[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6QHS0V8HZCD93MDA77KRV21":3,"skill-same:01M6QHS0V8HZCD93MDA77KRV21":80},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":38,"linkOnly":39,"markdownContent":40,"markdownBody":41,"frontMatter":42,"refName":51,"commitSha":52,"skillDirectory":53,"forksCount":14,"repositoryHtmlUrl":54,"repositoryHomepageUrl":55,"ownerHtmlUrl":56,"ownerBlogUrl":57,"ownerDescription":58,"updatedAt":59,"install":60,"resources":73},"01M6QHS0V8HZCD93MDA77KRV21","tiledbvcf",null,"Stores and retrieves genomic variant calls with TileDB-VCF. Use for indexed single-sample VCF\u002FBCF ingestion, incremental cohorts, region and sample queries, streaming results, allele statistics, QC, and VCF\u002FBCF export locally or through TileDB Cloud.","Tashima-Tarsh\u002FDisha6.6","Tashima-Tarsh","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F269906498?v=4","01M6PBH98X03XERH0HFT4PFB1P","NITISH KUMAR","01M6PBH9DA0M6KM3Z12P848F0K",0,"MIT","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkd-tiledbvcf\u002FSKILL.md",[],[5,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37],"for","storage","remote","native","scientific","bcftools","vcf","qc","input","conda","statistics","token","sample","ingestion","export","access","network","streaming","permissions","2026-11-10 10:12:30",false,"---\nname: tiledbvcf\ndescription: Stores and retrieves genomic variant calls with TileDB-VCF. Use for indexed single-sample VCF\u002FBCF ingestion, incremental cohorts, region and sample queries, streaming results, allele statistics, QC, and VCF\u002FBCF export locally or through TileDB Cloud.\nlicense: MIT license\ncompatibility: Requires a native TileDB-VCF installation and Python 3.9-3.12 for the reviewed Conda builds. bcftools compresses\u002Findexes input fixtures. Installation and remote storage need network access; TileDB Cloud needs its Python client, account token, and storage permissions.\nmetadata:\n  version: \"1.3\"\n  skill-author: Jeremy Leipzig\n  last-reviewed: \"2026-10-01\"\n  upstream-version: \"TileDB-VCF 0.40.3; tiledb-cloud 0.14.4\"\ncategory: \"scientific-computing\"\n---\n\n# TileDB-VCF\n\n## When to use\n\nUse for cohort variant storage, incremental sample ingestion, interval queries, and\nexporting subsets for downstream genomics. TileDB-VCF stores records; it does not\nperform joint variant calling, association testing, normalization, or population\nstructure adjustment. Dataset size alone does not determine whether cloud execution\nis appropriate: benchmark the intended sample, region, and attribute workload.\n\nThis skill targets released **TileDB-VCF 0.40.3**. Native Python\u002FCLI checks used two\nsynthetic single-sample VCFs on macOS ARM64. Cloud client **0.14.4** contracts were\nchecked against its released wheel and mocked dispatch; hosted queries and jobs\nwere not executed. See [installation and verification notes](references\u002Fverification.md).\n\n## Install and verify the native stack\n\nThe `tiledbvcf` Python distribution is not published on PyPI at this review. The\nofficial **`tiledb` Conda channel** supplies `tiledbvcf-py` 0.40.3 for macOS ARM64,\nmacOS x86-64, and Linux x86-64, with Python 3.9-3.12 builds. `tiledb` on PyPI is\nTileDB-Py and does not install TileDB-VCF. Native Apple Silicon no longer requires\nforcing `CONDA_SUBDIR=osx-64`.\n\nFor the tested macOS ARM64 environment, pins below avoid versioned-library import\nfailures in the otherwise successful current Conda solve:\n\n```bash\nconda create -n tiledb-vcf -c tiledb -c conda-forge \\\n  python=3.12 tiledbvcf-py=0.40.3 \\\n  azure-core-cpp=1.16.2 azure-storage-blobs-cpp=12.16.0 \\\n  azure-storage-files-datalake-cpp=12.14.0 capnproto=1.4.0 c-blosc2=2.23.1\nconda activate tiledb-vcf\npython -c 'import tiledbvcf; print(tiledbvcf.version)'\ntiledbvcf version\n```\n\nThe equivalent Micromamba solve\u002Finstall was executed. These extra ABI pins are a\nverified macOS ARM64 workaround, not a claim about all platforms. Preserve the\nresolved environment for production. Other-platform installation and official\nDocker images are alternatives documented upstream, not tested here.