-xxxx) will be printed to the terminal. Other dx commands will also accept the --brief flag and will also report only object IDs.

Examine Data

To take a quick look at the first few lines of the file you just uploaded, use the dx head command. By default, it prints the first 10 lines if given a file.

Let's run it on the file we just uploaded and use the -n flag to ask for the first 12 lines (the first 3 reads) of the FASTQ file.

$ dx head -n 12 small-celegans-sample.fastq
@SRR070372.1 FV5358E02GLGSF length=78
+SRR070372.1 FV5358E02GLGSF length=78
@SRR070372.2 FV5358E02FQJUJ length=177
+SRR070372.2 FV5358E02FQJUJ length=177
@SRR070372.3 FV5358E02GYL4S length=70
+SRR070372.3 FV5358E02GYL4S length=70
@@@@@[email protected]=?::[email protected]?DFFHHFDDDDFFFDDBBBB<<410

Download Data

If you'd like to download a file from the platform, just use the dx download command. This command will use the name of the file for the filename unless you specify your own with the -o/--output flag.

$ dx download small-celegans-sample.fastq
[ ] Downloaded 0 byte
[===========================================================>] Downloaded 16.02 of
[===========================================================>] Completed 16.02 of 16.02 bytes (100%) small-celegans-sample.fastq


Files have different available fields for metadata, such as "properties" (key-value pairs) and "tags".

Step 5: Analyze a Sample

For the next few steps, if you would like to follow along, you will need a C. elegans FASTQ file. We will map the reads against the ce10 genome. If you haven't already, you can download and use the following FASTQ file, which contains the first 25,000 reads from SRR070372: small-celegans-sample.fastq.

NOTE: If you'd like, you can also substitute your own reads file for a different species (though it may take longer to run through the example). For your convenience, DNAnexus has already imported a variety of reference genomes to the platform. If you have your own FASTA file that you would like to use, you can upload the file and create genome indices for either BWA or Bowtie2 using the BWA FASTA Indexer app or the Bowtie2 FASTA Indexer app (platform login required to access these links).

The following walkthrough is helpful if you would like to understand what all the commands do and take a look at what apps you're running, but if you're just interested in converting a FASTQ file to a VCF file via BWA and the FreeBayes variant caller, then you can skip ahead to the Automate It section, where you can see all the commands necessary for running apps.

Upload Reads

If you have not yet done so, you can upload a FASTQ file for analysis.

$ dx upload small-celegans-sample.fastq --wait

For more information about using the command dx upload, please see the dx upload page.

Map Reads

Next, use the BWA-MEM app (platform login required to access this link) to map the uploaded reads file to a reference genome.

Find the app name

First of all, if you don't know the command-line name of the app you would like to run, you have two options:

  1. You can navigate to its web page from the Apps page (platform login required to access this link) on the platform. The app's page will tell you how to run it from the command-line. You can find more information about the app we're running on the BWA-MEM FASTQ Read Mapper page (platform login required to access this link).

  2. Alternatively, you can search for apps from the command line by running the command dx find apps. You will find the name of the app that you can use on the command line in the parentheses (underlined below).

$ dx find apps
x BWA-MEM FASTQ Read Mapper (bwa_mem_fastq_read_mapper), v1.4.0

Install and run the app

Now, let's install the app using dx install and check that it has been installed. While you do not always need to install an app to run it, you may find it useful as a bookmarking tool.

$ dx install bwa_mem_fastq_read_mapper
Installed the bwa_mem_fastq_read_mapper app
$ dx find apps --installed
BWA-MEM FASTQ Read Mapper (bwa_mem_fastq_read_mapper), v1.4.0

We can now run the app using dx run. We will run it without any arguments; it will then prompt us for required and then optional arguments.

$ dx run bwa_mem_fastq_read_mapper
Entering interactive mode for input selection.
Input: Reads (reads_fastqgz)
Class: file
Enter file ID or path (<TAB> twice for compatible files in current directory,'?' for help)
reads_fastqgz: small-celegans-sample.fastq
Input: BWA reference genome index (genomeindex_targz)
Class: file
project-BQpp3Y804Y0xbyG4GJPQ01xv://file-\* (DNAnexus Reference Genomes)
Enter file ID or path (<TAB> twice for compatible files in current
directory,'?' for more options)
genomeindex_targz: "Reference Genome Files:/C. Elegans - Ce10/ce10.bwa-index.tar.gz"
Select an optional parameter to set by its # (^D or <ENTER> to finish):
[0] Reads (right mates) (reads2_fastqgz)
[1] Add read group information to the mappings (required by downstream GATK)? (add_read_group) [default=true]
[2] Read group id (read_group_id) [default={"$dnanexus_link": {"input": "reads_fastqgz", "metadata": "name"}}]
[3] Read group platform (read_group_platform) [default="ILLUMINA"]
[4] Read group platform unit (read_group_platform_unit) [default="None"]
[5] Read group library (read_group_library) [default="1"]
[6] Read group sample (read_group_sample) [default="1"]
[7] Output all alignments for single/unpaired reads? (all_alignments)
[8] Mark shorter split hits as secondary? (mark_as_secondary) [default=true]
[9] Advanced command line options (advanced_options)
Optional param #: <ENTER>
Using input JSON:
"reads_fastqgz": {
"$dnanexus_link": {
"project": "project-B3X8bjBqqBk1y7bVPkvQ0001",
"id": "file-B3P6v02KZbFFkQ2xj0JQ005Y"
"genomeindex_targz": {
"$dnanexus_link": {
"project": "project-BQpp3Y804Y0xbyG4GJPQ01xv",
"id": "file-BQbYJpQ09j3x9Fj30kf003JG"
Confirm running the applet/app with this input [Y/n]: <ENTER>
Calling app-BP2xVx80fVy0z92VYVXQ009j with output destination
Job ID: job-xxxx

Monitor your job

You can use the command dx watch to monitor jobs. The command will print out the log file of the job, including the STDOUT, STDERR, and INFO printouts.

