Skip to content

Practitioner Toolkit

Methods for separating personal observations, low-cost measurement capabilities, and research-quality evidence. This toolkit does not turn an observation or accessible assay into a treatment protocol.

Three-tier structure

  1. N-of-1 observations — structured records of direct observations; useful for generating study variables, not population or treatment claims.
  2. Measurement capability builds — low-cost research measurements. A capability build does not authorize ingestion, organism construction, or clinical use.
  3. Rigor Disciplines — literature mining, evidence verification, falsification, and AI-discovery controls that apply across research modes.

Self-Experiments (n=1)

Structured n=1 records can improve timing, endpoint, and confounder choices for later studies. They do not establish safety, causality, dose, or efficacy.

  • self-experiment-protocol.md — biomarker monitoring framework: blood panels (CBC / CMP / UA / hs-CRP / LDH / HbA1c), 16S stool, daily diary, red-flag halt criteria; specialty biomarkers (C3 / C4 / CH50 / C5a CP0; urinary LTE4 CP6a)
  • personal-genome-protocol.md — kitchen-table sequencing as both personal genome project AND Open Enzyme strain-QC infrastructure; gout pharmacogenomic query list (HLA-B*58:01, ABCG2 Q141K, SLC2A9, URAT1, PDZK1, MEFV)
  • PERT-timing self-experiment (April 2026 → ongoing) — a direct, single-subject record involving a commercial digestive-enzyme product, documented in digestive-enzyme-optimization.md. It describes only the product, subject, meals, timing, and outcomes actually observed. It does not establish timing, formulation, manufacturing process, dose, efficacy, or safety for an engineered UOX or koji platform.

DIY Capability Builds

Minimal-equipment research procedures. Some can support bench measurements for engineered-strain research; none is a personal-consumption protocol for an engineered strain.

  • koji-home-fermentation.md — wild-type food-fermentation reference. It may supply a process baseline for non-engineered koji, but it is not an ingestion or production protocol for engineered UOX material.
  • enzyme-quantification-protocol.md — tiered methods for measuring amylase / protease / lipase activity (kitchen → smartphone colorimetry → community-college bench → outsourced contract assay).
  • engineered-koji-protocol.md — engineering-stage construct, strain-QC, expression, and assay plan for A. oryzae. Its outputs require identity, containment, activity, coproduct, and safety evaluation; they are not for routine or personal consumption.
  • medicinal-mushroom-extract-sops.md — planned extract characterization SOPs (GLPP, cordycepin, ergothioneine); operator-independent reproducibility tolerances.
  • Low-cost liquid-handling automation (Picolab prior art, 2026-05-19). Picolab v2 is an MIT-licensed prototype that repurposes an Ender-style printer gantry into a syringe liquid handler with a FastAPI backend, React dashboard, G-code motion planner, calibration store, camera-assisted OpenAI agent workspace, and operator approval gate before hardware execution. For Open Enzyme, this belongs in the capability-build lane: plausible infrastructure for repeatable tube-scale colorimetric assays, serial dilutions, and quantification-ladder.md Tier ⅔ method development. It is not a sterile, clinical, or production liquid handler. Dedicated page threshold: create low-cost-lab-automation.md only after there is OE-specific analysis of bill of materials, positional/volume accuracy, contamination controls, assay compatibility, and first automatable protocols. (Engineering prior art; source: Picolab v2 repository)
  • Sequencing capability (sub-bucket of personal-genome-protocol.md) — candidate infrastructure for research-genome analysis and construct-identity verification. Clinical interpretation requires a clinically validated assay; engineered-strain release requires the applicable controlled QC system.

Rigor Disciplines (cross-cutting)

Methodology that applies to wet-lab, computational, literature, and n-of-1 observation work.

  • manual-literature-mining.md — five-rule protocol for safe LLM literature use (safe primitives only, anchor to meta.json, grep-verify all numbers, never propagate map / reduce summaries, cite line-anchored). Surfaced 2026-05-05 after documented Paperclip map operator hallucination.
  • chembl-cross-check.md — quarterly ChEMBL v34 cross-reference of stack compounds; separates direct-inhibitor claims from pathway-modulator claims (e.g., quercetin → 5-LOX, not NLRP3).
  • linter-design.md — two-linter architecture (Document Lint always-on; Falsification Lint on-demand per-hypothesis with killshot menus, failure-mode ontology, survival scoring).
  • tcm-modern-rigor-intersection.md — six-rule methodology for evidence-leveling traditional-medicine claims (chokepoint mapping, ChEMBL cross-check, bioavailability-honest framing, formula decomposition, standardized-extract specification, falsification-card discipline).
  • autonomous-screening-methodology.md — peer-track methodology page on ClockBase Agent; transferable patterns for comp-NNN: search-space sizing, provenance-bound property tables, hypothesis-then-verify, autonomy boundaries, and multi-model promotion only when predictors are calibrated to the same named outcome.
  • cross-validation.md — thesis stress-test discipline; risk matrix across all tracks; identifies true blockers vs. surmountable obstacles.

Section conventions

  • Evidence-level discipline applies. Use the standard wiki tags (Clinical Trial / Animal Model / In Vitro / Mechanistic Extrapolation). An n=1 observation is useful for generating study variables, not for population, safety, dose, or treatment claims.
  • Privacy-by-default. Self-experiment data is private unless explicitly published. Genetic data in particular is governed by the privacy gradient in personal-genome-protocol.md.
  • Project crossover where applicable. Sequencing and enzyme-assay capabilities may support platform measurements. Direct observations from commercial PERT use do not transfer a dose, timing rule, formulation, process, efficacy signal, or safety conclusion to engineered UOX.
  • Pre-commit grep-verify gate (CLAUDE.md §4) applies to load-bearing numbers in all three sub-buckets equally.

Open questions

  • Which low-cost measurements remain reproducible after blinded comparison with a reference laboratory?
  • Which observations are useful enough to justify a controlled study, and which are too confounded to carry forward?
  • Which capability gaps require a partner laboratory rather than an independent setup? Mammalian cell culture, pathogen work, engineered-organism production, and clinical testing require appropriate facilities, oversight, and containment.