The AADAC family has five human members and only two of them, AADAC and AADACL1/KIAA1363/NCEH1, have been functionally characterised; AADACL2, AADACL3 and AADACL4 have no assigned function in the literature.

AADAC, the family's name-giving member, is a type II membrane glycoprotein anchored in the ER with its catalytic domain in the lumen. This is the topology whose N-terminal signal anchor is the source of the propagated membrane annotation on AADACL2.

The family straddles bond types - AADAC hydrolyses both amide bonds of xenobiotic arylacetamides and ester bonds of endogenous diglycerides - so predicting a carboxylic ester hydrolase activity for AADACL2 from family membership does not exclude amidase activity.

The catalytic domains of both characterised family members are structurally homologous to hormone-sensitive lipase, which places AADACL2 in the GDXG/HSL-like branch of the alpha/beta-hydrolase fold.

An exome-wide insertion/deletion screen in 32,043 Chinese Han individuals named AADACL2 among previously unreported psoriasis susceptibility genes. This is a genetic association only, with no functional follow-up, but it is tissue-coherent with the gene's skin-restricted expression.

A genome-wide association study of ischemic stroke in West African populations found genome-wide significant signals near AADACL2 on chromosome 3, reported as protective and novel. The association is to a neighbouring locus rather than to the coding sequence.

The affinage record for AADACL2 is empty - zero discoveries and zero citations - despite passing its own quality gates. Consistent with the absence of any functional literature, but treated here as a lead rather than as evidence, and checked independently against PubMed and the UniProt reference list.

The catalytic apparatus is intact and independently corroborated. Eight of the fifteen WITH/FROM sources with annotated active sites - the seven closest by identity plus Mycobacterium tuberculosis LipN - map all three of their catalytic residues onto AADACL2's own annotated Ser189, Asp341 and His371. Those eight are human, mouse and rat AADAC, mouse Nceh1, the three M. tuberculosis carboxylesterases NlhH, LipI and LipN, and Arabidopsis CXE18. Every source that fails is at 26.5 percent identity or below, where global alignment loses register, so the failures are not evidence of residue loss.

Of the 14 reviewed UniProtKB entries carrying the IPR017157 family signature that generated the membrane annotation, 12 have a transmembrane segment - 11 of them annotated as a type-II signal anchor - and AADACL2 is the only one whose equivalent segment is annotated as a cleaved signal peptide, so the membrane call transfers a feature the target record says it lacks.

The sequence nevertheless does not settle the topology. The annotated cleavage site puts histidine at -1, which signal peptidase I does not accept, and AADACL2's N-terminal hydrophobic core scores higher on peak Kyte-Doolittle hydropathy (1.94) than AADAC's experimentally validated type-II signal anchor (1.81).

Mass spectrometry cannot adjudicate the topology either. All four observed peptides lie in the catalytic domain, the earliest starting at residue 124, so none spans the annotated cleavage site or begins at the putative mature N-terminus.

The PANTHER node placement is inverted with respect to what transfers safely - the well-conserved mechanistic activity sits only at ortholog-specific nodes while the poorly-conserved localisation sits at the shared family node, so moving GO:0017171 down and dropping GO:0016020 would correct AADACL2, AADACL3 and AADACL4 together.

No candidate term below GO:0016787 survives every donor of PTN009058710, so the generic term PAINT chose is the exact least common ancestor of the donor set rather than an under-specific approximation to it.

The two more specific candidates fail on different axes: the two kynurenine formamidases put GO:0052689 out of reach, and soybean HIDH's threonine nucleophile puts GO:0017171 out of reach.

Soybean HIDH is bifunctional rather than a pure lyase, holding an IDA carboxylesterase annotation as well as its dehydratase activity, so it does not threaten the hydrolase parent at all.

The mechanism term is true of every donor PAINT cites at the family node PTN009058713 and held by IDA in all three of them, which is where it should be placed instead of at the deep node.

The Thr164 call that the mechanism verdict turns on is a fold-position inference rather than a UniProt nucleophile label, but the elbow pentapeptide corroborates it at the sequence level and a sensitivity analysis shows the verdict does not depend on it.

The identity of the WITH/FROM set across all three paralogs is measured rather than assumed, from two independent sources for every gene.

The protein carries a complete serine-hydrolase catalytic triad, with the nucleophile at Ser189 in a GDSSG elbow, plus an oxyanion-hole motif at 111-113 and a conserved Cys116-Cys338 disulfide.

UniProt assigns AADACL2 to the GDXG lipolytic enzyme family, the hormone-sensitive-lipase-like branch of the alpha/beta-hydrolase fold, and gives it the incomplete EC number 3.1.1.- rather than a specific reaction.

The subcellular location is given as Secreted with no evidence tag, and rests on a predicted cleaved signal peptide covering residues 1-18.

Expression is skin-enriched, and the protein is detected at the protein level, but it remains pharmacologically undrugged and uncharacterised (Pharos Tdark).

Isoform 2 is expected to be degraded by nonsense-mediated decay and truncates before the catalytic serine, so it cannot carry the family activity.