# Core Workflows

## A) Single-molecule workflow

Use this when you want a focused readout for one gene, protein, or miRNA.

1. Search one molecule (`ATM`, `P05067`, `hsa-miR-107`).
2. Confirm node auto-selection in Summary.
3. Review EV Evidence first, then Pathway and Disease.
4. Use assistant prompts for narrow follow-up (`explain this pathway result`).

## B) Disease-first workflow

Use this for reverse lookup from a phenotype or diagnosis phrase.

1. Submit a disease-first natural-language query.
2. Review system summary before selecting nodes.
3. Open Disease to inspect ranked associations.
4. Cross-check with EV Evidence and Pathway.
5. Move to Cell, LR, and PPI for candidate communication context.

## C) Pathway-first workflow

Use this when your starting question is functional biology.

1. Query pathway intent (`PI3K-Akt`, `apoptosis`, `Notch`).
2. Review top entities in Summary.
3. Open Pathway and evaluate p-value and q-value in the same filter state.
4. Compare against Disease and EV Evidence before prioritizing follow-up.

## D) EV-evidence-first workflow

Use this when provenance and detectability are primary.

1. Start with an EV evidence question.
2. Inspect EV Evidence rows and source fields.
3. Check EV-TRACK fields when present.
4. Use Pathway and Disease only after evidence review.

## E) Collection workflow

Use this for panel-level hypothesis generation.

1. Submit a curated molecule set.
2. Start in system summary.
3. Recommended order: Summary -> Pathway -> Disease -> Cell -> LR/PPI.
4. Export machine-readable tables for downstream review.

## Workflow boundary

EVd3x ranks and organizes candidate evidence for review. It does not establish mechanism or causality without separate experimental validation.