Glycobiology and Glycomics (GlyGen)

Glycan structure and protein-glycosylation research using GlyGen's 5 ToolUniverse tools: GlyGen_get_glycan, GlyGen_search_glycans, GlyGen_get_glycoprotein, GlyGen_search_glycoproteins, GlyGen_get_site.

LOOK UP, DON'T GUESS: Glycan composition, mass, and site-level glycosylation data are highly specific and not reliably recalled from memory — always resolve through these tools rather than asserting a glycan's structure or a protein's glycosylation status from general knowledge.

---

When to Use This Skill

Apply when users ask about:

• A specific glycan's structure, mass, monosaccharide composition, or species distribution (given a GlyTouCan accession, e.g. G17689DH)
• Finding glycans that match a mass range, monosaccharide-count range, or glycan type (N-linked / O-linked)
• A protein's overall glycosylation profile (site count, site positions, attached glycans, associated diseases/pathways) given a UniProt accession
• Finding glycosylated proteins in an organism, by gene name, or by glycosylation-evidence level
• Detailed information about one specific glycosylation site: exact position, flanking amino-acid sequence, the glycan attached there, and any SNVs co-located at that position

NOT for (route elsewhere):

• Resolving a gene/protein name to its UniProt accession in the first place -> tooluniverse-sequence-retrieval or tooluniverse-protein-structure-retrieval (GlyGen tools require a UniProt accession as input, not a gene symbol, except GlyGen_search_glycoproteins's gene_name filter)
• 3D protein structure -> tooluniverse-protein-structure-retrieval
• Non-glycan post-translational modifications (phosphorylation as a topic in its own right, ubiquitination, etc.) -> tooluniverse-protein-modification-analysis
• General small-molecule/chemical compound lookup -> tooluniverse-chemical-compound-retrieval

---

Tool Chains for Common Questions

"What glycans are attached to protein X?"

1. If you only have a gene symbol, resolve it to a UniProt accession first (tooluniverse-sequence-retrieval / tooluniverse-protein-structure-retrieval, or pass the gene name straight into GlyGen_search_glycoproteins's gene_name filter if you just need to confirm the protein is glycosylated at all).
2. GlyGen_get_glycoprotein(uniprot_ac=...) — returns glycosylation_count and a glycosylation_sites list, each entry giving position, residue, type (N-linked/O-linked), and the glytoucan_ac of the attached glycan.
3. For full structural detail on any glycan found in step 2, call GlyGen_get_glycan(glytoucan_ac=...).
4. For full detail on one specific site (flanking sequence, co-located SNVs), build a site_id as <uniprot_ac>.<start_pos>.<end_pos> (the canonical isoform accession from step 2's response usually needs a -1 suffix, e.g. P14210-1.294.294 — confirm the exact accession format from the glycoprotein response rather than assuming) and call GlyGen_get_site.

"Find glycans matching a mass or composition range"

1. GlyGen_search_glycans(mass_min=..., mass_max=..., monosaccharide_min=..., monosaccharide_max=..., glycan_type=...) — all filters are optional and combine as AND; omit filters you don't need. Returns a ranked list with hit_score.
2. GlyGen_get_glycan(glytoucan_ac=...) on any hit for full structural detail (composition breakdown, IUPAC/WURCS notation, species).

"Which proteins are glycosylated in organism X / by gene name?"

1. GlyGen_search_glycoproteins(organism_id=..., gene_name=..., glycosylation_type=..., glycosylation_evidence="reported") — organism_id is an NCBI taxonomy ID (9606 = human, 10090 = mouse, 10029 = Chinese hamster). Returns proteins ranked by hit_score with a glycosylation_count per protein.
2. Drill into any hit with GlyGen_get_glycoprotein for the full site list.

---

Response Shape Notes (observed from live calls, not assumed from schema)

• Every tool wraps its payload as {"status": "success", "data": {...or [...]}, "metadata": {"source": "GlyGen", "query": ..., "endpoint": ...}} on success, or {"error": "..."} on failure — check status/error before reading data.
• GlyGen_get_glycan's data includes a monosaccharide-composition array (each entry has name, residue, count, and a PubChem cid/url) and a byonic composition string (e.g. HexNAc(4)Hex(5)dHex(1)NeuAc(2) % 2368.84091476) in addition to the fields in the JSON schema.
• GlyGen_get_glycoprotein's data includes sequence_length, function, snv_count, and publication_count alongside the documented fields. Each entry in glycosylation_sites has position, residue, type, glytoucan_ac, and site_category (e.g. "reported_with_glycan") — this is the richest per-site summary and is often enough without a follow-up GlyGen_get_site call unless you need flanking sequence or SNV detail.
• GlyGen_search_glycoproteins's metadata includes total_results (may exceed the returned page) and list_id — use total_results to know whether to page further with offset.
• GlyGen_get_site's data includes a species array with common_name / reference_species, an all_sites_count (other sites reported on the same protein), and a categories array (e.g. ["glycosylation_flag", "snv_flag"]) indicating what kinds of annotations exist at that position.

See references/glygen_tool_reference.md for the full parameter table and captured example responses for all 5 tools.

Output

State the GlyTouCan/UniProt accessions used, the resolved values (mass, site positions, etc.), and cite GlyGen as the source. Do not report a glycan composition or site position that wasn't actually returned by a tool call.