\n\n## Prepare and create a cohort\n\n1. Record reference assembly, contig naming\u002Flengths, callers, normalization rules,\n   sample identity, and file checksums. Do not mix assemblies or assume `1` and\n   `chr1` are equivalent. Confirm compatible headers across inputs.\n2. Each input must contain **one sample**, be coordinate sorted, and have an index.\n   Use BGZF-compressed VCF or BCF with a matching `.csi`\u002F`.tbi`. Inspect\n   `bcftools query -l sample1.vcf.gz`; filenames are not sample identifiers.\n3. Create the dataset explicitly, then ingest. Opening `mode=\"w\"` alone does not\n   create its schema. Materialize frequently read fields at creation time.\n4. Validate a small known interval and a round-trip export before scaling.\n\nFor existing sorted plain-text VCFs (run once per input; output names must be new):\n\n```bash\nbcftools view -Oz -o sample1.vcf.gz sample1.vcf\nbcftools index -c sample1.vcf.gz\nbcftools view -Oz -o sample2.vcf.gz sample2.vcf\nbcftools index -c sample2.vcf.gz\n```\n\nThe following local examples were exercised with sample IDs `S1` and `S2`, contig\n`chr1`, and coordinates 1-100; substitute the verified IDs and intervals in real data.\n\n```python\nimport tiledbvcf\n\nuri = \"cohort\"\nconfig = {\"sm.compute_concurrency_level\": \"2\", \"sm.io_concurrency_level\": \"2\"}\nwith tiledbvcf.Dataset(uri, mode=\"w\", tiledb_config=config) as ds:\n    ds.create_dataset(extra_attrs=[\"fmt_GT\", \"fmt_DP\"])\n    ds.ingest_samples(\n        [\"sample1.vcf.gz\"], threads=2, total_memory_budget_mb=512\n    )\n\n# Incremental ingestion uses the existing dataset; do not call create_dataset again.\nwith tiledbvcf.Dataset(uri, mode=\"w\", tiledb_config=config) as ds:\n    ds.ingest_samples(\n        [\"sample2.vcf.gz\"], threads=2, total_memory_budget_mb=512\n    )\n```\n\nThe schema-v4 data array has `contig`, start-coordinate, and sample **dimensions**;\nheaders and optional statistics are separate arrays in the dataset group. Fields\nnot explicitly materialized remain in the INFO\u002FFORMAT payloads. Query names such\nas `pos_start` need not match raw storage column names.\n\nUpstream supports parallel thread\u002Fprocess ingestion. Assign distinct sample work\nand coordinate lifecycle\u002Fmaintenance operations; do not blindly re-ingest samples.\n`resume=True` supports interrupted ingestion, not arbitrary duplicate correction.\n\n## Query completely and interpret coordinates correctly\n\n| Surface | Coordinate convention |\n| --- | --- |\n| Python `regions=[\"chr1:10-14\"]` | 1-based, closed interval |\n| Returned `pos_start`, `pos_end` | 1-based, inclusive record endpoints |\n| BED input and `query_bed_start`, `query_bed_end` | 0-based, half-open |\n| `read_variant_stats()` \u002F `read_allele_count()` column `pos` in 0.40.3 | **0-based**; add one before joining to VCF POS |\n\nQueries return **overlapping records**, not just records starting inside the\ninterval. A deletion spanning positions 10-14 appears in `chr1:14-14`, with\n`pos_start=10`. For BED `[13,14)`, the equivalent string is `chr1:14-14`.\nThe Python region parser requires explicit `contig:start-end`; bare `\"chr1\"`\nis rejected in this release. Use a known contig length for a whole-contig query.\n\n```python\ncfg = tiledbvcf.ReadConfig(memory_budget_mb=128, tiledb_config=config)\nattrs = [\"sample_name\", \"contig\", \"pos_start\", \"pos_end\", \"alleles\", \"fmt_GT\"]\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    assert {\"S1\", \"S2\"}.issubset(ds.samples())\n    assert set(attrs).issubset(ds.attributes())\n    for batch in ds.read_iter(\n        attrs=attrs, regions=[\"chr1:1-100\"], samples=[\"S1\", \"S2\"]\n    ):\n        print(batch)  # Replace with a bounded consumer or partitioned output.\n```\n\n- `read()` returns a pandas DataFrame; `read_arrow()` returns a PyArrow Table.\n  Either can return only the first memory-limited batch. Continue with\n  `continue_read()` \u002F `continue_read_arrow()` until `read_completed()`, or use\n  `read_iter()` for ordinary DataFrame queries. Avoid concatenating all batches\n  when the result cannot fit in memory.