# Core Workflows

## A) Single molecule workflow

Best for mechanistic deep dive of one miRNA, mRNA, or Protein.

1. Search one molecule (`hsa-miR-3910`, `ATM`, `P08962`).
2. Confirm node is auto-selected and **Summary** is populated.
3. Open EV Evidence, Pathway, then Disease tabs.
4. Use chat for focused interpretation:
   - `explain this pathway result`
   - `show dominant diseases linked to this node`

## B) Disease-first workflow

Best for reverse discovery from a disease phenotype.

1. Query disease-first in natural language.
2. Review collection/system summary first.
3. Open Disease tab to inspect ranked links.
4. Use Pathway and EV Evidence tabs to cross-validate biological plausibility.

## C) Pathway-first workflow

Best for functional biology exploration.

1. Query pathway intent (`PI3K-Akt`, `apoptosis`, etc.).
2. Review top entities in Summary.
3. Open Pathway tab and evaluate significance + source coverage.
4. Cross-check Disease and EV Evidence for translational relevance.

## D) EV-evidence-first workflow

Best for cargo evidence and isolation provenance.

1. Start with EV evidence question in natural language.
2. Inspect EV Evidence tab for source rows and EV-TRACK metadata.
3. Continue to Pathway and Disease only after confirming evidence quality.

## E) Collection workflow

Best for panel-level hypotheses.

1. Submit a curated molecule set.
2. Use system summary for broad context.
3. Open tabs in this order for speed: Summary -> Pathway -> Disease -> Cell -> L-R/PPI.
4. Use chat to refine the active dataset (`top 50`, `only miRNA`, `export selected columns`).