expression: GTEx has no thymus, so add the HPA consensus arm

Two gaps found while building the mouse side, both on the human side.

GTEx's 53 contexts contain no thymus, no lymph node and no bone marrow -- adult thymus
involutes and is not routinely sampled -- so the normal-tissue safety read was blind to
the tissue central tolerance happens in. Expression Atlas does not rescue it either: of
123 human baseline experiments only three carry a thymus group, and all three are
compromised (fetal tissue from fetuses with congenital defects, sorted thymocyte subsets,
and thymic epithelium cultured on a feeder layer).

And thymus/thymus_expression.tsv.gz was not quantitative: 10,367 of its 10,493 rows are
UniProt accessions of eluted-peptide source proteins with an EMPTY value, which records
provenance, not expression. Only 126 rows held anything, and that was a text label. The
mouse table had numbers and the human one did not.

HPA samples thymus directly, so its consensus set fixes both:

expression/reference_expression.tsv.gz 5,656,063 -> 6,681,814 rows, source=hpa_consensus
1,025,751 rows / 20,151 genes / 51 tissues
thymus/thymus_expression.tsv.gz 10,493 -> 26,772 rows, +16,279 measured nTPM

Added as a third `source`, not merged into GTEx and not replacing it; every gtex and tcga
row is untouched. The schema was designed with a `source` column for exactly this.

Checked, not assumed: of 51 tissues thymus ranks first for PRSS16 (770 nTPM), PSMB11,
DNTT, CD1A, RAG1, CCL25, CD3E and LCK, while ALB peaks in liver at 198,524 against 0.3 in
thymus.

Recorded in SOURCES.md: nTPM is not GTEx TPM, so compare within a source, not across; n=1
so q25/q75 carry no spread; and AIRE is detectable here (1.2 nTPM) but reads 0 in the
mouse CAGE atlas, which is assay depth and not a species difference.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

DESCRIPTION.md

@@ -110,6 +110,7 @@ Every tracked directory now carries one:
 | `ligandome/viral_orfs_gse272406.tsv.gz` | Pan-viral ORFs (GSE272406) | translated novel viral ORF proteins (foreign; *not* thymus self — spec mislabel) |
 | `ligandome/cancer_targets_tsarina.tsv.gz` | Cancer-testis antigens | curated shared tumor-antigen genes (tsarina) |
 | `ligandome/tissue_self_mmu.tsv.gz` | Tissue self-ligandome (mouse) | 46,334 rows / 17,256 distinct peptides across **18 non-thymic tissues**, Schuster murine MHC-I atlas (PXD008733) at a recomputed 1% peptide FDR. H-2Kb 8,829 + H-2Db 7,781 peptides (MHC-I) and 4,348 I-Ab (MHC-II); carries a `tissue` column. Tumour cell lines (EL4, B16F10, GL261, LLC) excluded — they are not normal tissue |
+| `expression/reference_expression.tsv.gz` | Reference expression (human) | 6,681,814 rows over three sources: `gtex` (3,955,284 / 53 tissues), **`hpa_consensus` (1,025,751 / 51 tissues — the only one containing thymus, added 2026-08-21)** and `tcga` (1,700,779 / 19 tumour types). GTEx has no thymus, lymph node or bone marrow |
 | `expression/reference_expression_mmu.tsv.gz` | Reference expression (mouse) | 659,050 rows / 18,830 genes across **35 adult tissues**, `source=fantom5_mouse`. Column for column identical to the human file — the mouse GTEx analogue, the normal-tissue safety read for a mouse epitope. FANTOM5 CAGE (E-MTAB-3579). `n`=1 per tissue: the IQR is across transcripts, not animals |
 | `expression/protein_abundance_mmu.tsv.gz` | Protein abundance (mouse) | 133,848 rows / 5,148 genes across 26 tissues, `source=geiger_silac_mouse`, `unit=ppb_ibaq` (E-PROT-11). Kept apart from the RNA table on purpose — a SILAC abundance is not a TPM |
 | `thymus/thymus_expression_mmu.tsv.gz` | Thymus expression (mouse) | 15,306 rows / 12,646 genes: adult-thymus CAGE TPM (11,862) + thymus SILAC protein (3,444). `value` carries the measurement so callers can threshold. `Aire` reads 0 — a detection floor, not biology |

