Alliance automated description regressions

Species: worm, yeast, DROME

Alliance automated descriptions that lost content

For these worm genes the current Alliance automated gene description (fetched 2026-10-29) is missing
function content that the RefSeq copy of an earlier Alliance release still carries. Texts are
quoted verbatim from genes/worm/<gene>/<gene>-descriptions.yaml. Found during the description
review in MOD_DESCRIPTIONS. The RefSeq bracketed provenance line gives the release.

Check each case before filing it. The automated text is generated from GO annotations, so a
loss can follow an annotation that was deliberately withdrawn. For example, daf-15's
"regulation of autophagosome assembly" is no longer in GOA. A real regression is one where the
annotations are still present but the sentences are gone (ced-5, che-2: the current text has
expression and orthology only).

ced-5 (WB:WBGene00000419)

che-2 (WB:WBGene00000484)

daf-15 (WB:WBGene00000911)

dpm-1 (WB:WBGene00022044)

J7SF87 (WB:WBGene00021183)

nsy-1 (WB:WBGene00003822)

nud-1 (WB:WBGene00003829)

ppgn-1 (WB:WBGene00021425)

wago-1 (WB:WBGene00011061)

Note: this is a divergence rather than a pure loss. The current text adds RNA endonuclease
and miRNA binding (which the wago-1 review removes or modifies) and drops DEAD/H-box RNA
helicase binding.

Yeast candidates

Found during the yeast follow-up. Fewer yeast genes have a RefSeq copy to compare
against, so most of these are omissions, not before/after losses.

Fly candidates

Found during the fly follow-up (texts fetched 2026-11-01, quoted from
genes/DROME/<gene>/<gene>-descriptions.yaml).

Ortholog names

Two automated texts, and their RefSeq copies, name a human ortholog that the gene
reviews disagree with. The ortholog sentence comes from the Alliance orthology data,
not from GO, so check the orthology call itself before filing: