Forty horse genes for ProtNLM evaluation
40 selected horse genes with 89 GO predictions and 17 function descriptions.
The paired reviews evaluate these horse predictions using exact sequence evidence
and characterized mammalian counterparts. The selection is based
on informative functional claims and biological diversity. Every gene has GO or
function-text output; protein-name predictions were not a selection criterion.
Download the 40-gene list ·
All 106 GO/function outputs ·
Current protein sequences ·
Selection manifest
27 genes have GO predictions, 17 have function descriptions, and four have both.
The list includes catalytic/regulatory distinctions, substrate and pathway
specificity, taxonomic context, localization and developmental claims. The focus
column records the review question, not a correctness verdict. Each target is paired
with a human review and evaluated against primary evidence and its exact current
horse sequence. See the review findings and evidence gaps.
| # | Horse review | Human review | Accession | GO | Function text | Review focus |
|---|---|---|---|---|---|---|
| 1 | VAPA | VAPA | A0A3Q2H1L9 | 0 | 1 | Test sperm-crawling narrative and taxonomic context |
| 2 | CAPSL | CAPSL | A0A3Q2I3U9 | 0 | 1 | Test venom-secretion narrative |
| 3 | MYL10 | MYL10 | A0A9L0TJE1 | 0 | 1 | Test venom-secretion narrative on a different protein family |
| 4 | KRIT1 | KRIT1 | A0A9L0SR44 | 0 | 1 | Test piRNA and germline narrative against target identity |
| 5 | CTDSP2 | CTDSP2 | F7A4N8 | 1 | 0 | Resolve predicted kinase activity versus target catalytic class |
| 6 | PEA15 | PEA15 | A0A9L0RWM8 | 1 | 0 | Test carbohydrate-transport process assignment |
| 7 | ALDH7A1 | ALDH7A1 | A0A9L0RRL6 | 1 | 0 | Resolve precise aldehyde-dehydrogenase substrate specificity |
| 8 | GHSR | GHSR | F6QF00 | 0 | 1 | Test oxytocin-receptor narrative and ligand specificity |
| 9 | ALG5 | ALG5 | A0A5F5PM72 | 0 | 1 | Resolve sugar donor and dolichol-glycosylation reaction |
| 10 | HSPD1 | HSPD1 | F6Z587 | 0 | 1 | Distinguish chaperonin complexes and transferred client biology |
| 11 | CXCR3 | CXCR3 | A0A9L0T1D1 | 0 | 1 | Resolve chemokine ligand class and specificity |
| 12 | DARS2 | DARS2 | A0A9L0SB67 | 0 | 1 | Test tRNA-Asn charging and organellar substrate specificity |
| 13 | GPAM | GPAM | A0A9L0TTC1 | 0 | 1 | Distinguish acyl-ACP from acyl-CoA donor chemistry |
| 14 | HSPA4 | HSPA4 | A0A9L0S5Z5 | 0 | 1 | Resolve ribosome-associated chaperone-family and complex transfer |
| 15 | DUOX1 | DUOX1 | A0A9L0SQG9 | 4 | 1 | Separate hydrogen peroxide production, catabolism and thyroid context |
| 16 | MTMR9 | MTMR9 | A0A9L0T3C1 | 7 | 0 | Separate phosphatase regulation and binding from intrinsic catalysis |
| 17 | PTPRN2 | PTPRN2 | A0A9L0T4W6 | 1 | 0 | Assess dephosphorylation process for a phosphatase-like protein |
| 18 | DNMT3L | DNMT3L | A0A9L0T837 | 2 | 0 | Compare regulatory function with active DNA methyltransferase |
| 19 | IRAK3 | IRAK3 | A0A3Q2HDT6 | 0 | 1 | Assess receptor-signaling narrative and inactive-kinase context |
| 20 | PPP4R4 | PPP4R4 | A0A9L0S961 | 3 | 0 | Separate phosphatase-regulator function from developmental transfer |
| 21 | CDK7 | CDK7 | A0A9L0R074 | 2 | 0 | Evaluate specific kinase activity and phosphorylation process |
| 22 | MAP2K2 | MAP2K2 | A0A9L0SHX8 | 1 | 0 | Evaluate conserved kinase process and specificity |
