Human Model · HM(t) at day 90

State-timeSynthetic demo data⟨ I, S_t, H_0:t, E_t, Θ_t, T, U_t ⟩ — belief, not fact; uncertainty shown, not hiddenSteering view

Localization — where is this individual biologically now?

Belief state b_t per dimension: value ± sd from within-person noise plus the observation model's semantic distance. Trajectory features (velocity, autocorrelation, recovery time) are state variables.

Dimension	b_t	dS/dt	autocorr	recovery	observation model
Therapy response	41 ± 10.1	-0.64/d	0.7	4 d	Imaging + tumour markers, interpolated · proxy · last read ≤ 14 d
Gastrointestinal function	61 ± 7	+0.12/d	0.63	1 d	Daily GI check-in · direct · today
Treatment toxicity	41 ± 7	-0.01/d	-0.17	2 d	Labs, interpolated between draws · proxy · last draw ≤ 7 d
Inflammatory activity	56 ± 12.6	+0.15/d	0.25	10 d	CRP/ferritin + immune signature · inferred · weeks
Immune / metabolic	48 ± 11.4	-0.42/d	0.06	1 d	Metabolic panel + wearable · inferred · weeks
Nutritional sufficiency	36 ± 8.6	+0.22/d	0.27	1 d	Weight, albumin, food log · proxy · days
Adherence	57 ± 6.9	-0.73/d	0.85	—	Dose + to-do completion · direct · today
Function / quality of life	36 ± 6.7	-0.5/d	0.41	1 d	PRO items · direct · today

High autocorrelation and slow recovery are early-warning signatures of a critical transition (Scheffer et al.), independent of the level.

Provenance — how did they arrive here?

The path H_0:t: baseline constraints, divergence, perturbations, actions and responses in order.

day 0baselineBaseline S14: EC-03 phenotype; Poor metabolizer; butyrate 28
day 14perturbationDaily walks resumed (3.3)
day 18perturbationDietary drift (-2.6)
day 31recommendedNP-02 recommended
day 32acceptedPatient accepted
day 42responseGI symptom burden −15% vs expected
day 48divergenceTrajectory diverged from the population model
day 48noteTrajectory diverged from prediction
day 48state changeTolerance state changed
day 53recommendedGI-04 recommended
day 54acceptedPatient accepted
day 62responseGI symptom burden −22% vs expected
day 68state changeTolerance state changed

Now: in the failing basin, heading towards failing.

Mechanism — which theories best explain the trajectory?

Versioned competing theories with supporting (E+), contradicting (E−) and missing-but-discriminative (E?) evidence. The contradiction engine lists what should NOT happen under each theory and whether it did.

T1 v3leading

confidence 47%

Mucosal barrier loss: drug-induced epithelial injury amplified by a FUT2-dependent glycan deficit; histamine and inflammatory load add on top.

E+

FUT2 non-secretor genotype (barrier glycan deficit)
DeepGene: host DNA fragmentation atypical
DeepGene: tissue recycling 60
Histamine load → GI +5.2 next day
Inflammatory load 56
GI-04 (barrier-support protocol) improved GI

E−

E? · would discriminate

Faecal calprotectin (would separate barrier from motility)
Serum DAO / histamine
Zonulin

Forbidden under this theory

✓ holdsGI should not improve while histamine intake stays high (GI velocity -0.09/day)

✓ holdsLow-residue meals should reduce GI within a week (NP-02 → improved)

v1 baseline: genotype + phenotype → v2 day 54: GI-04 response added to E+ → v3 dietary attribution from 60-day log

T2 v2competing

confidence 38%

Bile-acid diarrhoea on a low-butyrate ecology: reduced SCFA production leaves bile acids unconjugated after fat-rich meals; GI burden tracks fat intake with a one-day lag.

E+

Butyrate capacity 28/100
DeepGene: 6 keystone butyrate producers absent
DeepGene: H2S above band
Fat load → GI +6.5 next day
Diversity 1.9

E−

E? · would discriminate

Faecal bile acids
7α-hydroxy-4-cholesten-3-one (C4)
Rechallenge: 3 high-fat days under observation

Forbidden under this theory

✓ holdsGI should not rise on low-fat days (fat effect 6.5 over 32 low-fat days)

v1 baseline: microbiome → v2 dietary attribution

T3 v1falsified

confidence 9%

Drug-exposure toxicity: pharmacokinetic over-exposure drives GI and systemic toxicity together; dietary and microbial factors are secondary.

E+

Poor metabolizer: exposure above modelled range
GI autocorrelation 0.62 (slow recovery)

E−

Toxicity within range

E? · would discriminate

Trough drug level
Dose-timing rechallenge (evening vs morning)

Forbidden under this theory

✗ violatedGI should not move with diet if exposure is the driver (3 dietary detractors with ≥ 80% confidence)

v1 baseline: pharmacogenomics

T4 v1competing

confidence 5%

Behavioural spiral: symptomatic days cut adherence and sleep, which raise GI and stress, which cut adherence further; the biology is a follower, not the driver.

E+

Adherence 57%

E−

Sleep and stress stable

E? · would discriminate

Dose-taking timestamps vs symptom timestamps
Sleep-window rechallenge

Forbidden under this theory

✓ holdsGI should not worsen on days with full adherence and good sleep (insufficient paired days)

v1 from tracker trends

Forecast — which futures are likely under alternative actions?

