200-year-old death math could transform how protein drugs are built

Competing risks survival analysis (Fine & Gray 1999, actuarial roots >200y)
De novo protein design for therapeutics (RFdiffusion 2023, ProteinMPNN 2022, <4y)

Why This Matters

Actuaries have long used a statistical trick — separating the different ways something can fail — to predict human mortality, but researchers are now asking whether this same framework could decode how next-generation protein drugs fall apart inside the body. The surprising twist is that fixing one vulnerability in these drugs could mathematically guarantee another weakness gets worse, meaning the entire industry's approach to testing stability may be built on a blind spot. If these ideas hold up, borrowing an insurance mathematician's toolbox could reshape how AI-designed medicines are approved — and which ones actually make it to patients.

4 HYPOTHESESavg score 7.82 PASS2 CONDITIONAL
📐 Mathematics & Statistics🧬 Molecular & Cell Biology

Compare Hypotheses

All Hypotheses

Click any hypothesis to see the full mechanism, evidence, and test protocol.

📐 Mathematics & Statistics🧬 Molecular & Cell Biology

The Dominant Competing Risk Theorem -- Optimizing One Failure Mode Provably Accelerates Another

PASS
Competing risks survival analysis (Fine & Gray 1999, actuarial roots >200y)
De novo protein design for therapeutics (RFdiffusion 2023, ProteinMPNN 2022, <4y)
ScoutConverging Vocabularies

Fix one way a protein drug breaks, and you mathematically guarantee another weakness gets worse.

Score8.3
Confidence8
Grounded9
📐 Mathematics & Statistics🧬 Molecular & Cell Biology

Competing-Risk Cumulative Incidence Functions as a Unified Protein Therapeutic Lifetime Predictor

PASS
Competing risks survival analysis (Fine & Gray 1999, actuarial roots >200y)
De novo protein design for therapeutics (RFdiffusion 2023, ProteinMPNN 2022, <4y)
ScoutConverging Vocabularies

A survival statistics framework borrowed from actuaries could predict exactly how—and when—engineered protein drugs will break down in the body.

Score8
Confidence7
Grounded8
📐 Mathematics & Statistics🧬 Molecular & Cell Biology

Competing Risks Censoring Correction for Immunogenicity -- Anti-Drug Antibodies as Interval-Censored Competing Risk

CONDITIONAL
Competing risks survival analysis (Fine & Gray 1999, actuarial roots >200y)
De novo protein design for therapeutics (RFdiffusion 2023, ProteinMPNN 2022, <4y)
ScoutConverging Vocabularies

Fixing a hidden flaw in drug safety testing: fast-failing proteins mask their immune risks until it's too late.

Score7.7
Confidence7
Grounded7
📐 Mathematics & Statistics🧬 Molecular & Cell Biology

Nelson-Aalen Cumulative Hazard Decomposition Reveals Hidden Failure Modes in Accelerated Stability Studies

CONDITIONAL
Competing risks survival analysis (Fine & Gray 1999, actuarial roots >200y)
De novo protein design for therapeutics (RFdiffusion 2023, ProteinMPNN 2022, <4y)
ScoutConverging Vocabularies

Splitting protein drug degradation into its hidden failure modes could make shelf-life predictions far more accurate.

Score7.5
Confidence6
Grounded7