-
Notifications
You must be signed in to change notification settings - Fork 0
Pipeline Overview
西岡佳祐 edited this page Apr 10, 2026
·
1 revision
Patient biopsy
│
▼
NGS (Shotgun sequencing) ← PRJEB71108 cohort (peri-implantitis)
│ bowtie2 alignment
▼
MetaPhlAn 4 ← relative abundance per genus
│ metaphlan_feature_table_to_init_comp.py
▼
init_comp.json ← normalized fractions: Str/Act/Vel/Hae/Rot/Fus/Por
│
├─────────────────────────────────────────┐
│ │
▼ ▼
AGORA GEM library Hamilton ODE (0D)
(~800 reconstructions) TMCMC-calibrated parameters
│ select 5 species │
▼ ▼
GEM per species Community composition φ(t)
(~1000 rxns / ~800 metabolites) │
│ ▼
▼ Dysbiosis Index (DI)
dFBA (LP per species per Δt)
max μ s.t. stoichiometry + medium
│ medium update t→t+Δt
▼
COMETS (Java + cometspy) ← 2D spatial: 60×40 voxel grid
│ z=0: implant surface
│ z=top: GCF/saliva reservoir
▼
Metabolic profile (80 h)
· species biomass timecourse
· lactate, succinate, O₂ gradients
· pH proxy
| Component | Tool | Purpose |
|---|---|---|
| Sequencing | Illumina shotgun | Patient metagenome |
| Profiling | MetaPhlAn 4.2.4 | Relative abundance → genus fractions |
| GEM library | AGORA v1.03 | Genome-scale metabolic reconstructions |
| dFBA solver | COMETS (Java) | LP per species per Δt |
| 2D spatial | cometspy | Nutrient diffusion + growth on grid |
| Sensitivity | SALib (Sobol) | N=256, 12 kinetic params |
# Step A — 0D parameter sweep
python comets/run_comets_pipeline.py --step A
# Step B — 2D spatial (healthy vs diseased)
# submitted via PBS:
qsub comets/run_comets_BC.sh
# Step C — patient-specific (requires MetaPhlAn output)
python comets/run_comets_pipeline.py --step C \
--init-comp data/metaphlan_profiles/init_comp_ERR13166576_A_3.json| File | Content |
|---|---|
pipeline_results/A_0d_comparison.png |
0D healthy vs diseased timecourse |
pipeline_results/B_spatial_comparison.png |
2D spatial biomass + nutrients |
pipeline_results/sobol_sensitivity.png |
Sobol sensitivity indices |
pipeline_results/pipeline_overview.png |
Pipeline diagram |