\n- `ReadConfig(memory_budget_mb=...)` controls reads; use\n  `ingest_samples(total_memory_budget_mb=...)` for writes. A read budget is not\n  a process RSS cap. `limit` truncates total output; it is not a pagination size.\n- `sample_partition=(0, 2)` selects partition zero of two. Likewise,\n  `region_partition=(index, number_of_partitions)` partitions query regions;\n  these tuples are not coordinate bounds, tile extents, or sample capacities.\n- **0.40.3 retains BED selection state** when a later call omits `bed_file`.\n  Open a fresh `Dataset` for an independent query when changing BED\u002Fregion inputs.\n- `fmt_GT` contains allele indices: 0 refers to REF, positive values index ALT,\n  and -1 means missing. Preserve multiallelic identity and ploidy; do not treat\n  every positive integer as a biallelic dosage or missing calls as reference.\n  The integer list alone does not encode the original phased GT string.\n\n## Export and validate\n\n```python\nfrom pathlib import Path\n\nPath(\"exported\").mkdir(exist_ok=True)\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    ds.export(\n        samples=[\"S1\"], regions=[\"chr1:1-100\"],\n        output_format=\"v\", output_dir=\"exported\",\n    )\n```\n\nPython export formats: `v` VCF, `z` compressed VCF, `u` uncompressed BCF, `b`\ncompressed BCF. `merge=False` writes per-sample files; `merge=True` requires a\ncombined `output_path`. A combined export is not joint calling. Re-index exported\nfiles before indexed downstream access. Compare sample IDs, contigs, record\ncounts, alleles, INFO\u002FFORMAT values, and missing\u002Fphased genotypes; semantic\nround-trip preservation does not imply byte-identical compression or headers.\n\nThe CLI accepts positional sample paths or `--samples-file` (one URI per line),\nnot a comma-separated `--samples` argument:\n\n```bash\ntiledbvcf create --uri cli_cohort\ntiledbvcf store --uri cli_cohort --threads 2 --total-memory-budget-mb 512 \\\n  -- sample1.vcf.gz sample2.vcf.gz\ntiledbvcf list --uri cli_cohort\ntiledbvcf stat --uri cli_cohort\ntiledbvcf export --uri cli_cohort --regions chr1:1-100 \\\n  --sample-names S1,S2 -Ot --tsv-fields 'SAMPLE,CHR,POS,REF,ALT,F:GT' \\\n  --output-path variants.tsv\n```\n\n`--` protects positional paths from variable-length options such as\n`--tiledb-config`. Check expected artifacts and sample counts as well as the exit\ncode: an invalid store command returned exit zero while printing an error in the\nreviewed CLI. TSV fields use `F:GT` for FORMAT\u002FGT and `I:DP` for INFO\u002FDP.\n\n## Cohort statistics and QC\n\n```python\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    stats = ds.read_variant_stats(regions=[\"chr1:1-100\"], drop_ref=True)\n    stats[\"vcf_pos\"] = stats[\"pos\"] + 1\n    print(stats[[\"contig\", \"vcf_pos\", \"alleles\", \"ac\", \"an\", \"af\"]])\n\nqc = tiledbvcf.sample_qc(uri, samples=[\"S1\"], config=config)\nprint(qc)\n```\n\nStatistics arrays are enabled by default at creation; older\u002Fdisabled arrays may\nnot support these operations. `read_variant_stats` has no `samples` argument.\nThe older `read_allele_frequency(dataset_uri, region)` wrapper accepts a single\nregion and delegates to the deprecated singular argument; prefer the Dataset\nmethod above. `sample_qc` uses `dataset_uri`, not `uri`, as its first argument.\n\nThe `ac`, `an`, and `af` values summarize the ingested cohort, not a selected\nancestry or query sample subset. A `read(samples=[...], set_af_filter=\">0.6\")`\nfilter still used cohort AF in the tested release. Compute subgroup frequencies\nfrom correctly selected calls with explicit missingness, ploidy, callable-region,\nand gVCF reference-block policies. Use an appropriate called-allele denominator,\nnot universally `2 * number_of_samples`; a missing VCF record is not proof of a\nhomozygous-reference call. QC metrics and sparse storage do not establish GWAS\nreadiness or scientific validity.