# DKDComboMiner — top 8 unexplored cells

> Research-use only. Synthetic mock data. Not for clinical decision-making. IRB review required.

- Total cells scored: 235872
- Combo size up to: 2
- Weights: {'gap': 0.2, 'trend': 0.1, 'trial_gap': 0.1, 'kdigo': 0.15, 'mech': 0.2, 'feasibility': 0.2, 'faers_penalty': 0.15}

## Ranked hypothesis cards

## #1 (score 0.816)

### Hypothesis: Empagliflozin (EMPA), Liraglutide (LIRA) | KDIGO G2A1 | UACRV | T2DM_ASCVD

- Drug combo: Empagliflozin (EMPA), Liraglutide (LIRA)
- Class signature: GLP1RA + SGLT2i
- KDIGO stage: G2A1
- Primary biomarker: UACRV
- Target population: T2DM_ASCVD
- Composite score: **0.816**
- Synthetic literature gap (pubmed_n / trend_5y%): 1 / +55%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 0.925 | Mechanism: 0.88 | FAERS caution: 0.45 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=1 (gap=0.99), 5y trend +55% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 0.93 / Mechanism plausibility prior = 0.88; FAERS caution = 0.45 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #2 (score 0.803)

### Hypothesis: Dapagliflozin (DAPA), Dulaglutide (DULA) | KDIGO G3aA3 | UACR | T2DM_ASCVD

- Drug combo: Dapagliflozin (DAPA), Dulaglutide (DULA)
- Class signature: GLP1RA + SGLT2i
- KDIGO stage: G3aA3
- Primary biomarker: UACR
- Target population: T2DM_ASCVD
- Composite score: **0.803**
- Synthetic literature gap (pubmed_n / trend_5y%): 9 / +51%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 0.925 | Mechanism: 0.88 | FAERS caution: 0.45 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=9 (gap=0.95), 5y trend +51% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 0.93 / Mechanism plausibility prior = 0.88; FAERS caution = 0.45 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #3 (score 0.8029)

### Hypothesis: Finerenone (FINE), Semaglutide (SEMA) | KDIGO G1A1 | UACRV | T2DM_ASCVD

- Drug combo: Finerenone (FINE), Semaglutide (SEMA)
- Class signature: GLP1RA + MRA
- KDIGO stage: G1A1
- Primary biomarker: UACRV
- Target population: T2DM_ASCVD
- Composite score: **0.8029**
- Synthetic literature gap (pubmed_n / trend_5y%): 0 / +49%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 0.925 | Mechanism: 0.8 | FAERS caution: 0.388 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=0 (gap=1.00), 5y trend +49% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 0.93 / Mechanism plausibility prior = 0.80; FAERS caution = 0.39 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #4 (score 0.8011)

### Hypothesis: Dulaglutide (DULA), Finerenone (FINE) | KDIGO G3aA2 | UACR | T2DM_ASCVD

- Drug combo: Dulaglutide (DULA), Finerenone (FINE)
- Class signature: GLP1RA + MRA
- KDIGO stage: G3aA2
- Primary biomarker: UACR
- Target population: T2DM_ASCVD
- Composite score: **0.8011**
- Synthetic literature gap (pubmed_n / trend_5y%): 8 / +54%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 0.925 | Mechanism: 0.8 | FAERS caution: 0.388 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=8 (gap=0.96), 5y trend +54% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 0.93 / Mechanism plausibility prior = 0.80; FAERS caution = 0.39 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #5 (score 0.801)

### Hypothesis: Dulaglutide (DULA), Empagliflozin (EMPA) | KDIGO G3aA1 | EGFRSLOPE | T2DM_ASCVD

- Drug combo: Dulaglutide (DULA), Empagliflozin (EMPA)
- Class signature: GLP1RA + SGLT2i
- KDIGO stage: G3aA1
- Primary biomarker: EGFRSLOPE
- Target population: T2DM_ASCVD
- Composite score: **0.801**
- Synthetic literature gap (pubmed_n / trend_5y%): 6 / +57%
- Synthetic KR active trials: 1
- KDIGO 2024 strength avg: 0.925 | Mechanism: 0.88 | FAERS caution: 0.45 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=6 (gap=0.97), 5y trend +57% / Active KR trials (synthetic) = 1; KDIGO 2024 strength avg = 0.93 / Mechanism plausibility prior = 0.88; FAERS caution = 0.45 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #6 (score 0.7995)

### Hypothesis: Canagliflozin (CANA), Perindopril (PERI) | KDIGO G2A1 | UACRV | T2DM_ASCVD

- Drug combo: Canagliflozin (CANA), Perindopril (PERI)
- Class signature: RAS + SGLT2i
- KDIGO stage: G2A1
- Primary biomarker: UACRV
- Target population: T2DM_ASCVD
- Composite score: **0.7995**
- Synthetic literature gap (pubmed_n / trend_5y%): 8 / +46%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 1.0 | Mechanism: 0.72 | FAERS caution: 0.3 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=8 (gap=0.96), 5y trend +46% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 1.00 / Mechanism plausibility prior = 0.72; FAERS caution = 0.30 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #7 (score 0.7978)

### Hypothesis: Dapagliflozin (DAPA), Telmisartan (TELM) | KDIGO G3aA3 | EGFRSLOPE | T2DM_ASCVD

- Drug combo: Dapagliflozin (DAPA), Telmisartan (TELM)
- Class signature: RAS + SGLT2i
- KDIGO stage: G3aA3
- Primary biomarker: EGFRSLOPE
- Target population: T2DM_ASCVD
- Composite score: **0.7978**
- Synthetic literature gap (pubmed_n / trend_5y%): 1 / +49%
- Synthetic KR active trials: 1
- KDIGO 2024 strength avg: 1.0 | Mechanism: 0.72 | FAERS caution: 0.3 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=1 (gap=0.99), 5y trend +49% / Active KR trials (synthetic) = 1; KDIGO 2024 strength avg = 1.00 / Mechanism plausibility prior = 0.72; FAERS caution = 0.30 / Best matching KR cohort: KDA-REG (feasibility 0.88)


## #8 (score 0.796)

### Hypothesis: Dulaglutide (DULA) | KDIGO G2A2 | UACR | T2DM_LEAN

- Drug combo: Dulaglutide (DULA)
- Class signature: GLP1RA
- KDIGO stage: G2A2
- Primary biomarker: UACR
- Target population: T2DM_LEAN
- Composite score: **0.796**
- Synthetic literature gap (pubmed_n / trend_5y%): 30 / +58%
- Synthetic KR active trials: 0
- KDIGO 2024 strength avg: 0.85 | Mechanism: 0.7 | FAERS caution: 0.1 | KR IIS feasibility: 0.88
- Rationale: Synthetic PubMed n=30 (gap=0.85), 5y trend +58% / Active KR trials (synthetic) = 0; KDIGO 2024 strength avg = 0.85 / Mechanism plausibility prior = 0.70; FAERS caution = 0.10 / Best matching KR cohort: KDA-REG (feasibility 0.88)