Somatic Mosaicism

A disorder whose causal variant arose after fertilization is encoded through its mode of inheritance, the way a digenic one is, so the set can be recovered by query. Before this convention the flagship entries said the same thing four ways (an HP term, a block bound to Sporadic, a free-text "Not applicable" block, prose only) and none of them could be found.

The convention

1. Add an inheritance block whose inheritance_term is bound to HP:0001442 Typified by somatic mosaicism, with its own snippet-backed evidence. The usual quote is the paired lesion-versus-blood sequencing sentence or a tissue-restricted allele-fraction sentence.
2. Set variant_origin: SOMATIC on the causal genetic: row, or GERMLINE_AND_SOMATIC for a two-hit repressor. DE_NOVO is for a germline de novo variant present in every cell.
3. Keep category in the standard vocabulary (Genetic). Never Mendelian for an obligate-mosaic disorder, and never an ad hoc Somatic mosaic; the mosaic claim lives in the bound term.

HP:0003745 Sporadic is not a substitute: it describes the pedigree, not the mechanism.

Worked example: VEXAS_Syndrome (acquired clonal); Sturge-Weber_Syndrome (congenital obligate mosaic); Tuberous_Sclerosis_Complex (mosaic arm beside an autosomal dominant block).

Two tiers, and what is not a tier

Situation	Encode	Grouping member?
The disease is the mosaic (no germline form exists)	one HP:0001442 block	yes
A germline disease has a recognized mosaic arm described in the literature	a second HP:0001442 block beside the germline block; do not replace it	yes
One mosaic proband in a case report	bind HP:0001442 on that block if curated; it is a true observation	no
Parental or gonadal mosaicism explaining recurrence	bind HP:0001442 if curated; the child's disease is constitutional	no
Functional mosaicism from X inactivation	do not bind HP:0001442; free-text preferred_term and say why (CHILD_Syndrome)	no
Mosaic aneuploidy as a downstream phenotype of a germline gene (mosaic variegated aneuploidy)	ordinary germline inheritance	no

The term names DNA-level post-zygotic mosaicism and is true in the first four rows. Only the first two describe the disease's architecture, which is what kb/groupings/Somatic_Mosaic_Disorders.yaml collects under a NECESSARY HAS_INHERITANCE criterion. It is deliberately not NECESSARY_AND_SUFFICIENT, unlike the digenic grouping, because rows three and four are legitimate bindings that are not memberships; the evaluator would otherwise report them as candidates on every run.

Not yet modelled

There is no structured slot for variant allele fraction, affected tissue, or clone timing. Put those values in notes: for now. The schema question is tracked in projects/COMMONFUND/SMAHT.md and docs/todo/inheritance-enrichment.md.

Validate

just validate kb/disorders/<Entry>.yaml
just count-verified-snippets kb/disorders/<Entry>.yaml
just check-groupings --strict kb/groupings/Somatic_Mosaic_Disorders.yaml