Hyperprolinemia Type 2 (ALDH4A1, MONDO:0009401) — Claude Code literature sweep

Date: 2026-07-31 Tool: Claude Code (curation scanner, high_effort tier) Target entry: kb/disorders/Hyperprolinemia_Type_2.yaml Context: review follow-up on PR #7338 (issue #5558, IEMbase WP-003 row 1.7.06.01)

This artifact records a systematic literature sweep run after the initial entry was drafted, in response to review finding #1 on PR #7338 ("no deep-research artifact ships with this new entry, and findings 2–4 are all things a literature sweep would have surfaced"). It documents what was searched, what was found, and — importantly — which candidate papers were inspected and deliberately not used.

Method

PubMed E-utilities esearch, relevance-sorted, retmax=40, three complementary queries chosen to catch gene-anchored, disease-name-anchored, and enzyme-anchored literature respectively:

Titles for every PMID appearing in the union that was not already cited by the entry were retrieved via esummary and triaged. Every paper selected for use was then fetched with just fetch-reference and each snippet verified as an exact substring of the cached abstract via just validate-references.

Baseline check

GeneReviews: none exists. Confirmed independently in the PR review — both hyperprolinemia[TI] AND GeneReviews[Book] and (hyperprolinemia OR ALDH4A1 OR "pyrroline-5-carboxylate dehydrogenase") AND GeneReviews[Book] return zero results against a working control. No GeneReviews baseline requirement applies to this entry.

Orphanet: ORPHA:79101 cached via just structured-rebuild-orphanet --id 79101. The record carries a definition and cross-references but no HPO phenotype table, so it cannot supply frequency annotations for this disorder. Its definition sentence is used as an evidence snippet for the intellectual-deficit / developmental-delay phenotypes.

Papers newly incorporated

PMID:24173411 — van de Ven 2014, J Inherit Metab Dis

The largest HPII series in the literature: 4 metabolically confirmed patients ascertained from 20,991 urinary organic acid profiles. Supplies four things the entry lacked:

1. Counter-evidence to the pyridoxine claim — "The clinical course was non-progressive and independent from the B6 concentration and B6 therapy." Curated as supports: REFUTE on the treatment, against the single supporting case report. The treatment is now presented as genuinely conflicting rather than established.
2. Intellectual disability — 2 of 4 patients.
3. Behavioral phenotype — 4 of 4 with "significant behavioral problems, including anxiety and hallucinations."
4. The mitochondrial arm — 3 of 4 with biochemical markers of mitochondrial dysfunction, biopsy-confirmed in one.

PMID:34302426 — Namavar 2021, Am J Med Genet B

PRISMA systematic review, 1753 studies screened, 35 included. Reports a common psychiatric phenotype (developmental delay, intellectual disability, ASD, psychosis spectrum) and — critically — no biochemical–clinical phenotype correlation. Used both as phenotype support and as the reason the neurodevelopmental pathophysiology node is marked HYPOTHETICAL.

Scope caveat, applied throughout: this review pools hyperprolinemia types I and II. Every citation of it in the entry is therefore supports: PARTIAL, and the HPII-specific anchors (PMID:24173411, ORPHA:79101, PMID:41602883) carry the SUPPORT weight.

PMID:21168532 — He & DiMario 2011, Mitochondrion

Not flagged in the PR review; surfaced by the ALDH4A1 and enzyme-anchored queries. A Drosophila P5CDh-null model ("establishing a fly model for human type II hyperprolinemia") showing doubled proline together with swollen mitochondria and larval/pupal lethality. This is independent, cross-species corroboration of the mitochondrial arm from a completely different system, which is what justifies modeling that arm at all rather than treating the human n=4 finding as noise. Tagged MODEL_ORGANISM. The authors' own "first correlation between the loss of P5CDh and morphological defects in mitochondria" framing is quoted to keep the node PROVISIONAL.

PMID:41602883 — AlQurashi 2025, Front Pediatr

Also not flagged in the review; the most recent HPII case report. A consanguineous Saudi child presenting with global developmental delay and clinically diagnosed ASD, with a homozygous ALDH4A1 variant. Supplies the only HPII-specific primary citation for developmental delay (previously resting on the Orphanet definition alone).

Caveat recorded in the entry: the variant is of uncertain significance, not an established pathogenic allele. The ASD phenotype is therefore curated as two PARTIAL items with no frequency band, and the developmental-delay note states the VUS status explicitly.

Candidates inspected and NOT used

Recording these so a future curator does not re-triage them:

Named Entity Confusion (NEC) preflight

Re-verified for this sweep, since "hyperprolinemia type II" is a numbered-series label of exactly the kind flagged in research/nec_risk_disease_classes.md:

uv run runoak -i sqlite:obo:mondo info MONDO:0009401 -O obo
  relationship: RO:0004003 HGNC:406 ! ALDH4A1
  xref: OMIM:239510
  xref: Orphanet:79101

Gene, OMIM, and ORPHA all match WP-003 row 1.7.06.01. Every paper incorporated above was checked to be about ALDH4A1/HPII specifically, with the HPI-pooling caveat on PMID:34302426 handled by downgrading it to PARTIAL. PASS.

Residual gaps after this sweep

Both are tracked as KNOWLEDGE_GAP discussions in the entry rather than left implicit:

• gap_hpii_plp_seizure_causality — the PLP-depletion-causes-seizures model is now a matter of conflicting evidence, not absent evidence: one case reports B6-responsive seizures, the only cohort reports a course independent of B6 therapy. No controlled trial exists and none is likely at this prevalence.
• gap_hpii_penetrance_and_phenotype — no biochemical–clinical correlation has been demonstrated, so neither penetrance nor severity is predicted by proline level, and the modifiers remain unidentified.

A third gap is noted here but not curated as a discussion because it is a literature-coverage limitation rather than a mechanistic one: every clinical statement about HPII rests on a total of roughly a dozen patients worldwide. No frequency bands are curated anywhere in the entry for this reason.