NIH Grant Formatting Rules That Cause Desk Rejections

If your NIH grant application is administratively returned without review, a formatting violation was likely the reason. This is not a suggestion to reformat later — it is a complete loss of the funding opportunity that year. Once an application is returned, you cannot resubmit it for the same deadline.

The good news is that formatting compliance is mostly mechanical. It does not require insight or originality. The bad news is that the rules are scattered across multiple documents, they have changed in recent years, and the most expensive mistakes are often the ones that feel like minor details: a font that is one point too small, a figure legend that tips the page count over the limit, a required section that nobody told you was required.

This post covers the formatting rules that matter, the electronic submission systems that validate them (and sometimes contradict each other), the institutional timelines that determine whether your application makes the deadline at all, and the pre-submission checklist that PIs most often skip.

What counts as a formatting violation that triggers administrative return?

NIH publishes a detailed list of administrative defects in the relevant Funding Opportunity Announcement (FOA) for your grant mechanism (R01, R21, R03, etc.). The most common grounds for administrative return include:

• Font size below the minimum. The default minimum is 11 point for most fonts (Arial, Palatino, Georgia, Times Roman). A few fonts have exceptions — Computer Modern and TeX Gyre Termes can be 10 point — but these are specific and you cannot assume your font qualifies. The SF424 guide is the authority.
• Margins below 0.5 inches. All four margins must be at least half an inch. This includes margins around tables and figures.
• Exceeding the page limit for a section. The Research Strategy for an R01 is 12 pages. For an R21 it is 6 pages. Specific Aims is always 1 page. Every element on the page — figure legends, table titles, footnotes — counts toward the limit. Many PIs discover at the last minute that their figures pushed them over.
• Missing a required section. The most frequently omitted section is "Authentication of Key Biological and/or Chemical Resources" when your research uses them. Other commonly forgotten sections include Resource Sharing Plans, Multiple PI Leadership Plans (for multi-PI applications), and Select Agent Research disclosures (if applicable).
• File naming or format errors. Application files must follow a specific naming convention. PDF corruption or wrong file types will trigger administrative return.
• Biosketch formatting violations. Biosketches must be formatted according to the latest SciENcv template. Hand-written or outdated formats will be returned.

Font and formatting requirements — the details that matter

The font rules are more specific than most PIs realise, and they change with each application cycle. The current baseline is:

• Minimum 11 point for Arial, Calibri, Cambria, Courier New, Georgia, Helvetica, Palatino, Times New Roman, and Verdana.
• Minimum 10 point for Computer Modern and TeX Gyre Termes (the TeX fonts).
• No font smaller than 10 point, ever. Even if you think it is allowed, do not use it — the grant office will flag it.
• Fixed-width fonts must be at least 10 point. If you include code or technical output, increase the size by one point to be safe.
• Avoid decorative or obscure fonts. If it is not on the approved list, it will trigger a formatting check.

Margins are equally strict. All four margins must be exactly 0.5 inches or greater. This includes the space inside tables and figure borders. A common mistake is assuming that Word's default margins (usually 1 inch) apply — they do, but tables often override them, and when a table margin falls below 0.5 inches, the entire document fails the check.

Line spacing does not have a minimum, but 1.5-line spacing or double-spacing is standard in grant writing and improves readability for reviewers. Single spacing is technically compliant but makes your document harder to read on screen, which matters when reviewers have 50 applications to evaluate.

Page limits by section — and what counts toward them

The most expensive mistake is not knowing what counts toward a page limit. Here is the breakdown:

The phrase everything counts is load-bearing. Figure legends are not footnotes — they count. Table titles count. Footnotes count. Callout boxes and text callouts in figures count. A figure that is technically within the bounds of the page but has a caption that extends into the margin counts as an entire page. When you are close to the limit, count every word, including the captions.

Required sections that PIs frequently forget

The NIH publishes which sections are required in the relevant FOA, but many PIs do not read it carefully. Here are the sections most often omitted:

Authentication of Key Biological and/or Chemical Resources

If your research uses key biological resources (cell lines, organisms, organisms with special properties), chemical resources (compounds, reagents), or both, this section is required. It must describe how you will authenticate them — that is, verify they are what you claim they are. If you work with cell lines, you must state how you will authenticate them (STR profiling, genomic sequencing, etc.). If you use commercially purchased reagents, you typically describe the vendor and lot number. This section is not optional, and its omission is a compliance failure that can trigger administrative return.

Resource Sharing Plan

If your research generates data, software, materials, or other resources that would be of interest to the research community, NIH expects you to share them. You must describe how and when. If your answer is "we will not share," you must justify why (e.g., due to patient privacy, IP constraints, or proprietary agreements). A blank Resource Sharing Plan is a red flag.

Multiple PI Leadership Plan

If your application involves multiple PIs (a growing model in biomedical research), you must include a leadership plan describing how the PIs will coordinate, divide responsibilities, and make decisions. A one-paragraph statement is often sufficient, but it must be present.

Select Agent Research

If your research involves a select agent (a list of pathogens and toxins regulated by the CDC and USDA for national security reasons), you must disclose this and describe the biosafety and security measures you will use. If you are unsure whether your research involves a select agent, check the CDC and USDA lists — surprises at the last minute are expensive.

