🧬 Biohacking Wiki

Evidence Verification & Epistemic Hygiene Guide

4 min read

Evidence Verification & Epistemic Hygiene Guide

Live Citation Audit Status (2026-10-24)

  • Unique PubMed IDs (PMID) Audited via Official NCBI Entrez API: 309 / 309 (100.0% Resolved & Verified)
  • Total Linked Identifier Instances Across 98 - References/: 328 PMID links, 321 DOI links, and 74 PMCID open-access full-text links across 78 reference dossiers.
  • Parent Hub: 98 - References MOC · 00 - Biohacking Index

1. Direct Source Access Format (Preserved Raw IDs + One-Click Links)

Every reference in 98 - References/ keeps its raw copy-pasteable identifier inside backticks while providing immediate one-click verification links right beside it:

  • DOI Format: DOI: \10.xxxx/yyyy` (DOI · SciDB · Scholar)`
    • [DOI]: Official publisher landing page via doi.org.
    • [SciDB]: Direct full-text PDF lookup on Anna’s Archive SciDB (annas-archive.li/scidb/<doi>).
    • [Scholar]: Citation graph, citing papers, and related versions on Google Scholar.
  • PMID Format: PMID: \12345678` (PubMed · Europe PMC)`
    • [PubMed]: Official NIH / National Library of Medicine abstract, MeSH terms, and ClinicalTrials.gov links.
    • [Europe PMC]: European Bioinformatics Institute (EMBL-EBI) full-text reader and citation network.
  • PMCID Format: PMCID: \PMC1234567` (PMC Full Text)`
    • [PMC Full Text]: Free open-access full-text HTML/PDF hosted on PubMed Central.

2. Three-Tier Evidence Grading System (evidence_tier)

To prevent conflating preliminary animal/in-vitro mechanisms with proven human clinical outcomes, every compound, protocol, and diagnostic note in the vault is graded in its YAML frontmatter (evidence_tier: 1 | 2 | 3) and surfaced in Compounds & Protocols.base:

TierBadge in .baseCriteriaVault Examples
Tier 1🟢 Tier 1 (Human RCT)Multiple human Randomized Controlled Trials (RCTs), Cochrane/PRISMA meta-analyses, Mendelian randomization, or large prospective cohorts () with hard clinical endpoints.ApoB/Lp(a) ASCVD risk; Finnish KIHD Sauna mortality cohort; Creatine monohydrate; Zone 2 / ; Physiological sigh RCT.
Tier 2🟡 Tier 2 (Mechanistic/Clinical)Small-to-moderate human clinical trials (), human pharmacokinetic/biomarker trials (without hard mortality endpoints), or translational human tissue studies.NMN/NR elevation; Urolithin A mitophagy; Magnesium L-Threonate (MMFS-01); Bosman et al. UVB-microbiome trial.
Tier 3🟠 Tier 3 (Preclinical/N=1)Rodent/murine lifespan studies (e.g., NIA ITP), in-vitro receptor assays, regional non-ICH clinical monographs, or N=1 biohacking protocols lacking large western Phase III human RCTs.BPC-157 & TB-500 tissue repair (predominantly rodent data); Dihexa; Khavinson bioregulator peptides; systemic Rapamycin longevity dosing.

3. Five-Point Checklist to Audit Any Claim in This Vault

  1. Check the Species & Study Design: Always read the Study Design & Sample Size bullet in the 98 - References dossier—verify whether a reported percentage (e.g., LTP or interstitial clearance) was measured in mice vs. humans, and whether the sample size was or .
  2. Click the [PubMed] or [SciDB] Link: Verify that the quantitative effect size, -value, and confidence interval in the note match the original abstract/table.
  3. Check Surrogate vs. Hard Endpoints: Distinguish between a compound that moves a surrogate biomarker (e.g., raising PBMC by 40%) versus one proven to change a functional outcome (e.g., insulin clamp disposal, 6-minute walk distance, or all-cause mortality).
  4. Verify Third-Party Analytical COAs (97 - Sourcing & Supply Chain): For compounds and supplements, never rely on label claims—require batch-specific Eurofins, GBA, IFOS, or Janoshik Analytical (RP-HPLC + Mass Spec + LAL endotoxin) testing as documented in 97 - Sourcing & Supply Chain MOC.
  5. Close the Loop with N=1 Biomarkers (03 - Diagnostics & Quantified Self): Validate any intervention against your own baseline and follow-up telemetry (e.g., SYNLAB/Puhti blood panels, Kubios RMSSD HRV, or FreeStyle Libre 3 CGM).

4. Automated PowerShell Audit Script (Verify All Vault PMIDs Against Live NCBI API)

You can run this PowerShell command from d:\AI\biohacking at any time to verify every PMID in the vault against the live National Library of Medicine (eutils.ncbi.nlm.nih.gov) database:

$files = Get-ChildItem -Path "d:\AI\biohacking\Biohacking\98 - References" -Recurse -Filter "*.md"
$pmidMap = @{}
foreach ($f in $files) {
    $c = Get-Content -Path $f.FullName -Raw -Encoding UTF8
    foreach ($m in [regex]::Matches($c, 'PMID:?\s*`(\d{5,9})`')) {
        $pmidMap[$m.Groups[1].Value] = $f.Name
    }
}
$pmidList = @($pmidMap.Keys)
$verified = 0
for ($i = 0; $i -lt $pmidList.Count; $i += 100) {
    $batch = $pmidList[$i..([Math]::Min($i + 99, $pmidList.Count - 1))] -join ","
    $resp = Invoke-RestMethod -Uri "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi?db=pubmed&id=$batch&retmode=json"
    foreach ($id in $pmidList[$i..([Math]::Min($i + 99, $pmidList.Count - 1))]) {
        if ($resp.result.$id.title) { $verified++ }
        else { Write-Warning "Unverified PMID $id in $($pmidMap[$id])" }
    }
}
Write-Host "Verified $verified / $($pmidList.Count) unique PMIDs against live NCBI PubMed API."

Graph View

Backlinks