2026 Release Notes

Get details on new feature releases, improvements, and fixes across the DNAnexus Platform, related public tools, and public apps in 2026.

DNAnexus publishes release notes by deployment date. Most platform releases are deployed on Tuesdays throughout the year, which provides a predictable cadence for delivering security updates, reliability improvements, and new platform capabilities.

Product Versioning: Release versions are identified by deployment date. For example, a release deployed on June 17, 2025 can be identified as _20250617 in validation and compliance records. This format helps regulated customers identify exact production release versions for compliance.

New releaseImprovementsFixes

Release for 02-Jun-2026

Custom role permissions for organizations

With custom role permissions, organization admins can delegate specific administrative tasks to members without granting full org admin access. This supports least-privilege operations for org members who need focused administrative responsibilities.

Based on the permissions granted by an org admin, delegated members can archive and unarchive data across organization projects, demote project members, delete data, and search for data across the organization.

For details, see organization membership levels and custom role permissions.

Posit Workbench on the DNAnexus Platform

You can run Posit Workbench (formerly RStudio Workbench) on the DNAnexus Platform using your own Posit license. This integration lets your organization use existing Posit seats inside DNAnexus, so you can develop in R and Python with the full Posit Workbench environment alongside your DNAnexus data.

To enable Posit Workbench for your organization, contact DNAnexus with the required license details and seat counts. Once your organization is enabled, members can launch Posit Workbench and work in a session backed by your license.

For setup and usage instructions, see Posit Workbench (RStudio).

Improvements

  • When you reset your password, you are signed out of all active web sessions. You can optionally choose to also revoke active job and file tokens, which interrupts in-progress jobs and file transfers tied to the previous credentials.

  • The platform preserves your selected instance type when a job is automatically restarted after a non-AppInsufficientResourceError failure. Restarted jobs continue with the original instance type instead of falling back to the platform default.

Fixes

  • Fixed a bug that caused increased nginx 504 and 503 errors when listing executions through /system/findExecutions.

  • Fixed a bug where execution duration display sometimes uses inconsistent units.

  • Fixed a bug where estimated instance cost is shown inconsistently at launch time and in the execution info panel.

  • Fixed a bug in the Omics Data Assistant where a "prompt is too long" failure from the underlying model produced a broken or empty final message. The assistant shows the model error instead of summarizing it away, so you can see what happened and shorten the conversation or context.

  • Fixed a bug in the set-password flow where users provisioned by an org admin were shown a token-revocation control during their first sign-in. If users have no prior tokens to revoke, the control is hidden.

  • Fixed an Apollo query failure that could cause some Cohort Browser queries to fail before returning results.

  • Fixed an Apollo cohort-filter failure where filters that test for the presence or absence of a value might not evaluate correctly.

Improvements

Release for 19-May-2026

Improvements

  • Copy to Project in Omics Data Catalog is faster and more reliable when you copy a larger number of files into a destination project. It also preserves the source folder structure during the copy, which keeps copied data organized and consistent with the source.

  • dxCompiler 2.16.0 adds support for AppInsufficientResourceError in executionPolicy.restartOn, so compiled workflows can use automatic recovery from insufficient-resource failures.

Improvements

Release for 12-May-2026

This release includes a small usability improvement in the project creation dialog and routine infrastructure maintenance that does not change documented user workflows.

Improvements

  • When you create a new project, the default project name is fully selected in the New Project dialog, so you can immediately type a new name instead of editing the pre-filled text.

New releaseImprovementsFixes

Release for 05-May-2026

Protect projects from accidental deletion

You can protect projects from accidental deletion by enabling the destroyProtected setting. When deletion protection is on, a project cannot be deleted, even by project admins or org admins.

Project admins and org admins can turn this protection on or off in Project Settings or through the API. To delete a protected project, an admin must first remove the protection as a separate step. Deletion protection is optional and off by default for all projects.

For more details, see project data access controls.

nf-core Pipelines: RNA-seq, Sarek, and scRNA-seq

DNAnexus offers three production-ready nf-core pipeline apps available directly from the Tools Library. You can run them from the UI or using dx run:

  • nfcore_pipeline_rnaseq app (with nf-core/rnaseq v3.24.0) handles bulk RNA-seq data from FASTQ or BAM input through QC, trimming, alignment using STAR or HiSAT2, and quantification using Salmon or RSEM.

  • nfcore_pipeline_sarek app (with nf-core/sarek v3.8.1) supports germline and somatic variant calling on whole-genome or targeted sequencing data with callers including GATK HaplotypeCaller, Mutect2, DeepVariant, and Manta.

  • nfcore_pipeline_scrnaseq app (with nf-core/scrnaseq v4.1.0) processes 10x Genomics single-cell RNA-seq data with alignment strategies including STARsolo, SimpleAF/Alevin-Fry, Kallisto/BUStools, and CellRanger.

All three apps run on Nextflow v25.10.4, are pre-configured for the DNAnexus Platform with Docker enabled by default, and write outputs automatically to the job output folder. They are available across all DNAnexus AWS, Azure, and OCI regions. For CLI and API usage, see Running Nextflow pipelines.

Improvements

  • The Omics Data Assistant notifies you when a background task, such as dataset indexing or a long-running query, finishes while you are working in another tab. The browser tab title updates to show that a result is ready, and a badge appears on the Omics Data Assistant button.

  • The Omics Data Assistant conversation panel no longer scrolls while generating a response. When the assistant fails to produce a chart visualization, you can click the Retry button to regenerate the chart without restarting your query.

  • The Omics Data Assistant progress bar no longer shows 100% before dataset overview and sample prompts have finished generating, giving a more accurate picture of indexing status.

Fixes

  • Fixed a bug that caused dataset indexing in the Omics Data Assistant to fail when the dataset was too large. Large datasets that were previously blocked from indexing can be processed successfully.

  • Fixed a bug where partner branding logos and the Back to Log in link were missing on the password reset page when accessed through a partner portal.

  • Fixed a bug where the Create New User button did not appear consistently in the Invite Member modal.

ImprovementsFixes

Release for 28-Apr-2026

This week's release includes usability improvements to the interface and stability fixes across the platform.

Improvements

  • Refreshed UI for the password reset and token authentication pages.

  • Updated the object details display in the info panel for enhanced readability and to allow object details to be displayed in expanded view.

Fixes

  • Fixed a display bug in the Cohort Browser's Genomic Variant tab that showed variant data for cohorts containing no patients with genomic data.

  • Fixed a visual misalignment of the Copy to Clipboard button in the Cohort Browser's Data Preview tab when hovering over a sample ID.

  • Fixed the Run button on the tool details page invisibly overlapping content, which could cause unintended tool launches when clicking on that area.

Improvements

Release for 21-Apr-2026

  • Omics Data Assistant: The assistant can generate charts from query results alongside written responses, with PNG and SVG export.

  • Omics Data Assistant: Faster and more focused responses following a skills-based prompt hierarchy update.

  • Omics Data Assistant: Dataset overview is no longer shown as a placeholder during indexing. It appears only after indexing is fully complete.

  • Omics Data Catalog: Warning icon and tooltip added for Record ID fields that reference a non-existent record.

  • User interface: Type filter in Data Manager provides type suggestions based on the current data scope and accepts multiple values.

Improvements

Release for 14-Apr-2026

  • Omics Data Assistant: The assistant is available in projects with download restrictions

  • Omics Data Assistant: Referenced cohort details in conversations are collapsed by default and can be expanded on demand

  • User interface: Organization admin page layout improved for smaller screens

  • User interface: The info panel supports recursive expand and collapse of JSON content at any depth level

ImprovementsFixes

Release for 31-Mar-2026

  • Two-factor authentication (2FA): Disabling 2FA automatically revokes all active sessions

  • Omics Data Assistant: The thinking process panel shows a compact, high-level reasoning summary

  • User interface: Breadcrumbs navigation bar added to the job runner page

  • User interface: Improved job log viewer performance when displaying large messages

  • Fixed incorrect storage accounting when projects are transferred between billing organizations

