gad1a notes (Danio rerio, glutamate decarboxylase 1a; UniProt A0A8M1RDR2)

2026-09-28 — session log (DANRE_DUPLICATION batch 4, random TGD_tree sample)

Deep research: not available (Edison/Falcon 402 Payment Required; OpenAI key invalid). Not
attempted, per instructions. Literature searched by hand in Europe PMC (gad1a AND zebrafish,
"gad1b" AND zebrafish AND (mutant OR knockout OR morphant OR crispr),
TITLE:"glutamate decarboxylase" AND (zebrafish OR teleost OR fish), author searches for
VanLeuven/Lauderdale and Filippi) and by following ZFIN expression publications.
Pair material: gad1a-bioinformatics/RESULTS.md (protein, expression, synteny, probe identity);
paralog notes in ../gad1b/gad1b-notes.md.

Accession A0A8M1RDR2 (TrEMBL, RefSeq XP_002663350, 591 aa), ZFIN:ZDB-GENE-070912-472, Ensembl
ENSDARG00000093411, chr9. 8 GOA rows, all IBA or IEA; no experimental annotation.

Origin

• PANTHER TGD_tree (Neopterygii|Teleostei), 1:1, one gar co-ortholog; Ensembl Compara duplication
  node Clupeocephala; one gar GAD1 ortholog for both copies.
• Chromosomes: PMID:30200754
• Other fish also have two gad1 genes: PMID:30200754
• My synteny check: tlk1a/tlk1b and dync1i2a/dync1i2b flank both copies (RESULTS.md).
• Medaka: Ensembl lists a one-to-one medaka ortholog only for gad1a; PANTHER lists two medaka
  co-orthologs. Unresolved.

Protein (RESULTS.md)

• 86.3% identical to gad1b; 80.0% to human GAD1 (gad1b 82.7%).
• PLP lysine (human K405), GABA-binding residues and C-terminus kept. Human phospho-S78 is T in
  gad1a. Relative rate vs gar not significant (37 vs 23 unique changes).
• Mammalian GAD1 background: PMID:17384644
  PMID:1549570
  PMID:27461130
• No enzyme assay, antibody or tagged-protein data for Gad1a:
  PMID:30200754

Expression

• Public data (RESULTS.md): E-ERAD-475 whole embryo 0.4-4 TPM from segmentation to day 5 vs gad1b
  up to 84 TPM; Bgee: brain, liver, early embryo, blastula, gastrula, larva, head, bone (RNA-seq only).
• Single-cell RNA-seq of lateral-line efferent neurons:
  PMID:41950195
• Lueffe et al. compared both paralogs by in situ:
  PMID:34650032
  Their probe: PMID:35769333
  Caveat: the sequenced Martin et al. GAD67 cDNA (AF017266) is gad1b by sequence (99.6% vs 90.9%
  protein identity; probe_identity.py). Another lab calls the old gad67a/gad67b probes gad1b:
  PMID:26896392
  So the "gad1a" in situ data of Lueffe et al., and ZFIN's curation of gad1a brain expression from
  them, may be gad1b signal. Not resolvable without the probe sequence.
• Other statements of similarity are secondary: PMID:41097006;
  PMID:42466317
  The cited Filippi et al. 2014 (PMID:24374659) used gad1b and gad2 probes, not gad1a, so the second
  statement is not supported by its source.
• Differential regulation (qPCR, which uses paralog-specific primers):
  PMID:35769333
  PMID:34650032

Function

• No gad1a mutant, morphant or assay. A gad1b translation-blocking MO has 19/25 matches to gad1a:
  PMID:30200754
  The authors predict that double nulls would show the craniofacial MO phenotype:
  PMID:30200754

Annotation decisions

• MF rows (GAD activity IBA/IEA, carboxy-lyase, C-C lyase, PLP binding): ACCEPT; catalytic residues kept.
• GABA biosynthetic process (IBA): ACCEPT (the enzyme performs the step).
• Cytoplasm (IBA): ACCEPT. Presynaptic active zone (IBA): KEEP_AS_NON_CORE (rodent Gad1 donors; no
  Gad1a localization data).
• No NEW annotations: nothing gad1a-specific has been measured.