concept · intermediate · partly written
Gene and protein names: the decoder
A decoder for gene and protein names: who assigns them, what each pattern encodes, and where a name misleads.
Also called: gene nomenclature, how to read a gene name
The one rule
A name is a historical label under a priority rule, not a measurement. Join data on the stable ID (for human genes, the HGNC ID), and read the symbol as a readable layer on top. 12
Reported, not yet replicated
Who names what more
HGNC symbols use uppercase Latin letters and digits only; each gene gets one symbol, an HGNC ID and a descriptive name; names do not contain a species or the letter G; and HGNC no longer names phenotypes. Mouse symbols are maintained by MGI (first letter capital, the rest lowercase, with an exception for genes known only from a recessive mutant). Protein names are guided by the IUBMB/JCBN guidelines, which exclude molecular weight from a recommended name. 3145
Reported, not yet replicated
Three layers for one thing more
For one gene there is a symbol (TP53), a protein name (UniProt: Cellular tumor antigen p53) and lab names (p53, LFS1). HGNC keeps the lab names as aliases, and records previous symbols (CDK1 was CDC2 until 2009-12-16). 267
Reported, not yet replicated
Pattern: mass in the name (p53, p47) more
Encodes apparent gel mass from the 1970s-80s. Reliability: low. Example: p53 is 53,000 apparent against 43,653 Da calculated. Exception: p47 is one label on several genes. See the linked record. 879
Reported, not yet replicated
Pattern: family stem plus number (CDK1, CDK2) more
The stem is a usable guide to family; the number is a proposal-time label, not a similarity rank. Reliability: stem medium to high, number low. Exception: CDK1 was CDC2 and carried a yeast phenotype name before. See the linked record. 106
Reported, not yet replicated
Pattern: placeholders (C#orf#, FAM#, KIAA#) more
Encodes 'no better name was available at approval'. Reliable as a sign the symbol carries no function. 427 protein-coding genes still use one per HGNC's 2026 paper. See the linked record. 11
Reported, not yet replicated
Pattern: disease in the name more
TP53 has the alias LFS1 and the alias name Li-Fraumeni syndrome, and the p47 protein (NCF1) was found as a factor missing in autosomal chronic granulomatous disease. The name records how the gene was found, not what it does. Reliability: low. 212
Reported, not yet replicated
Pattern: cluster of differentiation (CD#) more
Encodes an antibody cluster from the HLDA workshops (first held in Paris in 1982), kept in gene names as an alias or previous symbol (CD20 is MS4A1). Reliability: medium as a cell-marker label, low as a gene-family guide; CD8 and CD1 each cover several genes. See the linked record. 1314
Reported, not yet replicated
Pattern: solute carriers (SLC) more
SLC + family + letter + member: the family is shared homology, not shared cargo, per HGNC's own post. Reliability: family medium, member low. SLC21 became SLCO. See the linked record. 15
Reported, not yet replicated
Pattern: receptor and ligand names more
Parallel stems (TNFSF10, TNFRSF10A) or a root plus R (IL6, IL6R), set family by family with specialist groups; no general rule found. Reliability: medium. See the linked record. 114
Reported, not yet replicated
Pattern: subunit and isoform suffixes more
A trailing letter can be a Greek subunit (PRKACA), a duplicate gene (CDK11A/B) or a split antigen (CD8A/B); isoforms of one gene usually share one symbol. Reliability: medium. See the linked record. 114
Reported, not yet replicated
Pattern: EC numbers more
A four-part code by the reaction catalysed (EC 2.7.11.22 is cyclin-dependent kinase); a real hierarchy of chemistry, not of sequence. Reliability: high for reaction, nothing for relatedness. See the linked record. 1617
Reported, not yet replicated
Pattern: capitals and italics more
Human symbols are uppercase: the HGNC guidelines page says symbols contain only uppercase Latin letters and Arabic numerals. What case means in other species is below. Other organisms have their own rules (worm, fly, yeast, zebrafish); they are listed in our own notes, but their authority pages were not opened, so none is stated here. 14
Reported, not yet replicated
Whether HGNC is the only namer, and since when deep
This record previously said HGNC "describes itself as the sole body responsible for naming human genes since 1979". The HGNC guidelines page was re-opened on 2026-11-19 and carries no such statement and no date: it was the wrong page for that claim. The wording may come from Bruford 2022, which has not been opened. Until a source is opened that carries it, neither the "sole body" wording nor the 1979 date is confirmed here, and this section teaches nothing. 3