Each of these sections is typically 1–2 pages, and each is required when applicable. Missing any of them gives the grants office grounds to return your application.

Electronic submission: three systems, three validation failures

Most PIs submit through Grants.gov, assuming that if the application passes validation there, it is done. This is incorrect. There are three validation systems, and they do not always agree.

Grants.gov validation

This is the federal submission portal. When you submit, Grants.gov validates the file structure, file names, and format. If validation passes, you get a receipt number and feel like you are done. You are not done.

eRA Commons validation

eRA Commons is NIH's internal system where grants office staff and NIH program officers see your application. An application that passes Grants.gov validation can still fail at eRA Commons. Validation errors at eRA Commons often do not surface for hours or days — sometimes not until the day after the deadline. By then it is too late to resubmit.

ASSIST

ASSIST is an internal NIH system that cross-checks applications against additional criteria. Some validations that pass Grants.gov and eRA Commons will fail ASSIST. ASSIST errors are rare but catastrophic because they often appear after you think your application is complete.

The practical implication: Do not wait until the deadline to submit electronically. Submit at least 2–3 days early, then check all three systems (or ask your grants office to do so) to confirm everything passed. If eRA Commons validation fails, you will have time to fix it and resubmit before the deadline.

The institutional submission timeline — the number one deadline that PIs miss

Every institution has an internal deadline — typically 5–10 business days before the NIH deadline — when your application must be in the hands of the grants office. This is not a suggestion. It is a hard gate.

Here is why: your institution's grants office does not process applications the moment they arrive. They are queued. When your application enters the queue, the office performs its own compliance check, confirms institutional information (your effort commitment, institutional signatures, cost-share if applicable), and then submits electronically. If you submit to your grants office on day 8 (three days before the NIH deadline) and there are 40 applications ahead of you, yours might not make it into the electronic queue until day 9. If NIH deadline is day 10, your application misses it entirely.

Late institutional submissions are the most common reason applications fail to make the deadline — more common than electronic submission errors, more common than formatting violations, more common than anything else. The solution is simple: mark your calendar for your internal deadline (call your grants office if you do not know what it is) and treat it as more urgent than the NIH deadline itself.

Notice the gaps. Notice that you are not in control of most of the timeline. Your job is to hit the institutional deadline early enough that your grants office has time to queue, review, and submit without rushing.

PDF preparation — the details that fail silently

A PDF that looks perfect on your machine can render differently on a reviewer's machine. The most common culprits are:

Font embedding

If a font is not embedded in your PDF, it will substitute with a different font on any machine that does not have the original installed. This can break carefully laid-out figures, push text over margins, or render mathematical symbols incorrectly. Always embed fonts when you export to PDF. In Microsoft Word, this is under File > Options > Trust Center > Trust Center Settings > Privacy Options > (check "Disable all untrusted document properties").

Figure resolution

Figures at 72 dpi (screen resolution) will look pixelated when printed or displayed on a high-resolution monitor. Use 300 dpi minimum for figures, especially if they contain fine detail (graphs, gel images, microscopy). A figure that looks fine in your document might be illegible to a reviewer on a different display.

Figure placement and overflow

A figure that appears on the correct page in Word might shift when the PDF is opened or printed due to font substitution or margin interpretation. Test your PDF carefully: open it on a different machine, zoom in and out, and confirm every figure is where you expect it. If a figure overflows into an adjacent page, the page count increases.

File size

NIH limits PDF file size to 100 MB per attachment. If your application contains high-resolution figures, it might exceed this. Compress image resolution to 150–200 dpi (below the 300 dpi recommendation but acceptable for complex multi-figure documents) or reduce the number of figures per page.

File naming

Grants.gov and eRA Commons require specific file names. A mismatch will trigger rejection during validation. Confirm the exact file naming convention in your FOA (typically something like "Nih_R01_ResearchStrategy_LastName.pdf") and follow it exactly. Do not deviate, even slightly.

The pre-submission checklist

Run through this checklist at least one week before your institutional submission deadline. Print it and check off each item as you go.

The less obvious rules that trip up experienced PIs

The expiry date on the funding opportunity announcement

Funding Opportunity Announcements (FOAs) have explicit expiry dates. If you submit to an expired FOA, your application will be administratively returned. This sounds obvious until you realise that some FOAs are re-issued annually with minor changes but the same URL, and the NIH website sometimes shows the old version first in search results. Always check the publication and expiry dates on any FOA you use.

Page count is measured from the PDF, not the source document

If your Word document shows 12 pages with single-spaced text, but your PDF renders as 13 pages due to font embedding or margin interpretation, the PDF count is what matters. Always measure from your final PDF, not from the Word file.

Effort commitment must match across all forms

Your effort commitment (the percentage of time you will spend on the project) must be consistent across the biosketch, the Research Strategy, the project summary, and any institutional forms. A mismatch triggers an administrative hold.

Budget must match justification

If your budget requests $50,000 for equipment but the justification describes renting equipment for $5,000, that mismatch will be flagged during the grants office review. Ensure consistency across all budget sections and the narrative justification.

Biosketches must use the current SciENcv template

NIH updated the biosketch format recently, and older formats are no longer accepted. If you have a biosketch template from more than two years ago, it is likely outdated. Always generate biosketches using the current SciENcv tool, never from an old template or a personal résumé.