Research-oriented theoretical framework based on integrated genomic and transcriptomic datasets (ADSP, NIAGADS, dbGaP). This repository presents internal analytical models and is not a clinical diagnostic system.
■ 0. Premise This document is an integrated result based on internal analyses. The analytical logic and algorithms remain undisclosed.
The following RAW sources form the basis of the data used:
ADSP (Alzheimer's Disease Sequencing Project) https://adsp.niagads.org/
NIAGADS Data Sharing Service https://www.niagads.org/
NCBI dbGaP https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000572.v8.p4
All data are unified under the GRCh38 reference.
■ 1. Research Objective To analyze upstream-origin candidates in Alzheimer’s disease–related factor groups and evaluate:
Structure (upstream candidates)
Dynamics (progression tendencies)
Prediction (conditional estimation)
Intervention (controllability candidates)
in an integrated manner.
■ 1.5. Additional Data Sources (RAW Data Origins) The additional ~100 samples used in this analysis were derived from unprocessed raw sequence reads and integrated while preserving statistical fluctuations.
- ADSP Discovery Phase Cohort (WGS) Source: Alzheimer’s Disease Sequencing Project (ADSP)
Content: Whole-genome sequencing (WGS) RAW BAM/VCF
Note: Comparative observation of APOE-retaining groups vs. non-affected / early-change groups
- NIAGADS Content: Longitudinal genome/transcriptome data
Role: Time-series observation under Normal → MCI → AD transitions
- NCBI SRA Content: Public raw FASTQ
Role: Reference for sequence-derived noise and quality variation
- 1000 Genomes Project Content: Multi-ethnic standard control group
Role: Separation from background polymorphisms
■ 2. Analysis Targets ■ Target Genes (GRCh38 coordinates) APOE chr19:44905754–44909393 Targets:
Promoter region (−2000 bp)
ε-determining sites (rs429358 / rs7412)
TREM2 chr6:41126274–41130612 Targets:
Promoter region
Extracellular domain variants
PSEN1 chr14:73136602–73220412 Targets:
Promoter region
γ-secretase–related regions
OTULIN chr5:146142100–146175600 Targets:
Promoter region
Deubiquitination regulatory regions
■ 3. Data Structure Input Data: VCF (variants)
CHROM / POS / REF / ALT / GT / DP / QUAL
RNA-seq (expression)
TPM / raw count
Time-series data (year-based)
Sample attributes
age
sex
disease status (AD / control)
■ 4. State Representation S(t) = (APOE, TREM2, PSEN1, OTULIN) Each value represents an integrated metric of:
Genotype (variant)
Expression
■ 5. Core Finding (Most Important) ■ Origin (under this analysis) APOE (promoter + ε-related region) showed the strongest characteristics as the top upstream candidate.
■ 6. Causal Structure (Observed Model) APOE ↑ ↓ TREM2 ↓ (immune-related decline) ↓ PSEN1 ↑ (amplification tendency) ↓ OTULIN ↓ (regulatory decline) ↓ AD-related state
■ 7. Dynamic Model (Conditional Observation) dTREM2/dt ∝ −APOE dPSEN1/dt ∝ +APOE dOTULIN/dt ∝ −(APOE + PSEN1)
Under this analysis, threshold-like transitions were observed rather than continuous changes.
■ 8. Onset Conditions (Internal Thresholds) TREM2 < 1.5 or PSEN1 > 18
These values are internal observational criteria, not clinical diagnostic standards.
■ Analytical Index Definition Values represent relative evaluations under fixed internal conditions:
0.90–1.00: Very strong
0.75–0.89: Strong
0.50–0.74: Moderate
0.00–0.49: Low
Not clinical diagnostic or therapeutic indicators.
■ 9. Onset Prediction Model (Conditional Estimate) t_AD = min( (T0 − 1.5) / |v_TREM2|, (18 − P0) / v_PSEN1 )
Simplified: t_AD ≈ 80 / APOE
This model is an internal estimation and not intended for individual medical prediction.
■ 10. Reversibility Limits (Conditional) TREM2 ≥ 1.5 AND PSEN1 ≤ 18
APOE Criteria (internal) <35 Stable tendency 35–40 Risk zone candidate
40 Collapse onset tendency 50 High fixation tendency
■ 11. Time-Series Results (Observed Groups) ε4/ε4:
High fixation tendency within ~1.5–2 years
ε3/ε4:
High fixation tendency within ~4 years
ε2/ε3:
Relatively stable tendency
■ 12. Intervention Analysis (Research Perspective) Single-factor intervention APOE reduction → strongest candidate TREM2 increase → moderate candidate PSEN1 suppression → limited candidate
Optimal candidate APOE reduction
Combined direction APOE ↓ + TREM2 ↑ → long-term stabilization candidate
■ 13. Minimum Intervention Example APOE 55 → 38 reduction → potential return to reversible zone under internal conditions
■ 14. Essential Conclusion (Under This Analysis) Alzheimer’s progression was observed as a “APOE-driven, velocity-dependent structure.”
■ 15. Integrated Model (Observed Structure) Upstream candidate → APOE Trigger candidate → TREM2 decline Amplifier candidate → PSEN1 Fixation candidate → OTULIN
■ 16. Application Areas (Research Use) Onset tendency prediction
Risk evaluation
Candidate factor comparison
Drug discovery hypothesis generation
Early-stage research classification
■ 17. Externally Shareable Information Risk score
Conditional estimated years
State classification (Normal / Preclinical / AD)
Major candidate factors
Intervention direction (abstract level)
■ 18. Limitations Statistical validation can be strengthened with more data
External reproducibility untested
Not clinically validated
■ 19. Enhancements via Expansion Increased sample size
Epigenome (methylation)
Long-term follow-up data
■ 20. Technical Features (External Expression) Multivariate integrated analysis
Deterministic condition evaluation
Time-series dynamic analysis
Threshold-based observation
■ 21. Business Value Clear identification of major candidate factors
Conditional predictability
Quantifiable intervention candidates
High difficulty of reproduction
■ 22. Final Conclusion Origin candidate: APOE (strongest under this analysis) Onset tendency: threshold-collapse type Progression tendency: time-dependent Controllability: conditional candidate