Help me find NIH grants for my research idea: [DESCRIBE IDEA IN A FEW SENTENCES] This focuses on NIH only (institutes, mechanisms, NOSIs, RePORTER). Not PCORI, DOD CDMRP, VA Merit, or foundations. Before searching, ask me three quick questions: 1. Career stage (determines mechanism fit): - Trainee (pre-doc / postdoc) → F31, F32, T32 - Early-career faculty (pre-tenure, no major grants) → K01, K08, K23, K99/R00, R21 - Mid-career (have or had a K or first R01) → R01, R21, R34 - Senior investigator → R01, R35, P01, U01 2. Preliminary data status: - Starting from scratch (no pilot data) - Have some pilot data or related publications - Resubmission — previously scored - Resubmission — previously triaged 3. Institutional environment: - Major research university (R1) with core facilities - Academic medical center / teaching hospital - Smaller institution / community-based - VA or military-affiliated Then run the analysis in this order: ═══════════════════════════════════════════════ PHASE 1 — POSITIONING ANALYSIS (5 Consensus searches) ═══════════════════════════════════════════════ The goal is NOT "is this novel?" — it's to produce draft Significance/Innovation language for the grant. Run 5 Consensus searches sequentially (1 query/sec rate limit), all with year_min: (current year - 6): 1. **What's established** — query the core concept. "What does the field agree on?" 2. **The stakes** — disease/condition prevalence, burden, outcomes, disparities in the target population. 3. **Current approaches & their limits** — clinical management, interventions, guidelines today. 4. **Method in adjacent contexts** — has this approach been applied elsewhere? Validation or whitespace. 5. **The gap (most important)** — explicitly search for "research gaps limitations unmet needs future directions" in the field. These quotes become the backbone. After all 5 complete: - Deduplicate papers by title/DOI across searches. Papers hitting multiple facets are high-signal. - **Extract 3-5 gap quotes** from abstracts — language like "future research should," "remains unclear," "no studies have," "limited evidence," "critical gap." These are the centerpiece. - Write a 2-3 paragraph **positioning narrative** the researcher can adapt directly: - ¶1: "The field has established [X, Y, Z]. The burden is significant because [stakes]." - ¶2: "Current approaches include [findings], but fall short because [gap quote 1] and [gap quote 2]." - ¶3: "This project proposes [the idea]. [Adjacent-context findings] suggest feasibility, and the gap shows it hasn't been applied to [this population/context]." ═══════════════════════════════════════════════ PHASE 2 — INSTITUTE & GRANT MAPPING (NIH RePORTER) ═══════════════════════════════════════════════ The NIH RePORTER API requires POST requests. Use bash_tool with curl, not web_fetch. Endpoint: `https://api.reporter.nih.gov/v2/projects/search` (no auth needed). Run two searches — narrow (AND) and broad (OR) — across the last 4 fiscal years. Computed dynamically: ```bash CURRENT_YEAR=$(date +%Y) FY_RANGE="[$((CURRENT_YEAR-3)), $((CURRENT_YEAR-2)), $((CURRENT_YEAR-1)), $CURRENT_YEAR]" curl -s -X POST "https://api.reporter.nih.gov/v2/projects/search" \ -H "Content-Type: application/json" \ -d '{ "criteria": { "advanced_text_search": { "operator": "and", "search_field": "projecttitle,terms,abstract", "search_text": "3-4 NARROW KEYWORDS" }, "fiscal_years": '"$FY_RANGE"', "include_active_projects": true }, "limit": 20, "offset": 0, "include_fields": ["ProjectTitle","AbstractText","ActivityCode","AwardAmount","FiscalYear","PrincipalInvestigators","Organization","OpportunityNumber","AgencyIcAdmin","ProjectNum","StudySection"] }' ``` Then run a broad version with `"operator": "or"` and 5-6 related terms. Combine + dedupe by project_num. From combined results, extract four things: **A. Institute targets** — tally `agency_ic_admin.abbreviation`. Top 2-3 = primary/secondary institutes. → "NIAMS is funding 8 projects in this space, NIA is funding 5 → primary target: NIAMS, secondary: NIA" **B. Study sections** — tally `study_section`. Top 2-3 panels with project counts. These are the review panels evaluating this work — most applicants don't think about this strategically. **C. NOSIs** — extract `opportunity_number` values starting with `NOT-`. Fetch each at `https://grants.nih.gov/grants/guide/notice-files/{NOSI_NUMBER}.html` (one at a time). Pull title, institute, purpose, deadline. **D. FOAs** — extract `opportunity_number` values starting with `PA-`, `PAR-`, `PAS-`, `RFA-`. Construct links: - PA/PAR/PAS: `https://grants.nih.gov/grants/guide/pa-files/{FOA}.html` - RFA: `https://grants.nih.gov/grants/guide/rfa-files/{FOA}.html` ═══════════════════════════════════════════════ PHASE 3 — MECHANISM MATCHING ═══════════════════════════════════════════════ Match mechanism to BOTH career stage AND project scope (size, prelim data). Reference: | Mechanism | Budget | Duration | Best for | Prelim data | |-----------|--------|----------|----------|-------------| | F31/F32 | Stipend + tuition | 1-3 yrs | Pre-doc / postdoc trainees | Some | | T32 | Institutional training | 2-5 yrs | Trainees via mentor's program | N/A | | R03 | ~$50K/yr direct | 2 yrs | Small self-contained / pilot | Minimal | | R21 | ~$275K total direct | 2 yrs | Exploratory/developmental | Minimal-moderate | | K01/K08/K23 | $100-250K/yr (incl. salary) | 3-5 yrs | Early-career mentored | Moderate | | K99/R00 | K99 ~$100K/yr; R00 ~$250K/yr | 2+3 yrs | Postdoc → faculty transition | Moderate | | R01 | ~$250K+/yr modular | 3-5 yrs | Full research project | Strong | | R34 | ~$450K total direct | 3 yrs | Clinical trial planning (no execution) | Moderate | | R61/R33 | Phased | 2+3 yrs | High-risk, milestone-gated | Moderate for R61 | | R35 | ~$750K/yr direct | 7 yrs | Outstanding investigator, broad program | Extensive | | P01 | $1M+/yr, multi-component | 5 yrs | Multi-PI program | Extensive | | U01 | ~$500K+/yr | 3-5 yrs | Cooperative with NIH involvement | Strong | | DP1/DP2 | ~$500K/yr | 5 yrs | Director's awards, bold ideas | Varies | Filter by career stage, then by scope. "Starting from scratch" → R21, R03, R61. "Have pilot data" → R01 realistic. "Resubmission" → match original mechanism unless score suggests mismatch. ═══════════════════════════════════════════════ OUTPUT (in chat — no .docx) ═══════════════════════════════════════════════ **1. Executive Summary** — 3-4 bullets: novelty level, top institute, top grant, key differentiator. **2. Your Positioning in the Field** - Lead with the 3-5 gap quotes (italicized, with linked citations). These are the most valuable output. - The 2-3 paragraph positioning narrative. **3. Where NIH Is Funding This Work** — institute ranking with project counts + 2-3 sentence interpretation. **4. Recommended Funding Opportunities** - If any NOSI found: bold callout at the top. - Top 3 grants: mechanism, FOA (linked), institute, budget/duration, fit. - For each: one short paragraph on why it fits the *project scope* and prelim data situation. **5. Similar Funded Research** — top 5 projects (PI, institution, award, year, link to reporter.nih.gov/project-details/{project_num}), plus a differentiation paragraph: "Your idea differs from existing funded work in [2-3 specific ways]." **6. Where This Work Gets Reviewed** — top 2-3 study sections with fit notes. Verify rosters at https://public.csr.nih.gov/studysections. **7. Strategic Recommendations & Next Steps** - 3-4 numbered recommendations. - **Always include a program officer recommendation** — single most valuable advice for any applicant. "Contact the PO at [institute] before submitting. POs confirm fit, suggest the right FOA, flag issues. Find contacts at https://www.nih.gov/institutes-nih/list-nih-institutes-centers-offices → [institute] → Staff." - If resubmission, include guidance on framing the response to reviewers. - **Submission timeline:** | Mechanism | Standard receipt dates | |-----------|----------------------| | R01, R21, R03 | Feb 5, Jun 5, Oct 5 | | K awards | Feb 12, Jun 12, Oct 12 | | R34, R61/R33 | Feb 16, Jun 16, Oct 16 | | F31, F32 | Apr 8, Aug 8, Dec 8 | Note which upcoming cycle is realistic given prelim data status. **8. References** — every paper cited, with clickable Consensus URLs. ═══════════════════════════════════════════════ RULES ═══════════════════════════════════════════════ - Only cite papers Consensus actually returned in this session, and only fund/FOA data RePORTER returned. If you include reference info (budget ranges, dates), label "[reference, not from tool call]." - If a search returns sparse results, say so — don't pad with model knowledge. - Consensus: typical caps are ~3 (unauthenticated), ~10 (free), ~20 (premium). If response says "showing top N of M," report it — sparse results may be a plan limit, not a literature gap. - Use full URLs, never truncated. - If no NOSIs surface, skip that callout — don't fabricate. - Career stage filtering is a signal, not a hard filter — flag if a grant is a stretch. - This is a starting point. As specific aims sharpen, run again — the positioning narrative gets sharper with a more specific idea.