Add sourced InterPro annotation datasets

README.md

@@ -0,0 +1,63 @@
+# interpro_repeat
+
+Protein-level multi-label dataset sourced from InterPro entry metadata and
+UniProtKB protein-to-InterPro cross-references.
+
+## Intended use
+
+Protein repeat annotation prediction from sequence. This evaluates
+recovery of curated InterPro classifications, not an experimental assay result.
+
+## Source and labels
+
+- InterPro metadata: `https://ftp.ebi.ac.uk/pub/databases/interpro/current_release/entry.list`
+- UniProt REST API: `https://rest.uniprot.org/uniprotkb/stream`
+- Card generated (UTC): `2026-10-05`
+- Organism Taxonomy ID: `9606` (`all` = all organisms).
+- Review status: `reviewed` (`reviewed` = Swiss-Prot, `unreviewed` = TrEMBL, `all` = both).
+- UniProt query: `(organism_id:9606) AND (reviewed:true)`
+- UniProt TSV fields: `accession,sequence,xref_interpro,organism_id,reviewed`
+- InterPro entry type: `Repeat`
+- Consider entries observed in at least 5 proteins.
+- After splitting, retain entries with at least 20
+    positives in training, 3 in validation,
+    and 3 in test.
+- `targets` is a multi-hot vector ordered by InterPro accession. Accession,
+  display name, and type are recorded in `label_vocabulary.json`.
+- Proteins without a retained entry before splitting are dropped; after
+    support filtering, other split variants may contain all-zero targets.
+
+## Splits
+
+Whole MMseqs2 `easy-linclust` clusters are assigned to splits targeting
+`{'train': 0.8, 'validation': 0.1, 'test': 0.1}`, with minimum identity
+`0.3`, minimum coverage
+`0.8`, and
+`1` thread(s). Seed: `1957723`.
+When `create_split_subsets` is enabled, pooled random, stratified, and
+hold-cluster-out subsets are also included. Hold-cluster-out subsets require
+enough MMseqs clusters to populate all three roles.
+The optional maximum sequence length is `None` and is
+applied before vocabulary construction and MMseqs2 clustering.
+
+## Split sizes
+
+- `test`: 156 rows
+- `test_cluster`: 156 rows
+- `test_random`: 156 rows
+- `test_stratified`: 156 rows
+- `train`: 1251 rows
+- `train_cluster`: 1251 rows
+- `train_random`: 1251 rows
+- `train_stratified`: 1251 rows
+- `validation`: 156 rows
+- `validation_cluster`: 156 rows
+- `validation_random`: 156 rows
+- `validation_stratified`: 156 rows
+
+Vocabulary size: 9.
+
+## Dataset statistics
+
+[`stats.json`](stats.json) at the dataset root contains row counts by split,
+columns, and SeqKit sequence-length metrics.

data/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0734d1a4bd8b579dcd4b9ce1b2f43c1402b39844ea3e3837fa133b9fd5da51fd
+size 90288

data/test_cluster-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f60f60a02d35610089812fd7ae811900a48af971958324157bc07424172f9fdd
+size 74409

data/test_random-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:688d0b39385dbf898732d4e578fb4ae5bcd5a5692b04ea0a87a0144376ad2440
+size 83226

data/test_stratified-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df7a941e67d7d28da1d3e8c89e2acf5e7b06c7a7901ab69a320331c22138bb70
+size 87774

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99db388eb927491831318dd2ae085ee09da4ee409a8804c7e1fe913db460fbc1
+size 563934

data/train_cluster-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dace85517df7ab5bef5411d8e71c73cdea33508d12455fa745437248412c3732
+size 589808

data/train_random-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aab36e0bddab34c4a94ae3fd07168c9ef937d161f879a0e2b18ea34d4505b33e
+size 593232

data/train_stratified-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21427c96287cee161d617da04fb09d583c5c74dc6baf95112f05fb8ced02dd27
+size 593920

data/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ee8c2aa625b38bc0b0b5b0e88ebf61b57f426a79a42bbe6b6ae79ca970a9c67
+size 79410

data/validation_cluster-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1927f58461fdf334b5d6903c4203376417fa1d556e0e8e9d1629922d4364f37d
+size 66700

data/validation_random-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:411da86f2b3eaf524143845f0252f97e979832c8cafbd55d69a70f6b0c17557b
+size 89303

data/validation_stratified-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f39ea77963aa2e0500e5d2eb8785548ecae2c24fc4b2011db79a8d2941cd5576
+size 84197

label_vocabulary.json

@@ -0,0 +1,51 @@
+{
+  "dataset_name": "interpro_repeat",
+  "vocabulary_version": 1,
+  "labels": [
+    {
+      "accession": "IPR000225",
+      "name": "Armadillo",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR000884",
+      "name": "Thrombospondin type-1 (TSP1) repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR001611",
+      "name": "Leucine-rich repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR001680",
+      "name": "WD40 repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR002110",
+      "name": "Ankyrin repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR003591",
+      "name": "Leucine-rich repeat, typical subtype",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR008160",
+      "name": "Collagen triple helix repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR019734",
+      "name": "Tetratricopeptide repeat",
+      "entry_type": "Repeat"
+    },
+    {
+      "accession": "IPR020472",
+      "name": "PAC1/LIS1-like, WD-40 repeat",
+      "entry_type": "Repeat"
+    }
+  ]
+}

