K-Dense-AI/scientific-agent-skills / skills/glycoengineering
glycoengineering
VerifiedAnalyzes and engineers protein glycosylation by scanning canonical N-glycosylation sequons, describing S/T-rich regions, checking curated glycan evidence, and preparing NetNGlyc, NetOGlyc and GlycoSHIELD workflows. Use for glycoprotein engineering, antibody Fc glycosylation, glycan shielding, and site-specific glycoproteomics interpretation.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.7 KB |
| references/glycan_databases.md | markdown | 8.2 KB |
| references/glycoshield.md | markdown | 6.4 KB |
| scripts/glycoengineering_tools.py | script | 4.8 KB |
| scripts/glytoucan_lookup.py | script | 2.6 KB |
references/glycan_databases.md
View on GitHub# Glycan databases, evidence and notation
Reviewed 2026-10-01. Public read-only probes below did not use credentials or
submit sequences, structures or private data. A successful transport probe is not
evidence that a database's scientific content is current or complete.
## GlyTouCan: accession to WURCS
[GlyTouCan](https://glytoucan.org/) assigns identifiers to glycan descriptions,
including descriptions with unresolved features. An accession alone does not
establish a fully resolved structure or biological occurrence.
The former example `https://api.glytoucan.org/glycan/{id}` is unsupported here
(the G00055MO probe returned 404). Use the project's documented
[SPARQL structure query](https://code.glytoucan.org/rdf-ontology/sparql-queries/),
restricted to the desired accession:
```python
from glytoucan_lookup import lookup_wurcs
sequences = lookup_wurcs("G00055MO")
print(sequences)
# Live review returned one sequence beginning WURCS=2.0/2,2,1/...
```
Import [scripts/glytoucan_lookup.py](../scripts/glytoucan_lookup.py) from the
skill's scripts directory. Tested with requests 2.34.2.
| Contract | Verified behavior |
| --- | --- |
| Request | Public `GET https://ts.glytoucan.org/sparql`; URL-encoded `query` and `format=application/sparql-results+json`, plus the matching `Accept` header. No API version path or credentials used. |
| Data selection | Named graphs `http://rdf.glytoucan.org/core` and `http://rdf.glytoucan.org/sequence/wurcs`; exact `VALUES ?PrimaryId` match. Accession validation prevents interpolating arbitrary SPARQL. |
| Response | JSON `head.vars` and `results.bindings`; each binding has `PrimaryId.value` and `Sequence.value`. Keep WURCS uncertainty. |
| Pagination | Single-accession lookup has no client pagination or artificial LIMIT. Bulk queries use SPARQL `ORDER BY` and explicit `LIMIT`/`OFFSET`, not a REST cursor. Do not claim a changing dataset was completely enumerated without a stable snapshot. |
| Failure | HTTP, timeout, invalid JSON and schema errors remain errors. An empty binding list means no matching WURCS in the queried graphs, not proof of absence from GlyTouCan. |
The helper sets connection/read timeouts, raises for HTTP failures, checks the
response type, validates the returned accession, and retains all distinct WURCS
sequences. It does not validate WURCS chemistry or register glycans.
## GlyConnect: protein, site and glycan evidence
[GlyConnect](https://glyconnect.expasy.org/) links glycan and protein evidence.
Use its search interface with a UniProt accession, then check species, isoform,
site, glycan/composition, experimental context and publication. The homepage
reports version 2.6.0 dated 2025-11-21; older linked pages still show 1.3.0.
Do not infer one synchronized release across these surfaces.
REST access is currently **unverified**: the official
[Swagger entry point](https://glyconnect.expasy.org/api/docs) redirects with a
schema URL `/api/assets/swagger.json`, which returned 404. The old
`/api/proteins/uniprot/P00533` returned 500; `/api/compositions/protein/1` returned
404. These failures establish neither a supported schema nor no biological data.
The previous functions returning `{}` or `[]` on these errors have been removed.
Do not invent replacements, response envelopes, auth requirements or pagination.
The official [RDF documentation](https://glyconnect.expasy.org/rdf) supports a
SPARQL alternative. This **transport check**, with explicit response negotiation,
was executed and returned one RDF schema triple:
```python
import requests
response = requests.post(
"https://glyconnect.expasy.org/sparql",
data="SELECT * { ?s ?p ?o } LIMIT 1",
headers={"Content-Type": "application/sparql-query",
"Accept": "application/sparql-results+json"},
timeout=(10, 45),
)
response.raise_for_status()
if not response.headers.get("Content-Type", "").startswith("application/sparql-results+json"):
raise ValueError("SPARQL service did not return negotiated JSON")
rows = response.json()["results"]["bindings"]
print(rows)
```
It is a read query despite HTTP POST; no credentials were needed. Form-encoded
`format=json` alone returned CSV during review, so retain the headers above.