Unverified: from our own report, source not opened
Mouse case, and why it is not stated here yet deep
This record previously said mouse symbols are a capital then lowercase, with a lowercase first letter marking a gene known only from a recessive mutant, and attributed it to "the two nomenclature pages". One of those two, the HGNC guidelines page, says nothing about non-human case; it was re-opened on 2026-11-19 to check. The convention should rest on the MGI page alone, which has not been re-opened for this. The rule is probably right and is not confirmed here, so nothing is taught from it. 4
Unverified: from our own report, source not opened
Where names fail more
Symbols change (CDC2 to CDK1; SEPT# to SEPTIN# in 2019), one label can name several genes (p47), aliases can collide (OX1R and OXR1 in at least 79 papers), and a spreadsheet can turn a symbol into a date. HGNC keeps a symbol even when its name is corrected (FMR1 stayed in use in over 4,000 publications). 6181939
Reported, not yet replicated
Looking a name up more
HGNC's tools include a search, HCOP (orthology predictions), a Multi-Symbol Checker and a BioMart service, all freely downloadable (REST included); UniProt gives protein names and calculated mass. Per-organism sites for worm, yeast and zebrafish were not opened in this run. 117
Reported, not yet replicated
His framing
He said that knowing a few naming patterns, like p53 or p47, and the repeating elements such as CDK1 and CDK2 would let a reader get insight faster, and that some very rare genes may not yet be understood or named. 20
Amit's hypothesis
The sourced account beside it more
The patterns he names are real and mostly follow the rule above: the p-number is an apparent mass that is not the real mass, and the CDK stem means cyclin-dependent kinase. The numbers after a stem are not a reliable similarity guide in the sources read. On unnamed genes, HGNC counts 427 protein-coding genes still on placeholder symbols. 871011
Reported, not yet replicated
Still open here
- Patterns still not written with sources: names from fly and worm mutant phenotypes carried into human genes, and the worm, yeast, fly and zebrafish nomenclature authorities (their pages block automated fetch and were not worked around). The IUPHAR/BPS Guide to Pharmacology pages for receptor names need a login and were not opened.
- Is there a published benchmark of how well a gene name predicts function? None was found.
- The claim that HGNC is "the sole body responsible for naming human genes since 1979" is not on the HGNC guidelines page, which was re-opened on 2026-11-19 and makes no such statement and gives no date. The claim may rest on Bruford 2022, which has not been opened. Until it is, treat the 1979 date and the "sole body" wording as unconfirmed.
- The mouse capital-then-lowercase convention is attributed here to "the two nomenclature pages", but the HGNC guidelines page does not address non-human case. It should rest on the MGI page alone, which has not yet been re-opened for this.
Not yet written: Plain answer, Why it matters, Equation, Real numbers.
Linked from
- Biomedical knowledge graphs: shapes, edges and what has worked
- Cluster of differentiation (CD) names
- Context-dependent edges: storing the conditions under which a statement holds
- EC numbers (enzyme classification)
- How entity grounding works and fails in knowledge graphs
- Family stem plus number (CDK1, CDK2)
- The Gene Ontology (GO)
- Joining host genetics to pathogen genomics: paired genome studies and where the join breaks
- How agentic science systems hold what they know
- Kosmos and its 'structured world model'
- Mass-based protein names (p53, p47)
- What the data object is, layer by layer
- How a pathway is encoded: BioPAX, SBML, GO-CAM and Reactome's own model
- Placeholder gene symbols (C#orf#, FAM#, KIAA#)
- "Quantum fingerprint": the metaphor, quantum-chemistry descriptors and quantum hardware, kept apart
- Receptor and ligand names
- Robin, FutureHouse's multi-agent discovery system
- Solute carrier (SLC) names
- Spreadsheet gene-name errors and the symbol renaming
- How MOFA+, WNN and totalVI join two layers, and how they are benchmarked
- Subunit and isoform suffixes
- The Virtual Biotech
- Translation initiation: scanning and start-codon choice
- HGNC (HUGO Gene Nomenclature Committee)
- MGI (Mouse Genome Informatics) nomenclature
- Elspeth A. Bruford
- Can paired host-genetics and pathogen-genomics results share one schema with interaction resources?
