RFWD3 (FANCW) — AIGR vs Affinage
Affinage record: run 2026-06-10 · 17 discoveries · self-eval win (faith 100%) · gates passed.
Agreement (brief)
Affinage and the AIGR review converge on the core biology, and the AIGR review was
already comprehensive here:
- RFWD3 is a RING-type E3 ubiquitin ligase (EC 2.3.2.27, catalytic Cys315) — AIGR core
molecular functionGO:0061630. - Recruited via its C-terminal WD40 domain to RPA2 on ssDNA at stalled forks / damage
sites; polyubiquitinates RPA and RAD51 to drive their VCP/p97-dependent turnover,
enabling HR (GO:0000724) and ICL repair (GO:0036297, RFWD3 = FANCW). - ATR/ATM phosphorylation-dependent activity; promotes ssDNA-protein ubiquitination →
PCNA ubiquitination / TLS. - Nuclear / site-of-DNA-damage localization; secondary MDM2-p53 stabilization role,
which AIGR correctly demotes toKEEP_AS_NON_CORE.
All 8 of Affinage's shared PMIDs (20173098, 21504906, 21558276, 26474068, 28575657,
28575658, 28691929, 33321094) were already cited and adjudicated in the AIGR review.
Disagreements
| Topic | Affinage says | AIGR review says | Verdict (who is right + why) |
|---|---|---|---|
| GO grounding of MF | mechanism_profile gives only coarse parents: GO:0016874 ligase activity, GO:0140096 catalytic activity acting on a protein |
Specific GO:0061630 ubiquitin protein ligase activity |
AIGR right. Affinage's own note flags the profile as coarse; GO:0016874 (generic ligase, C–O/C–N/C–S bond formation) is arguably a different branch from thioester-forming Ub transfer. Do not import. |
| Localization grounding | GO:0005634 nucleus + GO:0000228 nuclear chromosome |
nucleus + GO:0090734 site of DNA damage / GO:0035861 etc. |
AIGR right/finer. GO:0000228 (nuclear chromosome, the condensed structure) is not where an RPA/ssDNA-recruited fork factor is best placed; AIGR's damage-site/fork terms are more accurate. |
| Pathway layer | Reactome DNA Repair, DNA Replication, Disease, Metabolism of proteins | process terms grounded per-annotation | AIGR right. "Metabolism of proteins" is a frequency-biased over-general parent (any ubiquitination event); uninformative for this gene. |
| p53 / MDM2 axis | folded into the mechanistic narrative at similar prominence | KEEP_AS_NON_CORE (single-group, secondary) |
AIGR right. Correct scoping; the axis is real (PMID:20173098) but peripheral to the replication-stress core. |
| TREX1 / STING immune role (PMID:41117130) | listed as a 2025 finding | absent | AIGR defensibly conservative. Single 2025 paper, distinct pathway; not yet warranting a curated annotation. Noted, not imported. |
| ORC/ORCA stabilization (PMID:33044890) | listed (self-rated Low confidence) | absent | AIGR right to omit. Affinage itself rates it Low; p53-context-dependent, single group. |
| Drosophila Mus302 (PMID:31900333) | listed | absent | Correctly excluded — ortholog/evolution context, not human RFWD3 function; HR role is mammal-specific per that paper. |
No factual conflicts requiring an AIGR reversal; the disagreements are the expected
coarse/over-general GO layer plus a few recent single-group over-reaches AIGR rightly excludes.
Papers incorporated into the review
| PMID | Supports | How used |
|---|---|---|
| 30530694 | RFWD3 at unperturbed forks; PCNA/PIP binding stabilizes RFWD3; normal fork progression / DNA replication | Added to references (relevance MEDIUM, VERIFIED) + verbatim supported_by on the GO:0031297 replication fork processing annotation |
| 37036693 | RFWD3 recruits ZRANB3 and stimulates fork remodeling/reversal (ZRANB3-epistatic) via PCNA ubiquitination | Added to references (relevance MEDIUM, VERIFIED) + verbatim supported_by on the same GO:0031297 annotation |
Both are full-text, RFWD3-focused functional studies that were missing from the review and
that corroborate/extend the (already accepted) replication-fork-processing role. No existing
decisions were weakened. Other Affinage-only PMIDs (32391871, 35905994, 40940676, 41117130,
41372167, 33044890, 31900333) were reviewed and not incorporated: they are corroborating
context, single-group recent findings, or non-human, none altering a curation call.
Net assessment
The AIGR review is stronger than the Affinage record on GO grounding, scoping, and evidence
adjudication — Affinage's mechanism_profile collapses to over-general (sometimes wrong-branch)
parents and should not be imported. Affinage's value here is its dense, dated PMID narrative,
which surfaced two legitimate RFWD3-focused mechanistic papers (PCNA/fork progression; ZRANB3/fork
remodeling) absent from the review. These were incorporated conservatively as corroborating
evidence on the existing fork-processing annotation. 0 new annotations, 2 papers incorporated,
review remains ✓ Valid.