30-day rollout of GI burden under the baseline model T and under each safe candidate action (effect ramps over onset). Dashed = prediction, never fact.

No action

GI 61 → 52

Under ME-01

GI 61 → 50

Under SC-12

GI 61 → 45

Under CR-01

GI 61 → 47

Intervention as measurement

Each past action as a measurement of this patient's transition dynamics: predicted ΔGI vs observed, prediction error ε_t, and the within-patient evidence patterns → individualized response confidence (observational, kept separate from randomized evidence).

Action	pred ΔGI	obs ΔGI	ε	temporality	magnitude	dechallenge	rechallenge	cross-modal	confidence
NP-02 day 32	-2	34	+36	✗	✓	n/a	n/a	✗	18%
GI-04 day 54	-16	1	+17	✗	✗	n/a	n/a	✗	0%

Learning curve · |ε| per cycle:NP-02 36 (prior 36)GI-04 13 (prior 17)improving over cycles

Decision — the Restore controller

Safe set computed first, then A* = argmax E[HealthGain + λ₁·InformationGain + λ₂·ConfidenceGain − λ₃·Risk − λ₄·Burden − λ₅·Cost]. λ₁ = 0.6 (theories near-equiprobable), λ₃ = 1 (stable enough). Execution stays with the clinician or the informed patient (EU MDR Class I).

Class	Action	Health	Info	Conf.	Risk	Burden	Cost	Utility
Escalate	CR-01 Clinician Review 100% of 6 similar improved.	0.37	0.2	0.25	0.02	0.25	0.15	0.39
Measure	Faecal calprotectin (would separate barrier from motility) Discriminates T1 from its competitors (E?).	0	0.75	0.2	0.02	0.25	0.2	0.32
Test	Serum DAO / histamine Discriminates T1 from its competitors (E?).	0	0.75	0.2	0.02	0.25	0.5	0.27
Escalate	ME-01 Monitoring Escalation 100% of 66 similar improved.	0.22	0.2	0.25	0.08	0.1	0.15	0.26
Treat	SC-12 Supportive Compound 67% of 45 similar improved.	0.35	0.2	0.25	0.08	0.25	0.4	0.26
Treat	HY-01 Hydration Protocol 0% of 0 similar improved.	0.15	0.2	0.25	0.08	0.25	0.15	0.11
Treat	BR-03 Behavioural Recommendation 0% of 0 similar improved.	0.14	0.2	0.25	0.08	0.25	0.15	0.1
Observe	Continue passive collection, no perturbation A clean transition can be observed only without a new perturbation.	0	0.15	0.05	0	0.05	0	0.08
Treat	AD-05 Adherence Support 0% of 0 similar improved.	0.1	0.2	0.25	0.08	0.25	0.15	0.06
Maintain	Hold current regimen and content Protects nothing while the state deteriorates.	0.05	0.05	0.05	0.4	0.05	0	-0.33

Excluded from the safe set

GI-04 Not indicated for the current state.
NP-02 Not indicated for the current state.
DT-01 Dose-timing change requires investigator authorization in poor metabolizers (contraindication from I).

Evidence weights · w = Authority × Relevance × Persistence

Biological identity Poor metabolizer; FUT2, IL6, CYP2D6
Germline panel, baseline
0.86 (0.95·0.9·1)
High-authority clinical Toxicity 41, response 41
Labs / imaging, ≤ 14 d
0.77 (0.9·1·0.85)
Intervention response 2 within-patient responses recorded
ΔS after each action, accumulating
0.68 (0.75·1·0.9)
Deep biological state DeepGene retest day 90: damage 85/100, Shannon 1.64, butyrate 282 CPM, 3 opportunists above, 6 keystone absent
Synthetic DeepGene metagenomics + host DNA, sample 909862-R (compositional proxy)
0.63 (0.7·1·0.9)
Dynamic state GI 61, sleep 6.9 h, HRV 46
Companion + wearable, today
0.18 (0.6·1·0.3)
Exposure context Prepare a plain, low-fat pasta or rice meal and note how it feels afterwards; dosing per protocol; adherence 57%
Food log + dose log, event-level
0.1 (0.55·0.9·0.2)

LLIFE OS insight · the five questions, in prose

grounded · add an LLM key for prose

P-01332 is on day 90 of NBL-101 (EC-03, baseline S14): response unstable, tolerance stable, GI unstable. Over 14 days response moved -14 and GI -2; discontinuation risk is 67%. The state is in the failing basin, heading towards failing, after diverging from the model on day 48.

Biology sets the gain: FUT2 non-secretor, IL6 high-expression variant, Poor metabolizer, Low butyrate capacity, Low microbial diversity, Baseline state S14, Cluster EC-03 (high response / high intolerance), Adherence < 60%; butyrate capacity 28/100, diversity 1.9. Pushing it down: Enterococcus faecium 16.88% (LPS load) (-28.5), Low butyrate capacity (-23.3), Dietary fat load (-20.5). Pulling it up: GI Support Protocol GI-04 (+28.4), Dietary Support Protocol NP-02 (+24.5). From the daily log: fat +6.5 GI next day, simple carbs +6 GI next day, histamine +5.2 GI next day.

Past actions: NP-02 on day 32 → improved (15%); GI-04 on day 54 → improved (22%).

Next: ME-01 — 100% of 284 similar trajectories improved, 83% probability here. This would be contradicted by a further GI rise within the onset window or falling adherence.