NIH Oligonucleotide Toxicity (OligoTox) Open Data Challenge

NIH Oligonucleotide Toxicity (OligoTox) Open Data Challenge

The Oligonucleotide Toxicity (OligoTox) Open Data Challenge is a two-phase NIH challenge run by the National Center for Advancing Translational Sciences (NCATS), with a total prize pool of up to $500,000 across both phases. The challenge is designed to stimulate the generation of publicly accessible, high-quality datasets from in vitro human systems to improve machine-learning and other in silico methods for predicting oligonucleotide therapeutic toxicity.

The page is structured as a prize competition, not a grant, not an appointment, and not a fellowship. The practical consequence is important: this is a deliverable-driven submission model where applicants are judged on how well they can define, generate, document, and open-share datasets and methods that improve predictive toxicology workflows.

As of the last NIH review update in May 2026, Phase 1 winners have already been announced, but Phase 2 remains open to new and existing participants from May 1, 2026 through December 31, 2026, with winner announcements expected in March 2027. That makes it a relevant 2026/2027 opportunity, especially for teams that can mobilize a complete dataset pipeline during the second window.

Key details

What this challenge is for, and what it is not

OligoTox is specifically about data quality and public usability, not about proposing a concept alone. The challenge asks teams to contribute usable evidence that can directly improve models of oligonucleotide toxicity.

The challenge has two phases:

• Phase 1 Ideation: propose a defensible dataset strategy, including indicators, model system choices, controls, predictor features, and utility for downstream predictive modeling.
• Phase 2 Data Generation: collect, generate, or contribute a real dataset in the target format and quality band, with documentation and public dissemination plans.

That structure means Phase 1 success depends heavily on rigor in experimental design framing and data strategy clarity, while Phase 2 success depends on execution quality and full package compliance.

Important distinction: this is a prize challenge:

• There is no budget request form like R01/R01-equivalent submissions.
• There is no direct funding allocation for labor categories, fringe rates, and facilities in the typical grant sense.
• Your outputs are judged on dataset value, reproducibility, scientific relevance, and openness.

Given this, the right participants are usually one of these combinations:

1. research groups with in-house wet-lab and assay capacity who can generate reliable toxicology signals,
2. teams pairing modelers, domain experts, and data engineers who can convert assays into FAIR-ready machine learning inputs,
3. institutions with good compliance workflows able to manage licensing and public access documents.

Who should apply: practical fit criteria

The official rules include several eligibility gates that create a practical shortlist.

Required fit categories

The challenge permits participation by individuals, teams, or entities. It is intentionally broad, but the strongest fits are:

• institutions that already operate oligonucleotide assay pipelines,
• teams with prior experience in dataset design (especially toxicity, omics, assay metadata, and model-grounded features),
• participants who can provide open data artifacts with clear provenance and legal sharing conditions.

This challenge is particularly suitable if your team already has:

• a clear plan for toxicity endpoint definitions (for example hepatotoxicity, kidney toxicity, thrombocytopenia, complement activation, coagulopathy, immunotoxicity, chronic neurotoxicity, or hydrocephalus),
• a realistic source of sequence and chemistry diversity for oligo testing,
• capacity to produce metadata-rich outputs and consistent controls.

Geographic and eligibility boundaries

For entities and citizens of record, location rules are straightforward:

• For private entity participation the organization must be incorporated and have a primary place of business in the U.S.; this is a central eligibility condition.
• For cash-prize eligibility the individual must be a U.S. citizen or permanent resident.
• Team participation may include non-U.S. scientists in supporting roles, but non-citizen/non-resident participants are not eligible to receive cash from the winner payout.
• Minimum age: 18.

This makes it a challenge that can still support global collaboration but with a clear U.S.-centric funding outcome. If your group expects to include international partners, plan financial distribution and IP language early so there is no confusion at award stage.

Non-obvious participation constraint: federal funding sources

If your work uses federal grants/cooperative agreements/contracts, you cannot simply assume you can repurpose those budgets for challenge execution. Rules specifically require consistency with the terms of that federal support and, in some cases, handling winners’ disbursements as program income in compliance with federal award rules.

This is one of the strongest sources of late-stage disqualification in many challenge contexts: technically strong submissions can be invalidated by a compliance mismatch if documentation is late or misaligned. Treat this as part of your initial eligibility filter, especially if you are a subawarded team.