JSON follows SPARQL binding objects, not a `data` list. Pagination, when needed,
belongs in the query with stable ordering and `LIMIT`/`OFFSET`; no cursor was
advertised. The live endpoint returned no rows for a bounded UniProt-object probe
and no named graphs for a bounded graph query. Linked protein/site sample query
files also failed. **Protein/site/composition retrieval was not established**;
a successful generic query must not be reported as a successful glycoprotein
lookup. Consult the current ontology/examples and verify actual records before
automating scientific extraction.
## Other resources and tools
| Resource | Use and limits |
| --- | --- |
| [KEGG GLYCAN](https://www.genome.jp/kegg/glycan/) | Glycan G-number records linked to biosynthesis/pathway maps; composition, name and ID search. A pathway association does not demonstrate occupancy on a given protein. |
| [CAZy](https://www.cazy.org/) | Carbohydrate-active enzyme families for enzyme candidate research. Family membership does not uniquely determine substrate specificity. Direct page access returned 403 during review, so current data/release was not verified. |
| [UniCarb-DR](https://unicarb-dr.glycosmos.org/) | Non-curated repository of annotated glycomic MS/MS data with MIRAGE metadata, distinct from curated UniCarb-DB. Review original spectra and experimental context. |
| [GlycoWorkbench archive](https://code.google.com/archive/p/glycoworkbench/) | Legacy Java glycan drawing/mass-spectrum annotation tool; use the official archive linked by UniCarb-DR's [MIRAGE protocol](https://unicarb-dr.glycosmos.org/assets/samplefiles/MIRAGEprotocolv2.0.pdf), not third-party installer mirrors. Modern OS/Java execution was not tested. |
| [Byonic](https://www.proteinmetrics.com/products/byonic) | Commercial glycopeptide database search and spectral annotation. It is not generally a de novo full glycan-structure solver; see the entrypoint's composition/localization caveats. |
| [Mascot](https://www.matrixscience.com/help/o-fucosylated_cid_spectra.html) | Modification-aware peptide search; glycan neutral-loss behavior and residue specificity must be modeled. A generic modification hit is not complete glycan structural assignment. |
| [GlycoMine study](https://pubmed.ncbi.nlm.nih.gov/25568279/) | Historical N-, C- and O-linked site predictor for the human proteome. The publication does not justify claiming separate validated predictions for every O-glycan initiation type; a current hosted service was not verified. |
The former `unicarbkb.org` URL now serves unrelated educational/product content;
it was not retained as an active curated glycan database. The previous unsourced
“SymLink” predictor and generic Skyline glycan-database integration claim could
not be verified from primary documentation and are not actionable workflows here.
No REST routes are asserted for these catalog resources.
## Nomenclature and reporting
Use [SNFG](https://www.ncbi.nlm.nih.gov/glycans/snfg.html) for symbolic figures:
| Residue | Symbol |
| --- | --- |
| Glc / glucose | Blue circle |
| Man / mannose | Green circle |
| Gal / galactose | Yellow circle |
| GlcNAc / N-acetylglucosamine | Blue square |
| GalNAc / N-acetylgalactosamine | Yellow square |
| Neu5Ac / N-acetylneuraminic acid | Purple diamond |
| Fuc / fucose | Red triangle |
Oxford-style Fc shorthand such as G0F, G1F, G2F and G2FS1 describes common glycan
classes. Define the convention, antennae, core fucose, bisection and sialic-acid
linkages; shorthand can leave isomers unresolved. M5 denotes Man5GlcNAc2, whereas
a generic Hex5HexNAc2 composition alone does not identify every monosaccharide or
linkage. Store full WURCS/GlycoCT/IUPAC descriptions when supported by evidence.
For protein sites, always state the numbering system and sequence accession:
precursor, mature protein, isolated Fc, fusion construct and EU positions differ.
Avoid transferring therapeutic site numbers or glycosite counts between constructs
without sequence alignment and experimental evidence. Record observed occupancy,
possible sequons and ambiguous glycopeptide localizations separately.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |