mikessh Claude Opus 4.8 commited on
Commit
89a5bee
·
1 Parent(s): 1ebb8a6

Rework #9 into workhorse neoag DB (tested + candidates) + HLA population freqs

Browse files

- immunogenicity/neoag_tested.tsv.gz: 526k tested patient-derived neoantigens
(immunogenicity 0/1) from TESLA, Neopep, CEDAR, Gfeller, ITSNdb, GBM, VACCIMEL,
Bjerregaard; adds patient_id, cancer_type, validated, affinity_wt
- immunogenicity/neoag_candidates.tsv.gz: 1.36M untested candidates (Neopep not_tested)
- immunogenicity/hla_pop_freqs.tsv.gz: EUR/ASN/AFR class-I allele frequencies (AFND)
- removed benchmarks.tsv.gz (superseded by the split)
- DESCRIPTION.md: brief per-dataset index, now tracked

Pipeline, dashboard, methods & roadmap moved to private repseq/gamaleya-cancer.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

.gitignore CHANGED
@@ -11,8 +11,5 @@ raw/
11
  gamaleya/
12
  tcga/
13
 
14
- # local-only deliverables (kept local for now)
15
- DESCRIPTION.md
16
- ROADMAP.md
17
- dashboard/
18
  summary_all_datasets.tsv
 
11
  gamaleya/
12
  tcga/
13
 
14
+ # internal spec (not published); docs/dashboard live in repseq/gamaleya-cancer
 
 
 
15
  summary_all_datasets.tsv
DESCRIPTION.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # pMHC Data Compendium
2
+
3
+ Harmonized data substrate for a personalized cancer-vaccine immunogenicity pipeline:
4
+ epitope–MHC binding, immunogenicity labels & a tested-neoantigen workhorse DB, self/foreign
5
+ similarity references, and HLA population frequencies. Public data is here; per-patient data
6
+ stays private (gitignored). Pipeline, methods, roadmap & dashboard live in the private
7
+ infrastructure repo (`repseq/gamaleya-cancer`).
8
+
9
+ All tables are gzip-TSV. The immunogenicity tables share one schema: `peptide, mhc_a, mhc_b,
10
+ mhc_class, mhc_species, source_species, host_species, assay_type, immunogenicity{1,0,null},
11
+ expression, affinity, affinity_wt, agretopicity, foreignness, validated, dataset_origin,
12
+ reference_id, n_references` (+ `patient_id, cancer_type` for the neoag tables). `affinity` is
13
+ as-reported per source (units differ); MHC alleles are as-reported.
14
+
15
+ ## Public datasets (this repo)
16
+
17
+ | Path | # | Dataset | One-line |
18
+ |------|---|---------|----------|
19
+ | `pmhc/pmhc_full.tsv.gz` | 6 | IEDB epitope–MHC | positive IEDB MHC-ligand binding records (human IMGT/HLA + mouse H-2) |
20
+ | `pmhc/pmhc_shortlist.tsv.gz` | 6 | IEDB epitope–MHC (≥2 refs) | high-confidence pairs (≥2 references) |
21
+ | `immunogenicity/iedb_labeled.tsv.gz` | 7 | IEDB immunogenicity | immunogenic (T-cell+) vs non-immunogenic (MHC-ligand+, healthy self) |
22
+ | `immunogenicity/neoag_tested.tsv.gz` | 9 | Neoantigens — tested (workhorse) | tested patient-derived neoantigens, immunogenicity 0/1 (TESLA, Neopep, CEDAR, Gfeller, ITSNdb, GBM, VACCIMEL, Bjerregaard) |
23
+ | `immunogenicity/neoag_candidates.tsv.gz` | 9 | Neoantigens — candidates | untested neoantigen candidates (Neopep `not_tested`) |
24
+ | `immunogenicity/hla_pop_freqs.tsv.gz` | — | HLA population freqs | EUR/ASN/AFR class-I (A/B/C) allele frequencies (AFND) |
25
+ | `proteome/human.fasta.gz` | 4 | Human proteome (UP000005640) | self-reference proteome for similarity search |
26
+ | `proteome/mouse.fasta.gz` | 5 | Mouse proteome (UP000000589) | C57BL/6 self-reference proteome |
27
+ | `ligandome/viral_foreign_iedb.tsv.gz` | 2 | Viral ligandome (IEDB) | viral-source presented peptides (foreign reference) |
28
+ | `ligandome/viral_orfs_gse272406.tsv.gz` | 1 | Pan-viral ORFs (GSE272406) | translated novel viral ORF proteins (foreign; *not* thymus self — spec mislabel) |
29
+ | `ligandome/cancer_targets_tsarina.tsv.gz` | 3 | Cancer-testis antigens | curated shared tumor-antigen genes (tsarina) |
30
+ | `summary.tsv` | — | Pipeline stats | row counts & extraction metrics |
31
+
32
+ ## Private (gitignored — NOT in this repo)
33
+
34
+ | Path | # | Dataset |
35
+ |------|---|---------|
36
+ | `gamaleya/patient_features.tsv.gz`, `gamaleya/hla_coverage.tsv.gz` | 8 | Gamaleya patient candidates + features (17 patients) |
37
+ | `tcga/tcga_neoantigens.tsv.gz`, `tcga/hla_coverage.tsv.gz` | 10 | TCGA expressed neoantigen binders + tumor type (controlled-access) |
38
+
39
+ Detailed schemas, methods, and caveats: see the private `gamaleya-cancer` repo
40
+ (`DESCRIPTION` here is intentionally brief).
immunogenicity/{benchmarks.tsv.gz → hla_pop_freqs.tsv.gz} RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:921b1530dae81bd0c0169271a1e5c3e39c8652ee9583ff82520d97251d1027d3
3
- size 6226745
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0beccfd30d62e5934f6335c971416ce165c84301e8e262a0c418d7cc9f8ba5de
3
+ size 56665
immunogenicity/neoag_candidates.tsv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d0b0e00c10c56c27a79722507e0e21ae35a8bce8c5181034cabadc9ce8a547b
3
+ size 17532582
immunogenicity/neoag_tested.tsv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:118658608f7046d5f007a7fb76ef3716eb4467a73474b8b32ffb7e87f44d6167
3
+ size 7119906