\n\n## Object storage and TileDB Cloud\n\nDirect storage uses `s3:\u002F\u002Fbucket\u002Fpath`, `azure:\u002F\u002Fcontainer\u002Fpath`, or\n`gcs:\u002F\u002Fbucket\u002Fpath`, supported by the installed TileDB backend and its provider\ncredentials. This does not automatically distribute the computation. Use the\nprovider's credential chain or scoped configuration, and verify access to both\nthe dataset and source indexes. A TileDB Cloud token is distinct from bucket\ncredentials. Remote storage examples below are illustrative, source-verified,\nand not authenticated end-to-end tests.\n\nInstall `tiledb-cloud==0.14.4` in a compatible environment. Its `life-sciences`\nextra adds TileDB-SOMA, **not TileDB-VCF**. Supply `TILEDB_REST_TOKEN` through the\nexecution environment before importing the client; `TILEDB_REST_HOST` selects a\ncustom deployment if required. Do not embed tokens in code or print configuration.\n\n```python\n# Illustrative: requires an accessible registered cohort and a billing namespace.\nimport tiledb.cloud\nimport tiledb.cloud.vcf\n\ncloud_cfg = tiledb.cloud.Config()\nwith tiledbvcf.Dataset(\"tiledb:\u002F\u002Fmy-namespace\u002Fcohort\", tiledb_config=cloud_cfg) as ds:\n    sample_names = ds.samples()\n\nresult = tiledb.cloud.vcf.read(\n    dataset_uri=\"tiledb:\u002F\u002Fmy-namespace\u002Fcohort\",\n    config=cloud_cfg, attrs=[\"sample_name\", \"pos_start\", \"fmt_GT\"],\n    regions=[\"chr1:1-100\"], samples=sample_names[:2],\n    num_region_partitions=1, namespace=\"my-namespace\", max_workers=2,\n)\nframe = result.to_pandas()  # read returns an Arrow table, not a DataFrame.\n```\n\nDistributed reads assemble an Arrow result and may use significant worker\u002Fclient\nmemory. Bound the requested regions\u002Fsamples and worker count; partitioning does\nnot make the final concatenated result memory-free. This is SDK task execution,\nnot a paginated VCF REST-list API.\n\n```python\n# Illustrative, mutating cloud job: use only for an authorized ingestion task.\nsubmission = tiledb.cloud.vcf.ingest(\n    dataset_uri=\"s3:\u002F\u002Fmy-bucket\u002Fcohort\",\n    sample_list_uri=\"s3:\u002F\u002Fmy-bucket\u002Finputs\u002Fsample-uris.txt\",\n    namespace=\"my-namespace\", acn=\"registered-storage-role\",\n    register_name=\"cohort\", max_samples=2,\n    ingest_resources={\"cpu\": \"2\", \"memory\": \"4Gi\"},\n)\nprint(submission[\"graph_id\"])\n```\n\nUse exactly one of `search_uri`, `sample_list_uri`, or `metadata_uri`; file-search\npatterns apply to `search_uri`. Registration requires the access credential name\n(`acn`). `vcf.ingest` submits asynchronously and returns\n`{\"status\": \"started\", \"graph_id\": ...}`; track that graph to terminal success\nand verify ingested samples before claiming completion. There is no released\n`ingest_vcf_dataset(source=..., output=...)` API. Pricing, availability, security\ncontrols, and deployment obligations require the current account\u002Fservice terms.\n\n## Official references\n\n- [TileDB-VCF 0.40.3 release](https:\u002F\u002Fgithub.com\u002FTileDB-Inc\u002FTileDB-VCF\u002Freleases\u002Ftag\u002F0.40.3)\n- [Dataset API](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Freference\u002FDataset.html)\n- [Ingestion](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fingest-samples.html)\n- [Large queries](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fhandle-large-queries.html)\n- [Cloud storage](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fwork-with-cloud-object-stores.html)\n- [TileDB Cloud client release](https:\u002F\u002Fpypi.org\u002Fproject\u002Ftiledb-cloud\u002F0.14.4\u002F)\n- [Verification details and known release issues](references\u002Fverification.md)\n","# TileDB-VCF\n\n## When to use\n\nUse for cohort variant storage, incremental sample ingestion, interval queries, and\nexporting subsets for downstream genomics. TileDB-VCF stores records; it does not\nperform joint variant calling, association testing, normalization, or population\nstructure adjustment. Dataset size alone does not determine whether cloud execution\nis appropriate: benchmark the intended sample, region, and attribute workload.