P4 Batch 1 Ingest Report
Date: 2026-05-12 Tier: P4 — post-2020 peer-reviewed, high-evidence Handles processed: 200 (sorted year-descending, positions 1–200 of 6,571) Groups: 4 parallel agents of 50 handles each
1. Summary
| Category | Count |
|---|---|
| Handles in this batch | 200 |
| Source pages created | 12 |
| False positives skipped | ~188 |
| Food / food-pathway concentration papers | 7 |
| Biomonitoring / exposure papers | 3 |
| Environmental (agricultural soil) | 1 |
| Review / context papers | 1 |
Cumulative source pages after batch 1: 294
False positive rate: ~94%. The top 200 P4 handles (sorted by manifest year descending) are dominated by papers with OCR year artifacts and by analytical chemistry / materials science papers with incidental metal mentions. The yield improves as ingest moves into papers with accurate 2025/2024/2023 dates.
2. Year Correction — Systematic OCR Artifact
The manifest assigned years of 2026–2029 to the top 200+ P4 handles. These are OCR artifacts from Marker conversion: numbers like 2021, 2023, 2025 in footers, reference lists, or header text were extracted as publication years. The actual years in this batch ranged from 2020–2026.
Impact: The P4 sort order by year-descending is unreliable for the first ~200 entries. The yield will likely improve once ingest enters the 2025 and 2024 papers, which will have more accurate year labels and more food-focused content.
3. Source Pages Created
Food / food-pathway concentration papers
reksten2021-bay-bengal-fish-metals (FM_8160839) Reksten et al. 2021. 24 fish species from Bay of Bengal; n=1,111 fish; ICP-MS at IMR Bergen. Mean Cd 0.19 mg/kg. 58% of small whole-consumed fish (≤25 cm) exceed EU maximum of 0.050 mg/kg Cd. Full target hazard quotient (THQ) and carcinogenic risk assessment. Highest-priority Cd finding in this batch — directly relevant to any product using fish meal, fish powder, or dried small fish from South/Southeast Asian fisheries.
reksten2020-angola-fish-metals (FM_7278876) Reksten et al. 2020. tAs, Cd, tHg, Pb in 5 Angolan marine fish species (n=25 composites); all below EU limits; ICP-MS at IMR Bergen. Baseline comparison for reksten2021.
albuquerque2026-fish-toxic-elements-western-para (FM_12947147) Albuquerque et al. 2026, ACS Omega. 398 fish from 6 species in 5 Amazonian municipalities. Hg exceeds Brazilian limits in most carnivorous species. 25% of samples exceed 10⁻⁴ carcinogenic risk threshold (As dominant). THQ > 1 at local consumption rates. CC BY.
adelusi2024-dairy-feed-south-africa (FM_11167146) Adelusi et al. 2024. As, Cd, Pb all <LOD in 70 South African dairy cattle feed samples (Free State + Limpopo); Cr 0.032–1.459 mg/kg. Supply-chain baseline for dairy products.
abeslami2025-moroccan-honey-minerals (FM_11721970) Abeslami et al. 2025. Cd 0.0017–0.018 mg/kg, Pb 0.13–0.19 mg/kg in 7 Moroccan honey types. Pb values exceed EU recommended limit of 0.10 mg/kg (non-mandatory). First Morocco-jurisdiction honey metals data in corpus.
porwollik2026-rhodiola-supplements-us-market (FM_12810810) Porwollik & Jafari 2026, PLoS One. 10 US-market Rhodiola rosea supplements. All 7 capsular products: detectable tAs (21–393 ppb), Co (18–733 ppb), Pb (9–88 ppb). Tinctures all <LOQ for all metals. tAs not speciated; iAs/tAs split unknown. ICP-MS at Eurofins. CC BY.
ji2026-agricultural-soil-metals-zhejiang (FM_12962467) Ji & Wu 2026, PLOS ONE. 877 agricultural soil samples from coastal eastern Zhejiang (paddy fields 48%); Cr, Pb, Cd, Hg, As by ICP/AFS. Traffic sources dominate (52.5%); max values exceed background 7–13×. Data from 2013 soil survey. Relevant to paddy rice supply-chain contamination context for Zhejiang-sourced rice.
Biomonitoring / exposure papers
taylor2025-seafood-benefits-contaminants (FM_12071223) Taylor et al. 2025. B-tier broad regulatory review of seafood benefits vs. contaminant concerns. No primary data; regulatory context synthesis only.
scovronick2025-glynn-county-exposure (FM_12887143 equivalent) Scovronick et al. 2025, Environmental Pollution. Community biomonitoring near EPA Superfund sites, Glynn County GA (n=96). Pb, Cd, tHg all comparable to US general population. PCB and toxaphene exposures elevated; fishing predicted Aroclor 1268 body burden. CC BY-NC-ND.
lepak2026-mehg-depuration-fish-consumption (FM_12930318) Lepak et al. 2026, ACS EnvHealth. MeHg depuration rate 8.3 ± 1.1 ng/g/day measured in two travelers after fish consumption in Gabon; both exceeded EPA RfD 2–4×. Relevant for fish consumption advisory context.
uzomah2021-nigeria-fish-contaminants (FM_8465269) B-tier review; Pb potentially toxic in fish from industrial/oil-extraction sites in Nigeria. No pooled concentration tables.
arain2026-groundwater-arsenic-dadu-pakistan (FM_12946953) Arain et al. 2026, ACS Omega. 159 groundwater samples from flood-prone Dadu district, Sindh; As up to 500 µg/L (50× WHO guideline); 59.7% E. coli positive. Drinking water pathway; relevant to formula reconstitution water risk for Pakistan.
4. False Positives — By Category
The ~188 false positives fell into these categories:
- Materials science / sensor fabrication (~90): RSC Advances sensor papers (MOFs, photocatalysts, perovskites, supercapacitors), nanomaterial synthesis, photovoltaics. These appeared in P4 because they contain metal names in a fabrication context.
- Drug development / clinical pharmacology (~40): Pharmaceutical compound synthesis, cancer drug design, clinical drug trials. Metal mentions from chelating groups or ligand chemistry.
- Plant physiology / environmental biology (~30): Cd/Pb/As stress responses in plants (spinach, wheat, rice) without human food pathway. Phytoremediation papers without agricultural food-crop link.
- Medical case reports / clinical epidemiology (~20): Cardiac events, neurology, orthopedics, ophthalmology — papers where metal mention is from unrelated biochemical or environmental context.
- Miscellaneous (~8): Fermentation aroma profiles, veterinary medicine, aquaculture nutrition without metals data.
5. New-Page Proposals
No new ingredient, product, or regulation pages proposed. Frontmatter links existing pages or leaves arrays empty where appropriate.
Standing proposals from this batch:
wiki/testing/icp-ms-food-matrices.md— methods page stub; multiple LOD/LOQ values captured across P2 and P4 ingest would populate it.
6. Commits
a0f6a8e— P4 group 3: 1 source page (scovronick2025)a0d9a61— P4 group 2: 1 source page (porwollik2026)313c29c— P4 group 1: 6 source pagesc1d623a— P4 group 4: 4 source pages- (this commit) — P4 batch 1 close: batch report + log entry