- How many human protein-coding genes still have no real name?
- What fraction of genes and proteins is lost or split when mapping between Ensembl, RefSeq, HGNC and UniProt?
Sources
- HGNC. HGNC guidelines for human gene nomenclature. genenames.org (2026). organisation-page, not peer reviewed; opened 2026-11-19 all records citing this
- HGNC. HGNC record for TP53 (REST fetch). rest.genenames.org (2026). database, not peer reviewed; opened 2026-11-19 all records citing this
- Bruford, E., on behalf of the HGNC. Comment on Herring et al. The Use of Retardation in FRAXA, FMRP, FMR1 and Other Designations. Cells 11 (2022). paper; opened 2026-11-17 all records citing this
- MGI. Mouse Genome Informatics: guidelines for nomenclature of genes. informatics.jax.org. organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- IUBMB. Guidelines for recommended protein names (IUBMB/JCBN). iubmb.qmul.ac.uk. organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- HGNC. HGNC record for CDK1 (REST fetch). rest.genenames.org (2026). database, not peer reviewed; opened 2026-11-19 all records citing this
- UniProt. UniProtKB P04637 (TP53, human): Cellular tumor antigen p53. rest.uniprot.org (2026). database, not peer reviewed; opened 2026-11-19 all records citing this
- DeLeo, A. B., et al.. Detection of a transformation-related antigen in chemically induced sarcomas and other transformed cells of the mouse. PNAS 76 (1979). paper; opened 2026-11-17 all records citing this
- HGNC. HGNC search for alias symbol p47 (REST). rest.genenames.org (2026). database, not peer reviewed; opened 2026-11-19 all records citing this
- Malumbres, M., Harlow, E., Hunt, T., et al.. Cyclin-dependent kinases: a family portrait. Nature Cell Biology 11 (2009). paper; opened 2026-11-17 all records citing this
- Seal, R. L., Braschi, B., Gray, K., et al., Bruford, E. A.. Genenames.org: the HGNC and PGNC resources in 2026. Nucleic Acids Research 54 (2025). paper; opened 2026-11-17 all records citing this
- Volpp, B. D., Nauseef, W. M., Clark, R. A.. Two cytosolic neutrophil oxidase components absent in autosomal chronic granulomatous disease. Science 242 (1988). paper; opened 2026-11-17 all records citing this
- HCDM. HCDM: Human Cell Differentiation Molecules (home and workshop history). hcdm.org (2026). organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- HGNC. HGNC REST symbol records read on 2026-11-17 (SLC, CD, receptor, subunit and CDK genes). rest.genenames.org (2026). database, not peer reviewed; opened 2026-11-19 all records citing this
- HGNC. HGNC blog: solute carrier families (31 March 2023). HGNC blog (2023). organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- IUBMB. Enzyme nomenclature: introduction and rules (IUBMB). iubmb.qmul.ac.uk. organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- IUBMB. EC 2.7.11.22: cyclin-dependent kinase (IUBMB enzyme list). iubmb.qmul.ac.uk. database, not peer reviewed; opened 2026-11-17 all records citing this
- HGNC. The risks of using unapproved gene symbols. HGNC blog (2021). organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- Amit's words about the atlas, kept verbatim in this repository. Kept in the atlas repository; not published.
Where it sits on the map
History of this record (11)
This record has been changed 11 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.
- Status
- partly written
- Last updated
- 2026-11-19
- Record
- content/concepts/gene-and-protein-naming-decoder.json