\n\nThis skill targets released **TileDB-VCF 0.40.3**. Native Python\u002FCLI checks used two\nsynthetic single-sample VCFs on macOS ARM64. Cloud client **0.14.4** contracts were\nchecked against its released wheel and mocked dispatch; hosted queries and jobs\nwere not executed. See [installation and verification notes](references\u002Fverification.md).\n\n## Install and verify the native stack\n\nThe `tiledbvcf` Python distribution is not published on PyPI at this review. The\nofficial **`tiledb` Conda channel** supplies `tiledbvcf-py` 0.40.3 for macOS ARM64,\nmacOS x86-64, and Linux x86-64, with Python 3.9-3.12 builds. `tiledb` on PyPI is\nTileDB-Py and does not install TileDB-VCF. Native Apple Silicon no longer requires\nforcing `CONDA_SUBDIR=osx-64`.\n\nFor the tested macOS ARM64 environment, pins below avoid versioned-library import\nfailures in the otherwise successful current Conda solve:\n\n```bash\nconda create -n tiledb-vcf -c tiledb -c conda-forge \\\n  python=3.12 tiledbvcf-py=0.40.3 \\\n  azure-core-cpp=1.16.2 azure-storage-blobs-cpp=12.16.0 \\\n  azure-storage-files-datalake-cpp=12.14.0 capnproto=1.4.0 c-blosc2=2.23.1\nconda activate tiledb-vcf\npython -c 'import tiledbvcf; print(tiledbvcf.version)'\ntiledbvcf version\n```\n\nThe equivalent Micromamba solve\u002Finstall was executed. These extra ABI pins are a\nverified macOS ARM64 workaround, not a claim about all platforms. Preserve the\nresolved environment for production. Other-platform installation and official\nDocker images are alternatives documented upstream, not tested here.\n\n## Prepare and create a cohort\n\n1. Record reference assembly, contig naming\u002Flengths, callers, normalization rules,\n   sample identity, and file checksums. Do not mix assemblies or assume `1` and\n   `chr1` are equivalent. Confirm compatible headers across inputs.\n2. Each input must contain **one sample**, be coordinate sorted, and have an index.\n   Use BGZF-compressed VCF or BCF with a matching `.csi`\u002F`.tbi`. Inspect\n   `bcftools query -l sample1.vcf.gz`; filenames are not sample identifiers.\n3. Create the dataset explicitly, then ingest. Opening `mode=\"w\"` alone does not\n   create its schema. Materialize frequently read fields at creation time.\n4. Validate a small known interval and a round-trip export before scaling.\n\nFor existing sorted plain-text VCFs (run once per input; output names must be new):\n\n```bash\nbcftools view -Oz -o sample1.vcf.gz sample1.vcf\nbcftools index -c sample1.vcf.gz\nbcftools view -Oz -o sample2.vcf.gz sample2.vcf\nbcftools index -c sample2.vcf.gz\n```\n\nThe following local examples were exercised with sample IDs `S1` and `S2`, contig\n`chr1`, and coordinates 1-100; substitute the verified IDs and intervals in real data.\n\n```python\nimport tiledbvcf\n\nuri = \"cohort\"\nconfig = {\"sm.compute_concurrency_level\": \"2\", \"sm.io_concurrency_level\": \"2\"}\nwith tiledbvcf.Dataset(uri, mode=\"w\", tiledb_config=config) as ds:\n    ds.create_dataset(extra_attrs=[\"fmt_GT\", \"fmt_DP\"])\n    ds.ingest_samples(\n        [\"sample1.vcf.gz\"], threads=2, total_memory_budget_mb=512\n    )\n\n# Incremental ingestion uses the existing dataset; do not call create_dataset again.\nwith tiledbvcf.Dataset(uri, mode=\"w\", tiledb_config=config) as ds:\n    ds.ingest_samples(\n        [\"sample2.vcf.gz\"], threads=2, total_memory_budget_mb=512\n    )\n```\n\nThe schema-v4 data array has `contig`, start-coordinate, and sample **dimensions**;\nheaders and optional statistics are separate arrays in the dataset group. Fields\nnot explicitly materialized remain in the INFO\u002FFORMAT payloads. Query names such\nas `pos_start` need not match raw storage column names.\n\nUpstream supports parallel thread\u002Fprocess ingestion. Assign distinct sample work\nand coordinate lifecycle\u002Fmaintenance operations; do not blindly re-ingest samples.\n`resume=True` supports interrupted ingestion, not arbitrary duplicate correction.