Timeline and status you should plan around

The full published timeline contains both phases and was useful to convert into a planning schedule:

• Phase 1 launch: September 30, 2025.
• Phase 1 submission: 12/19/2025 → 02/28/2026.
• Phase 1 judging: early March to mid-April 2026.
• Phase 1 winner announcements: 04/30/2026.
• Phase 2 launch: 05/01/2026.
• Phase 2 submission: 05/01/2026 → 12/31/2026.
• Phase 2 judging: 01/08/2027 → 02/28/2027.
• Phase 2 winner announcements: 03/15/2027.

As of this page’s review date, Phase 1 is closed and Phase 2 is the active competitive window. If your team is already through a similar ideation cycle, Phase 2 still allows direct data submissions without requiring Phase 1 entry.

Phase 1 vs Phase 2 deliverables at a glance

• Phase 1: conceptual + technical design + data plan + controls + replicates + collection logic.
• Phase 2: complete dataset packet: narrative, methods, public access plan, and the dataset itself.

A practical planning insight: Phase 2 is easier only if you treat it like a grant-like production plan with deliverables, timeline, and review criteria mapped at week-level granularity. You do not want to discover near the end that your format or metadata schema is incomplete for FAIRness and reproducibility.

What you must submit, and how evaluation works

The submission structure is strict. You are not graded by intent; you are graded by what you file.

Required eligibility package pieces

All phases require registration completion and inclusion in submission package.

The challenge lists explicit requirements including:

• Completed registration form from NCATS challenge page,
• clear participant type and responsible point of contact,
• complete submission package by phase,
• materials directly aligned with required page limits and formatting.

Phase 1 submission materials

For Phase 1, participants must explain:

• which toxicity indicators and predictor variables they target,
• existing public data landscape and gap analysis,
• why the new data materially adds to public resources,
• planned data source and ownership constraints,
• model system choice,
• toxicity readouts,
• sample size strategy,
• controls and validation logic,
• and methods for oligo identity verification.

The package should be concise and complete; for Phase 1 the page limit is 10 pages plus references.

Phase 2 submission materials

Phase 2 is heavier. NIH expects a package with:

• a narrative PDF (up to 12 pages),
• a methodology PDF (up to 5 pages),
• Public Access and Dissemination Plan (up to 5 pages),
• the dataset package with metadata schema, raw/access instructions, and sequence/modification context for each oligo.

There is no page limit for raw dataset files, but there is still a quality ceiling in evaluation: reproducibility and usability.

Scoring logic and review criteria

The review criteria are published and are not generic language. They include:

• scientific value,
• experimental approach,
• experimental design quality,
• data translatability,
• dataset management and documentation,
• and for Phase 2, explicit strength of PADP execution plan.

Confidence ratings (low/some/high) are applied by criterion, and this makes a direct difference when judging quality. The practical lesson: strong methods without metadata maturity can underperform a technically moderate dataset with better FAIR and governance.

PADP and openness are not optional

Public Access and Dissemination is central. Participants must include licensing terms and fallback plans so external users can access and use the dataset if the original winner is unable to support continued access. This requirement is often overlooked in teams that focus only on assay throughput.

If your team has an IP policy but no ready public-access path, Phase 2 submission becomes an administrative risk even if model quality is high.

Submission process in practice

NIH/NCATS process details for this challenge are unusual:

• Registration and submissions are made through email (not Grants.gov, not an NIH proposal system portal).
• Email used is the official challenge address; teams should send complete packages to the specified mailbox.
• Registration form links are maintained on the NCATS challenges page.

Because the process is email-led, practical execution risk is highest around inbox timing and version control:

1. Use one submission version per phase.
2. Use consistent filenames with phase/date/version tags.
3. Include every required attachment in the first submission packet.
4. Preserve PDF readability and page limits exactly.
5. Send with enough time for internal review and timezone coordination.

If you need a concrete template, build internal folders like:

• phase2-narrative-v1.pdf
• phase2-methods-v1.pdf
• phase2-padp-v1.pdf
• phase2-dataset-metadata-schema
• phase2-raw-data.zip

and keep a “submission manifest” that confirms every required piece.