You can also use the command dx describe job-xxxx to learn more about your job. If you don't know the job's ID, you can use the command dx find jobs to list all the jobs run in the current project, along with their status and when they began.

$ dx find jobs
* BWA-MEM FASTQ Read Mapper (bwa_mem_fastq_read_mapper:main)(done) job-xxxx
amy 2017-01-01 09:00:00 (runtime 0:00:27)
$ dx describe job-xxxx

There are also additional options that you can use to restrict your search of previous jobs, such as by their names or when they were run.

After your job finishes

You should now see two new files in your project: the mapped reads in a BAM file, and an index of that BAM file with a .bai extension. You can refer to the output file by name or by the job that produced it using the syntax job-xxxx:<output field>. Try it yourself with the job ID you got from calling the BWA-MEM app!

$ dx ls
$ dx describe small-celegans-sample.bam
$ dx describe job-xxxx:sorted_bam

Call Variants

You can use the FreeBayes Variant Caller app (platform login required to access this link) to call variants on your BAM file.

This time, we won't rely on the interactive mode to enter our inputs. Instead, we will provide them directly. But first, let's look up the app's spec so we know what the inputs are called. For this, let's run the command dx run freebayes -h.

$ dx run freebayes -h
usage: dx run freebayes [-iINPUT_NAME=VALUE ...]
App: FreeBayes Variant Caller
Calls variants (SNPs, indels, and other events) using FreeBayes
See the app page for more information:
Sorted mappings: -isorted_bams=(file) [-isorted_bams=... [...]]
One or more coordinate-sorted BAM files containing mappings to call
variants for.
Genome: -igenome_fastagz=(file)
A file, in gzipped FASTA format, with the reference genome that the
reads were mapped against.

Optional inputs are shown using square brackets ([]) around the command-line syntax for each input. You'll notice that there are two required inputs that must be specified:

  1. Sorted mappings (sorted_bams): A list of files with a .bam extension.

  2. Genome (genome_fastagz): A reference genome in FASTA format that has been gzipped.

NOTE: You can also run dx describe freebayes for a more compact view of the input and output specifications. By default, it will hide the advanced input options, but you can view them using the --verbose flag.

Run the app with a one-liner using a job-based object reference

It is sometimes more convenient to run apps using a single one-line command. You can do this by specifying all the necessary inputs either via the command line or in a prepared file. We will use the -i flag to specify inputs as suggested by the output of dx run freebayes ‑h:

  • sorted_bams: The output of the previous BWA step (see the Map Reads section for more information).

  • genome_fastagz: The ce10 genome in the Reference Genomes project.

To specify new job input using the output of a previous job, we'll use a [job-based object reference](/Developer-Tutorials/Sample-Code?bash#Use-job-based-object-references-(JBORs)) via the job-xxxx:<output field> syntax we used earlier.

NOTE: You can use job-based object references as input even before the referenced jobs have finished. The system will simply wait until the input is ready to begin the new job.

Replace the job ID below with that generated by the BWA app you ran earlier. The -y flag skips the input confirmation.

$ dx run freebayes -y \
-igenome_fastagz=Reference\ Genome\ Files:/C.\ Elegans\ -\ Ce10/ce10.fasta.gz \
Using input JSON:
"genome_fastagz": {
"$dnanexus_link": {
"project": "project-xxxx",
"id": "file-xxxx"
"sorted_bams": {
"field": "sorted_bam",
"job": "job-xxxx"
Calling app-BFG5k2009PxyvYXBBJY00BK1 with output destination
Job ID: job-xxxx

Automatically run a command after a job finishes

You can use the command dx wait to wait for a job to finish. If we run the following command right after running the Freebayes app, it will show you the recent jobs only after the job has finished.

$ dx wait job-xxxx && dx find jobs
Waiting for job-xxxx to finish running...
* FreeBayes Variant Caller (done) job-xxxx
amy 2017-01-01 09:00:00 (runtime 0:05:24)

Congratulations! You have now called variants on a reads sample, and you did it all on the command line. Now let's look at how you can automate this process.

Automate It

The beauty of the command-line interface is the ability to automate processes. In fact, we can automate everything we just did. The following script assumes that you've already logged in and is hardcoded to use the ce10 genome and takes in a local FASTQ file as its command-line argument.

# Usage: <> local_fastq_filename.fastq
reference="Reference Genome Files:/C. Elegans - Ce10/ce10.fasta.gz"
bwa_indexed_reference="Reference Genome Files:/C. Elegans - Ce10/ce10.bwa-index.tar.gz"
reads_file_id=$(dx upload "$local_reads_file" --brief)
bwa_job=$(dx run bwa_mem_fastq_read_mapper -ireads_fastqgz=$reads_file_id -igenomeindex_targz="$bwa_indexed_reference" -y --brief)
freebayes_job=$(dx run freebayes -isorted_bams=$bwa_job:sorted_bam -igenome_fastagz="$reference" -y --brief)
dx wait $freebayes_job
dx download $freebayes_job:variants_vcfgz -o "$local_reads_file".vcf

What's Next

You're now ready to start scripting using dx. As shown in some of the examples above, the --brief flag can come in handy for scripting. A list of all dx commands and flags is on the Index of dx Commands page.

For more detailed information about running apps and applets from the command line, see the Running Apps and Applets page.

For a comprehensive guide to the DNAnexus SDK, see the SDK documentation.

Want to start writing your own apps? Check out the Developer Portal for some useful tutorials.