expression/SOURCES.md

@@ -6,15 +6,18 @@ gene-keyed table and a peptide-keyed one are not the same object.
 | | |
 |---|---|
 | file | `reference_expression.tsv.gz` |
-| rows | 5,656,063 |
+| rows | 6,681,814 |
 | schema | `key`, `key_type`, `source`, `context`, `median_tpm`, `q25_tpm`, `q75_tpm`, `n` |
 | provenance | **experimental** (both sources are measured RNA-seq); the per-context median/IQR/`n` are **computed** summaries of it |
 
 | `source` | `key_type` | rows | contexts | what a row is |
 |---|---|--:|--:|---|
 | `gtex` | `gene` | 3,955,284 | 53 | a gene's expression across normal tissue — the **safety** read |
+| `hpa_consensus` | `gene` | 1,025,751 | 51 | the same read over HPA's tissue panel — **the only one that contains thymus** |
 | `tcga` | `peptide` | 1,700,779 | 19 | a mutant peptide's expression by tumour type — the **target** read |
 
+Total 6,681,814 rows.
+
 `key_type` is the join key and it differs by source on purpose: GTEx answers "is this gene on in
 normal tissue", TCGA answers "is this neoantigen transcribed in this tumour". A caller that treats
 them as one column is asking a question neither answers.
@@ -119,3 +122,57 @@ the only two of those with broad tissue coverage.
 ### Regenerate
 
     cd ~/vcs/projects/2026-mouse-thymus && ./src/build_expression_mmu.py
+
+---
+
+## Human — the `hpa_consensus` arm, and why GTEx alone was not enough
+
+Added 2026-08-21, in the same pass that built the mouse references.
+
+**GTEx has no thymus.** Its 53 contexts contain no thymus, no lymph node and no bone marrow — adult
+thymus involutes and is not routinely sampled, so the normal-tissue safety read was blind to the one
+tissue central tolerance actually happens in. Expression Atlas is no better: of its **123 human
+baseline experiments only three carry any thymus group**, and all three are compromised for this
+purpose — E-MTAB-3871 is fetal tissue from fetuses with congenital defects, E-MTAB-3827 is sorted
+thymocyte subsets rather than tissue, and E-MTAB-9641 is thymic epithelium cultured on a 3T3-J2
+feeder layer.
+
+HPA samples thymus directly, so its consensus tissue set is what fills the hole.
+
+| | |
+|---|---|
+| rows | **1,025,751** |
+| keys | 20,151 gene symbols (`key_type = gene`) |
+| contexts | **51 tissues, including thymus** |
+| `source` | `hpa_consensus` |
+| provenance | **experimental** — RNA-seq; the HPA consensus normalisation is **computed** |
+
+**Origin.** [`rna_tissue_consensus.tsv.zip`](https://www.proteinatlas.org/download/tsv/rna_tissue_consensus.tsv.zip)
+from the Human Protein Atlas.
+
+> Uhlén M, Fagerberg L, Hallström BM, Lindskog C, Oksvold P, Mardinoglu A, Sivertsson Å, Kampf C,
+> Sjöstedt E, Asplund A, Olsson I, Edlund K, Lundberg E, Navani S, Szigyarto CA, Odeberg J,
+> Djureinovic D, Takanen JO, Hober S, Alm T, Edqvist PH, Berling H, Tegel H, Mulder J, Rockberg J,
+> Nilsson P, Schwenk JM, Hamsten M, von Feilitzen K, Forsberg M, Persson L, Johansson F, Zwahlen M,
+> von Heijne G, Nielsen J, Pontén F.
+> **Tissue-based map of the human proteome.**
+> *Science* 2015;347(6220):1260419.
+> PMID [25613900](https://pubmed.ncbi.nlm.nih.gov/25613900/) ·
+> doi:[10.1126/science.1260419](https://doi.org/10.1126/science.1260419)
+
+**It is added as a third `source`, not merged into GTEx and not replacing it.** The schema was
+designed with a `source` column for exactly this; every existing `gtex` and `tcga` row is untouched.
+
+**Checked against known biology.** Of 51 tissues, thymus ranks **first** for `PRSS16` (770 nTPM),
+`PSMB11` (28.5), `DNTT` (306), `CD1A` (296), `RAG1` (368), `CCL25` (718), `CD3E` (596) and `LCK`
+(326). Negative controls behave: `ALB` peaks in liver at 198,524 nTPM against 0.3 in thymus, `INS` in
+pancreas, `MYH7` in skeletal muscle.
+
+**Caveats.**
+- **nTPM is not GTEx TPM.** Both are TPM-family units but the normalisation differs, so compare
+  *within* a `source`, not across. This is what the `source` column is for.
+- `n` is 1 — HPA consensus publishes one value per gene × tissue, so `q25_tpm` and `q75_tpm` equal
+  `median_tpm` and carry no spread. GTEx `n` runs to several hundred donors; the two `n` columns are
+  not the same quantity.
+- **`AIRE` is detectable here (1.2 nTPM, rank 2 of 51) but reads 0 in the mouse CAGE atlas.** That is
+  a difference in assay depth, not in biology — do not read it as a species difference.