Improvements

Release for 24-Mar-2026

  • Session revocation is active for all security events, including password changes, email updates, and MFA changes

  • Omics Data Assistant: Referenced cohorts can be passed into conversations for contextual queries using the Ask About Cohort button

  • Omics Data Assistant: Datasets containing multiple assays indicate which assay is in use

New releaseImprovements

Release for 17-Mar-2026

Automatically recover jobs from insufficient-resource failures

Jobs that fail due to out-of-memory or out-of-disk conditions can recover automatically by restarting on a larger instance. When a job fails with AppInsufficientResourceError, the platform detects the failure, restarts the job on the next-tier instance within the same instance family, and continues retrying until the job succeeds, the retry limit defined in the restart policy is reached, or no larger instance is available in the family.

To use this capability, enable the Instance Upgrade on Job Restart policy for your organization and configure executionPolicy.restartOn for AppInsufficientResourceError (or "*") in your run specification's execution policy.

This capability helps identify insufficient resource errors, reduces cascading failures, cuts down on manual reruns, and improves overall pipeline reliability.

Improvements

  • User interface: Files display a warning when they contain suspicious content

  • User interface: "New Project" dialog project name input fields improved

Improvements

Release for 10-Mar-2026

  • User interface: Cookies information page added

  • Cohort Browser: Copy Cohort action is more accessible

  • The mem3_ssd1_gpu_x8, mem3_ssd1_gpu_x32, and mem3_ssd1_gpu_x64 NVIDIA V100 GPU instances are no longer available for new jobs

  • Apollo: Direct access to the Thrift Server for SQL queries has been removed. To run SQL queries against Apollo data, launch your own Spark cluster.

Improvements

Release for 03-Mar-2026

  • Omics Data Assistant: Users can rate assistant messages to provide feedback

  • Omics Data Assistant: Conversations are automatically named when created

  • User interface: Tool detail page sections can be expanded and collapsed

  • User interface: Custom table columns are sortable alphanumerically

  • User interface: Breadcrumbs truncate and abbreviate long path names

  • User interface: Folder name validation message is visible for top-level folders

  • Cohort Browser: The interaction for ANY/ALL filters is more intuitive

New releaseFixes

Release for 24-Feb-2026

  • Omics Data Catalog is available. The catalog is a metadata management system that makes research data discoverable across projects and organizations.

  • Fixed a display issue where the mem2_ssd2_gpu1_v2_x32 instance type showed incorrect storage specifications on the organization billing page.

New releaseImprovementsFixes

Release for 10-Feb-2026

  • Omics Data Assistant is available. The assistant is a GenAI-powered conversational interface integrated into Cohort Browser, so you can explore and analyze biomedical and clinical datasets using natural language. You can define cohorts by describing criteria in plain English, request AI-generated dataset overviews, and ask follow-up questions.

  • User interface: Affected sessions are shown when changing password or updating account email

  • Cohort Browser: Fixed Locus Details page showing incorrect locus name in some situations

Improvements

Release for 03-Feb-2026

  • Creating or updating applets and apps with retired instance types returns a validation error

New releaseImprovements

Release for 27-Jan-2026

  • Dynamic Instance Type Selection is available for apps and applets. By specifying instanceTypeSelector in the run specification, developers can define an ordered list of preferred instance types and the platform automatically selects the first available one, enabling graceful fallback when spot instances are unavailable.

  • Cohort Browser: "Without" filter added for the Germline Variant Assay

  • Cohort Browser: Boxplot visualization fixed to include occurrence counts in grouped responses

  • The /project-xxxx/describe API method no longer returns retired instance types

ImprovementsFixes

Release for 13-Jan-2026

  • The /org-xxxx/describe API method returns an expiresAt field. For details, see how to check org expiration status.

  • Cohort Browser: log₂ and log₁₀ axis transformations are available for histograms and box plots, which help visualize data with wide value ranges or skewed distributions

  • Cohort Browser: Data Preview table displays human-readable meaning labels for coded fields

  • Cohort Browser: The Save Cohort action is more accessible

  • Cohort Browser: Fixed missing assay ID on locus details pages in specific cases

Last updated

Was this helpful?