Skills › Scientific Agent Skills › gget › common_workflows.md

common_workflows.md

32 lines · skills/gget/references/common_workflows.md at 5f2bc03

Runs code

Open on GitHub

Most marks come from matching words in the file, so treat them as pointers.

1# Common gget workflows

3Use [workflows.md](workflows.md) for current 0.30.8 integration examples and their

4verification boundaries. Use [database_info.md](database_info.md) for service,

5authentication, pagination, and return-shape details.

7| Task | Sequence of operations | Check before interpretation |

8| --- | --- | --- |

9| Gene characterization | `search` -> unique exact `gene_name` -> `info` / `seq` | Preserve Ensembl index; inspect service failures; distinguish genomic sequence from CDS. |

10| Ortholog comparison | `bgee` -> select species/ortholog -> `seq` -> FASTA file -> `muscle(out=...)` | Reject ambiguous orthology and nested FASTA lists. MUSCLE returns None. |

11| Structures and residue features | `g2p` with exact accession -> inspect isoform/PDB map -> `pdb(resource="mmcif")` | Match chain/residue coverage; predicted features are not experimental validation. |

12| Expression | `archs4` or scoped `cellxgene` query | Human-only ARCHS4 correlations; species-specific tissue query; Census metadata counts cells. |

13| Enrichment | selected genes + measured background -> `enrichr` with full library -> adjusted-p-value filter | Species/library compatibility, mapped IDs, selection universe, and returned `adj_p_val`. |

14| Disease/drug exploration | human Ensembl ID -> `opentargets` resources | Dot-separated columns, local filtering after limit, partial pages, association versus causation. |

15| Cancer data | chosen cBioPortal studies -> `cbio_plot`; licensed local COSMIC TSV -> `cosmic` | Study denominators, project/release/assembly, download completeness. |

16| Reference preparation | `ref` with release -> CLI `-d` download -> checksums -> downstream index | `which` is comma-separated on CLI; Python has no download parameter. |

18Bundled scripts (run from the skill root):

20```bash

21python scripts/gene_analysis.py TP53

22python scripts/enrichment_pipeline.py selected.txt --background tested.txt --no-plot

23python scripts/enrichment_pipeline.py fly_genes.txt --species fly --database SPECIES_SPECIFIC_LIBRARY --no-plot

24python scripts/batch_sequence_analysis.py proteins.fasta --database swissprot

25```

27These network invocations are illustrative. Confirm that a species-specific

28library exists before submitting it; its example name is not a freshness claim.

29The script suite verifies local control flow with mocked responses. The gene

30pipeline depends on the Ensembl REST adapter, which failed during the current

31public probe. The batch script is for small exploratory queries and never runs

32AlphaFold; large searches should use local alignment tooling.

Common gget workflows

Use workflows.md for current 0.30.8 integration examples and their verification boundaries. Use database_info.md for service, authentication, pagination, and return-shape details.

Task	Sequence of operations	Check before interpretation
Gene characterization	`search` -> unique exact `gene_name` -> `info` / `seq`	Preserve Ensembl index; inspect service failures; distinguish genomic sequence from CDS.
Ortholog comparison	`bgee` -> select species/ortholog -> `seq` -> FASTA file -> `muscle(out=...)`	Reject ambiguous orthology and nested FASTA lists. MUSCLE returns None.
Structures and residue features	`g2p` with exact accession -> inspect isoform/PDB map -> `pdb(resource="mmcif")`	Match chain/residue coverage; predicted features are not experimental validation.
Expression	`archs4` or scoped `cellxgene` query	Human-only ARCHS4 correlations; species-specific tissue query; Census metadata counts cells.
Enrichment	selected genes + measured background -> `enrichr` with full library -> adjusted-p-value filter	Species/library compatibility, mapped IDs, selection universe, and returned `adj_p_val`.
Disease/drug exploration	human Ensembl ID -> `opentargets` resources	Dot-separated columns, local filtering after limit, partial pages, association versus causation.
Cancer data	chosen cBioPortal studies -> `cbio_plot`; licensed local COSMIC TSV -> `cosmic`	Study denominators, project/release/assembly, download completeness.
Reference preparation	`ref` with release -> CLI `-d` download -> checksums -> downstream index	`which` is comma-separated on CLI; Python has no download parameter.

Bundled scripts (run from the skill root):

python scripts/gene_analysis.py TP53
python scripts/enrichment_pipeline.py selected.txt --background tested.txt --no-plot
python scripts/enrichment_pipeline.py fly_genes.txt --species fly --database SPECIES_SPECIFIC_LIBRARY --no-plot
python scripts/batch_sequence_analysis.py proteins.fasta --database swissprot

These network invocations are illustrative. Confirm that a species-specific library exists before submitting it; its example name is not a freshness claim. The script suite verifies local control flow with mocked responses. The gene pipeline depends on the Ensembl REST adapter, which failed during the current public probe. The batch script is for small exploratory queries and never runs AlphaFold; large searches should use local alignment tooling.

Raw lines 1 to 32

Runs code python scripts/gene_analysis.py :21

Runs code python scripts/enrichment_pipeline.py :22

Runs code python scripts/batch_sequence_analysis.py :24