expression/reference_expression.tsv.gz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6274d089b63fa85d81f4c82acf3b63e4d63a1ca56683f76c5e3f2ba56ff20d46
-size 105486870
+oid sha256:b66301491e4fb0945eebfafa267dc287c8d4e9ffe2e81d0c8f351972f509605e
+size 114886559

thymus/SOURCES.md

@@ -135,12 +135,21 @@ Thymically expressed self, by gene. Schema: `gene`, `species`, `evidence`, `valu
 
 | `evidence` | rows | what it is |
 |---|--:|---|
+| `hpa_thymus_ntpm` | 16,279 | **HPA thymus nTPM > 0 — the measured value**, added 2026-08-21 |
 | `hla_ligand_atlas_thymus_source` | 10,367 | source proteins of the peptides above |
 | `hpa_thymus_cluster` | 126 | Human Protein Atlas thymus-cluster genes |
 
-The two lines of evidence are kept in a column rather than pooled: one is "a peptide from this
-protein was eluted from thymus", the other is "this gene is thymus-enriched by expression". They
-support each other; they are not interchangeable. Human only — there is no mouse counterpart yet.
+26,772 rows. The evidence lines are kept in a column rather than pooled, because they are three
+different statements: a gene is *transcribed* in thymus, a peptide from its protein was *eluted*
+from thymus, and it is thymus-*enriched* by cluster assignment. They support each other; they are
+not interchangeable.
+
+**Why the `hpa_thymus_ntpm` line was added.** Before it, this table was not quantitative: 10,367 of
+its 10,493 rows carried an **empty `value`** — they record which proteins peptides came from, which
+is not an expression measurement — and the only rows with content held a text cluster label. The
+mouse table had numbers and the human one did not. Note also that **GTEx has no thymus at all**, so
+the compendium's main safety read cannot answer thymic questions; `expression/SOURCES.md` records
+what was searched and why HPA is the source that works.
 
 ## Caveats on the mouse file
 

thymus/thymus_expression.tsv.gz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7ffca9961b75c8341ec0740f7af14c88b6064c5a72409c7b169d40ea7959857
-size 42774
+oid sha256:4de6b82d1f57859e4059249eae201246e1b227bab241b85e205b21f919cef10f
+size 114898