stats.json

@@ -0,0 +1,207 @@
+{
+  "rows_total": 6252,
+  "rows_by_split": {
+    "train": 1251,
+    "validation": 156,
+    "test": 156,
+    "train_random": 1251,
+    "validation_random": 156,
+    "test_random": 156,
+    "train_stratified": 1251,
+    "validation_stratified": 156,
+    "test_stratified": 156,
+    "train_cluster": 1251,
+    "validation_cluster": 156,
+    "test_cluster": 156
+  },
+  "columns": [
+    "id",
+    "sequence",
+    "targets",
+    "split"
+  ],
+  "seqkit_stats": [
+    {
+      "file": "test.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 161592,
+      "min_len": 138,
+      "avg_len": 1035.8,
+      "max_len": 4377,
+      "Q1": 468.5,
+      "Q2": 854.5,
+      "Q3": 1349.5,
+      "sum_gap": 0,
+      "sum_n": 6523
+    },
+    {
+      "file": "test_cluster.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 133420,
+      "min_len": 114,
+      "avg_len": 855.3,
+      "max_len": 5635,
+      "Q1": 380.0,
+      "Q2": 641.5,
+      "Q3": 1002.0,
+      "sum_gap": 0,
+      "sum_n": 5316
+    },
+    {
+      "file": "test_random.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 139996,
+      "min_len": 154,
+      "avg_len": 897.4,
+      "max_len": 6306,
+      "Q1": 422.0,
+      "Q2": 699.0,
+      "Q3": 992.0,
+      "sum_gap": 0,
+      "sum_n": 5602
+    },
+    {
+      "file": "test_stratified.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 145841,
+      "min_len": 114,
+      "avg_len": 934.9,
+      "max_len": 6885,
+      "Q1": 425.0,
+      "Q2": 638.0,
+      "Q3": 1057.0,
+      "sum_gap": 0,
+      "sum_n": 5696
+    },
+    {
+      "file": "train.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 1251,
+      "sum_len": 1092684,
+      "min_len": 79,
+      "avg_len": 873.4,
+      "max_len": 8797,
+      "Q1": 444.5,
+      "Q2": 661.0,
+      "Q3": 1031.0,
+      "sum_gap": 0,
+      "sum_n": 42863
+    },
+    {
+      "file": "train_cluster.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 1251,
+      "sum_len": 1137714,
+      "min_len": 79,
+      "avg_len": 909.4,
+      "max_len": 8797,
+      "Q1": 463.0,
+      "Q2": 686.0,
+      "Q3": 1050.5,
+      "sum_gap": 0,
+      "sum_n": 44368
+    },
+    {
+      "file": "train_random.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 1251,
+      "sum_len": 1104675,
+      "min_len": 79,
+      "avg_len": 883.0,
+      "max_len": 8797,
+      "Q1": 446.5,
+      "Q2": 669.0,
+      "Q3": 1045.5,
+      "sum_gap": 0,
+      "sum_n": 43068
+    },
+    {
+      "file": "train_stratified.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 1251,
+      "sum_len": 1104009,
+      "min_len": 79,
+      "avg_len": 882.5,
+      "max_len": 8797,
+      "Q1": 453.0,
+      "Q2": 673.0,
+      "Q3": 1045.0,
+      "sum_gap": 0,
+      "sum_n": 43026
+    },
+    {
+      "file": "validation.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 139158,
+      "min_len": 123,
+      "avg_len": 892.0,
+      "max_len": 7388,
+      "Q1": 449.0,
+      "Q2": 667.5,
+      "Q3": 1006.0,
+      "sum_gap": 0,
+      "sum_n": 5050
+    },
+    {
+      "file": "validation_cluster.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 122300,
+      "min_len": 138,
+      "avg_len": 784.0,
+      "max_len": 2694,
+      "Q1": 345.5,
+      "Q2": 601.0,
+      "Q3": 921.5,
+      "sum_gap": 0,
+      "sum_n": 4752
+    },
+    {
+      "file": "validation_random.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 148763,
+      "min_len": 165,
+      "avg_len": 953.6,
+      "max_len": 7388,
+      "Q1": 476.5,
+      "Q2": 668.0,
+      "Q3": 1050.5,
+      "sum_gap": 0,
+      "sum_n": 5766
+    },
+    {
+      "file": "validation_stratified.fasta.gz",
+      "type": "Protein",
+      "num_seqs": 156,
+      "sum_len": 143584,
+      "min_len": 86,
+      "avg_len": 920.4,
+      "max_len": 7570,
+      "Q1": 430.0,
+      "Q2": 726.5,
+      "Q3": 1028.0,
+      "sum_gap": 0,
+      "sum_n": 5714
+    },
+    {
+      "file": "overall",
+      "type": "Protein",
+      "num_seqs": 6252,
+      "sum_len": 5573736,
+      "min_len": 79,
+      "avg_len": 891.5,
+      "max_len": 8797,
+      "Q1": 447.0,
+      "Q2": 672.0,
+      "Q3": 1046.0,
+      "sum_gap": 0,
+      "sum_n": 217744
+    }
+  ]
+}