| 23 | ZDHHC23 | ZDHHC23 | A0A9L0T4E4 | 1 | 0 | Evaluate specific protein S-palmitoyltransferase activity |
| 24 | CH25H | CH25H | F6T000 | 3 | 0 | Evaluate oxidoreductase and lipid-biosynthesis specificity |
| 25 | BCAT2 | BCAT2 | A0A9L0TSN4 | 0 | 1 | Evaluate branched-chain amino-acid catabolic narrative |
| 26 | SIRT5 | SIRT5 | F6S899 | 1 | 1 | Evaluate deacylation substrate classes and weak in-vitro activity caveat |
| 27 | USP8 | USP8 | A0A9L0T7K6 | 1 | 1 | Separate deubiquitination function from protein-catabolism process |
| 28 | OMA1 | OMA1 | A0A9L0R9P8 | 1 | 0 | Evaluate proteolysis prediction and more specific mitochondrial role |
| 29 | EFR3A | EFR3A | A0A9L0S4L8 | 1 | 0 | Evaluate protein recruitment to plasma membrane |
| 30 | CACNB3 | CACNB3 | A0A5F5PZM5 | 1 | 0 | Distinguish ion-channel regulation from ion transport catalysis |
| 31 | SHLD2 | SHLD2 | A0A9L0RGD6 | 6 | 1 | Evaluate DNA-repair directionality, class switching and localization |
| 32 | GEMIN5 | GEMIN5 | A0A9L0R5P7 | 6 | 0 | Compare RNA-binding and translation claims with ubiquitination |
| 33 | AFAP1L2 | AFAP1L2 | A0A9L0RQI4 | 9 | 0 | Evaluate adaptor binding, kinase activation and cytokine processes |
| 34 | OLFML2A | OLFML2A | A0A9L0SKW1 | 4 | 0 | Separate extracellular-matrix location, binding and organization |
| 35 | CC2D2A | CC2D2A | A0A5F5PJ44 | 4 | 0 | Evaluate developmental-process transfer and ciliary mechanism |
| 36 | WDPCP | WDPCP | A0A3Q2KRK8 | 4 | 0 | Evaluate cilium and cytoskeleton localization with projection organization |
| 37 | TRAF2 | TRAF2 | F7BIV4 | 5 | 0 | Evaluate receptor binding, complex membership and immune context |
| 38 | WEE1 | WEE1 | F6TY09 | 7 | 0 | Separate kinase-related processes from oocyte-specific context |
| 39 | DYNLT2B | DYNLT2B | A0A9L0SWY1 | 10 | 0 | Evaluate ciliary transport, dynein binding and localization claims |
| 40 | DNMT3A | DNMT3A | A0A9L0TK01 | 2 | 0 | Compare gene-expression regulation with DNMT3L on the same output terms |
Identity and evidence
There is one accession per selected gene label and no duplicate current protein
sequences. SHLD2 uses A0A9L0RGD6; the additional release record A0A3Q2HUD4 is not a
second benchmark gene. Symbols come from the frozen official accession list.
Some current UniProt records omit their gene-name field; that absence is retained
in the CSV rather than filled from the model's predicted name.
The ordinary UniProt record snapshot
provides species, sequence versions, lengths and sequence checksums. All 40 are
horse records and all current lengths match the release list. This does not prove
that every record is a complete protein or that the prediction-time sequence is
unchanged; examine gene models and domain completeness during review.
The CSV identifies existing human, mouse and rat review files with matching gene
symbols as evidence-search leads. These are not verified orthology assignments or
independent biological validation. Establish the counterpart and trace its
experimental/analytical evidence before transferring a claim to horse.
This is a targeted, retrospective cohort, not a random horse sample or an estimate
of whole-proteome accuracy. A later mammalian benchmark can grow from these
horse-anchored cases after orthology and evidence are established. The original
ARGO-ProtNLM-50 is unchanged.