Common mistakes seen in these challenges

Even high-quality teams trip over recurring errors:

• submitting incomplete documents (especially registration form omission),
• missing page format/limit constraints,
• leaving ambiguity in participant type (individual/team/entity) and winner payment rules,
• underbuilding metadata (no schema, missing chemistry positions, weak provenance fields),
• insufficient control design,
• missing positive/negative controls,
• PADP too thin or missing fallback licensing scenario,
• assumptions about federal funding usage not checked against award conditions,
• and late submission timing with last-minute package reformatting.

The strongest mitigation is a pre-submission compliance checklist with hard owners: one person handles scientific design, one handles documentation and formatting, one handles legal/compliance, and one owns submission mailbox coordination.

Reviewers’ expectation model (what to optimize first)

For Phase 2 judges, this challenge behaves like a practical utility test. They look for:

• whether the dataset fills an actionable gap,
• whether indicators and predictor variables are selected with biological and modeling logic,
• whether there is enough quality signal for downstream model improvement,
• whether controls and replicates are adequate,
• whether files are complete enough that another group can reproduce usage,
• whether the dissemination plan is legally and operationally usable.

If a team does only one thing early, it should be to lock a clean data dictionary with stable naming, units, and provenance fields. Most of the judging criteria point back to this.

Who this challenge is especially not for

It is not ideal if your team:

• cannot generate or access oligo testing data in a realistic timeframe,
• cannot produce FAIR-aligned documentation,
• is mostly concept-driven with no concrete dataset execution path,
• relies on unpublished commercial data with unresolved sharing restrictions,
• is only looking for small symbolic recognition and not a practical submission.

If your organization depends on strict NDA restrictions that limit open dissemination, this challenge is typically not a fit unless your legal path to licensing is already clear before submission.

FAQ

Is this open to teams outside the United States?

Participation can include non-U.S. collaborators in teams, but only U.S. citizens/permanent residents and U.S.-based private entities are eligible to receive cash awards. Non-U.S. team members can still contribute scientifically to the submission.

Can non-NIH teams apply as entities?

Yes, if the entity is properly formed in the U.S. and aligned with the eligibility rules.

Is participation limited to universities?

No. Individuals, teams, and entities can all participate if they meet rules and submission requirements.

Can you use federal grant funds to support challenge preparation?

Generally only where consistent with the underlying award terms; if using federal funds, teams must follow applicable federal award compliance, including treatment of any prize-related funds as program income where relevant.

Are Phase 1 and Phase 2 independent?

Yes. Phase 2 participation does not require winning Phase 1, and participants can submit directly in Phase 2 if they have a qualifying dataset package.

What is the most important operational step before submission?

Confirm all required fields for participant, registration, licensing, metadata, and documentation are in one packet. In this format, an otherwise strong dataset can be disqualified by missing paperwork.

Practical preparation plan for Phase 2 (high-trust version)

Weeks 1–2: Compliance and structure

• Confirm team roster and roles.
• Lock legal and compliance position for data release.
• Draft a minimal data dictionary and define variable naming conventions.
• Confirm eligibility edge cases for each participant.

Weeks 3–5: Experimental and curation buildout

• Finalize toxicity endpoints and controls.
• Define sampling and replicate plan.
• Start raw data logging in a standardized schema.
• Establish versioning for sequence and chemistry metadata.

Weeks 6–8: Documentation engine

• Draft narrative and methods sections simultaneously.
• Prepare PADP language with fallback scenarios for access and licensing.
• Convert to required format and check page limits.

Weeks 9–10: Dry run and compliance audit

• Run an internal review against every published Phase 2 rule.
• Verify formatting, attachments, and signatures.
• Confirm email submission format and subject details.
• Finalize and submit early in deadline window.

Official links

• Official challenge page (primary): https://www.nih.gov/challenges/oligonucleotide-toxicity-oligotox-open-data-challenge
• NCATS challenge portal (registration form and linked challenge info): https://ncats.nih.gov/funding/challenges
• Contact: [email protected]

Final note

This challenge is worth tracking if your team works at the intersection of oligonucleotide therapeutics, toxicology data, and predictive modeling. It is most valuable for teams that can produce a dataset package that is scientifically useful beyond a single lab benchmark and clearly open to external use.

The page indicates a strong practical orientation: the judges are effectively asking for reusable, transparent toxicology resources, not just promising ideas. If your team can deliver both data quality and public-access readiness, this challenge can convert effort into both visibility and concrete prize outcomes.