Written and maintained by CASRAI Editorial Board
Last updated
Who this is for: researchers and lab or core-facility staff who have been told to use DNAnexus and need to know what it is and how a first analysis runs from the command line. After reading, you should be able to install the tools, log in, upload data, launch an app and read the job output.
What does DNAnexus do?
DNAnexus is a cloud platform for storing, analysing and sharing omics data. Its own About page describes the mission as building “the trusted infrastructure where the world’s omics data and AI come together to advance human health,” and lists pharma, diagnostic labs, health systems, biobanks and regulators among the organisations it serves (DNAnexus, About, read 2026-11-07). In practice, a research group uploads sequence or other omics files into a project, runs apps and workflows against them on platform compute, and keeps the inputs, outputs and job records together. The platform documentation is organised into Getting Started, User (web interface and command line), Developer (API, apps, workflows), Administrator (organisation permissions, billing, authentication), Downloads (SDK) and Release Notes.
It is a hosted platform, not a single statistical package: the analysis is done by apps and workflows that run inside it. For the wider field, see what bioinformatics is.
Access and install the dx command-line tools
Access to the platform is through an account on the vendor’s site; the documentation covers the web interface and the command line. The command-line client and Python SDK are distributed as the dxpy package. According to the DNAnexus Downloads page (read 2026-11-07):
- Python 3 must be installed first.
- Install with
pip3 install dxpy; upgrade withpip3 install --upgrade dxpy. - On Windows, the SDK requires Windows 10 or later.
- On macOS, a failed compile of the psutil dependency may need
export ARCHFLAGS="-arch x86_64"before reinstalling.
First analysis, step by step
The sequence below follows the official CLI quickstart (DNAnexus documentation, read 2026-11-07). Names such as job-xxxx are placeholders for IDs the platform returns.
- Log in. Run
dx login. The documentation states that this saves your authentication token and current project settings to a local configuration file. - Select a project. The quickstart selects the public reference-genome project with
dx select --public --name "Reference Genome Files*". Use your own project when working with your data. - Upload data.
dx upload --wait small-celegans-sample.fastq. The--waitflag makes the command wait until the upload is complete. - Find an app.
dx find appslists available apps. - Run it.
dx run bwa_mem_fastq_read_mapperprompts for inputs interactively. For a non-interactive run, such asdx run freebayes -y -i<input>=..., list the input names withdx run freebayes -h. - Follow the job.
dx watch job-xxxxstreams the job log, including STDOUT, STDERR and INFO lines. - Download results.
dx download small-celegans-sample.fastqkeeps the platform file name unless you pass-oor--output.
Reading job output
Three commands cover job inspection, all from the same quickstart: dx watch job-xxxx for the live log, dx describe job-xxxx for the job’s recorded details, and dx find jobs to list jobs in the current project. Check STDERR in the watch output first when a job does not behave as expected. Output files are stored in the project and retrieved with dx download. If your results are alignments, BAM index files explain the companion .bai files you may see.
Troubleshooting
dxnot found or install fails: confirm Python 3 is the interpreter pip3 belongs to; the Downloads page requires Python 3.- macOS build error for psutil: set
ARCHFLAGSas above. - Windows: the SDK requires Windows 10 or later.
- Run command rejects inputs: input names are app-specific; print them with
dx run <app> -h. - File not yet present after upload: use
dx upload --waitso the command returns only when the upload is complete. - Wrong project or expired session: run
dx loginagain and re-select the project.
Free alternatives
These are generic alternatives for running bioinformatics analyses. Descriptions come from each project’s own site (read 2026-11-07); they differ in scope and are not drop-in replacements.
| Option | What it is, per its own site | Cost statement on its site |
|---|---|---|
| Galaxy | Open source platform for accessible, reproducible, and transparent computational research; runs in a web browser without programming or local installation | Free to use on public servers worldwide (usegalaxy.org, usegalaxy.eu, usegalaxy.fr, usegalaxy.org.au, usegalaxy.ca) |
| Nextflow | Workflow language for scalable, reproducible, portable pipelines; supports Docker and Singularity, resumable execution, and executors for SLURM, LSF, Kubernetes and major clouds | Described as open-source software |
Nextflow is a workflow engine you run on your own or cloud compute, whereas DNAnexus and Galaxy are hosted platforms. See Snakemake vs Nextflow for workflow-engine differences, and CyVerse for another research cyberinfrastructure.
FAQ
Is DNAnexus free?
The documentation pages used for this guide do not state pricing, so none is given here. Account and billing terms are set out by the vendor at dnanexus.com; the Administrator section of the documentation covers billing for organisations.
Who is the CEO of DNAnexus?
Leadership is listed by the company itself; check the leadership section of dnanexus.com for the current officers, since this changes over time.
Where is DNAnexus located?
Office and contact details are published on the vendor’s website at dnanexus.com. This guide does not restate an address.
How do I log in to DNAnexus?
From the command line, run dx login after installing dxpy; the web interface is reached through your account on the vendor’s site.
Sources
- DNAnexus Documentation, overview (read 2026-11-07)
- DNAnexus Documentation: Downloads (read 2026-11-07)
- DNAnexus CLI quickstart (read 2026-11-07)
- DNAnexus: About (read 2026-11-07)
- Galaxy Project and Nextflow (read 2026-11-07)
