How selbioinfonivedi-web/g4_ews_pipeline works
G4-WATCH is a research framework for evaluating whether G-quadruplex loci behave differently from matched controls over viral evolutionary history, and—only if the evidence gate is open—producing surveillance scores. Its main workflow constructs an atlas, processes sequence data through QC, alignment and phylogenetic analysis, tests locus-control differences, then conditionally scores and reports. It also provides a separate operator console and a read-only dashboard/workstation. Pipeline stages and web boundaries follow the README; unsampled stage internals are represented only where their role is documented.
Main entry points: cli.py Stage CLI · stage45_dh1.py D.H1 test · app.py Operator console · gating.py Scoring gate
Components 25
Atlas and inputs
- G4 atlas
g4watch/pipeline/stage0_atlas.py - G4 prediction
g4watch/atlas/stage0.py - Sequence corpus
g4watch/io/corpus.py - QC and alignment
g4watch/pipeline/stage1_align.py
Evolution analysis
- Recombination screen
g4watch/pipeline/stage15_recombination.py - Phylogenetic states
g4watch/phylo/ancestral_states.py - Variant analysis
g4watch/variants/alignment_variant_caller.py
Evidence and gating
- D.H1 test
g4watch/pipeline/stage45_dh1.py - Surveillance metrics
g4watch/metrics/surveillance_metrics.py - Scoring gate
g4watch/gating.py - Verdict ledger
g4watch/ledger.py
Scoring and views
- Gated scoring
g4watch/pipeline/stage5_scoring.py - Score models
g4watch/scoring/integrated_score.py - Scored report
g4watch/pipeline/stage6_reporting.py - Dashboard builder
g4watch/reporting/dashboard.py
Web applications
- Operator console
web/runner/app.py - Command registry
web/runner/commands.py - Job manager
web/runner/jobs.py - Analysis database
web/store.py - Read-only dashboard API
web/backend/app.py - Linked-view workstation
web/workstation/dataset.py
Other components
- Research operator
- Stage CLI
g4watch/cli.py - Pipeline user
- Web user
Connections 29
- Research operator → Stage CLI: runs stages
- Stage CLI → G4 atlas: dispatches
- G4 atlas → G4 prediction: scans genome
- Stage CLI → QC and alignment: dispatches
- QC and alignment → Sequence corpus: processes sequences
- Stage CLI → Recombination screen: dispatches
- Recombination screen → Phylogenetic states: precedes
- Stage CLI → Phylogenetic states: dispatches
- Stage CLI → Variant analysis: dispatches
- Stage CLI → D.H1 test: dispatches
- D.H1 test → Surveillance metrics: computes
- D.H1 test → Scoring gate: evaluates
- D.H1 test → Verdict ledger: appends verdicts
- Stage CLI → Gated scoring: dispatches
- Gated scoring → Scoring gate: checks permission
- Scoring gate → Verdict ledger: reads verdicts
- Gated scoring → Score models: computes scores
- Stage CLI → Scored report: dispatches
- Scored report → Scoring gate: checks permission
- Pipeline user → Operator console: uses console
- Operator console → Command registry: builds command
- Operator console → Job manager: manages jobs
- Operator console → Analysis database: persists analyses
- Job manager → Stage CLI: launches commands
- Web user → Read-only dashboard API: views results
- Read-only dashboard API → Dashboard builder: builds dashboard
- Dashboard builder → Verdict ledger: reads verdicts
- Web user → Linked-view workstation: opens linked views
- Linked-view workstation → Verdict ledger: reads verdicts