\n\n## Query completely and interpret coordinates correctly\n\n| Surface | Coordinate convention |\n| --- | --- |\n| Python `regions=[\"chr1:10-14\"]` | 1-based, closed interval |\n| Returned `pos_start`, `pos_end` | 1-based, inclusive record endpoints |\n| BED input and `query_bed_start`, `query_bed_end` | 0-based, half-open |\n| `read_variant_stats()` \u002F `read_allele_count()` column `pos` in 0.40.3 | **0-based**; add one before joining to VCF POS |\n\nQueries return **overlapping records**, not just records starting inside the\ninterval. A deletion spanning positions 10-14 appears in `chr1:14-14`, with\n`pos_start=10`. For BED `[13,14)`, the equivalent string is `chr1:14-14`.\nThe Python region parser requires explicit `contig:start-end`; bare `\"chr1\"`\nis rejected in this release. Use a known contig length for a whole-contig query.\n\n```python\ncfg = tiledbvcf.ReadConfig(memory_budget_mb=128, tiledb_config=config)\nattrs = [\"sample_name\", \"contig\", \"pos_start\", \"pos_end\", \"alleles\", \"fmt_GT\"]\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    assert {\"S1\", \"S2\"}.issubset(ds.samples())\n    assert set(attrs).issubset(ds.attributes())\n    for batch in ds.read_iter(\n        attrs=attrs, regions=[\"chr1:1-100\"], samples=[\"S1\", \"S2\"]\n    ):\n        print(batch)  # Replace with a bounded consumer or partitioned output.\n```\n\n- `read()` returns a pandas DataFrame; `read_arrow()` returns a PyArrow Table.\n  Either can return only the first memory-limited batch. Continue with\n  `continue_read()` \u002F `continue_read_arrow()` until `read_completed()`, or use\n  `read_iter()` for ordinary DataFrame queries. Avoid concatenating all batches\n  when the result cannot fit in memory.\n- `ReadConfig(memory_budget_mb=...)` controls reads; use\n  `ingest_samples(total_memory_budget_mb=...)` for writes. A read budget is not\n  a process RSS cap. `limit` truncates total output; it is not a pagination size.\n- `sample_partition=(0, 2)` selects partition zero of two. Likewise,\n  `region_partition=(index, number_of_partitions)` partitions query regions;\n  these tuples are not coordinate bounds, tile extents, or sample capacities.\n- **0.40.3 retains BED selection state** when a later call omits `bed_file`.\n  Open a fresh `Dataset` for an independent query when changing BED\u002Fregion inputs.\n- `fmt_GT` contains allele indices: 0 refers to REF, positive values index ALT,\n  and -1 means missing. Preserve multiallelic identity and ploidy; do not treat\n  every positive integer as a biallelic dosage or missing calls as reference.\n  The integer list alone does not encode the original phased GT string.\n\n## Export and validate\n\n```python\nfrom pathlib import Path\n\nPath(\"exported\").mkdir(exist_ok=True)\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    ds.export(\n        samples=[\"S1\"], regions=[\"chr1:1-100\"],\n        output_format=\"v\", output_dir=\"exported\",\n    )\n```\n\nPython export formats: `v` VCF, `z` compressed VCF, `u` uncompressed BCF, `b`\ncompressed BCF. `merge=False` writes per-sample files; `merge=True` requires a\ncombined `output_path`. A combined export is not joint calling. Re-index exported\nfiles before indexed downstream access. Compare sample IDs, contigs, record\ncounts, alleles, INFO\u002FFORMAT values, and missing\u002Fphased genotypes; semantic\nround-trip preservation does not imply byte-identical compression or headers.\n\nThe CLI accepts positional sample paths or `--samples-file` (one URI per line),\nnot a comma-separated `--samples` argument:\n\n```bash\ntiledbvcf create --uri cli_cohort\ntiledbvcf store --uri cli_cohort --threads 2 --total-memory-budget-mb 512 \\\n  -- sample1.vcf.gz sample2.vcf.gz\ntiledbvcf list --uri cli_cohort\ntiledbvcf stat --uri cli_cohort\ntiledbvcf export --uri cli_cohort --regions chr1:1-100 \\\n  --sample-names S1,S2 -Ot --tsv-fields 'SAMPLE,CHR,POS,REF,ALT,F:GT' \\\n  --output-path variants.tsv\n```\n\n`--` protects positional paths from variable-length options such as\n`--tiledb-config`. Check expected artifacts and sample counts as well as the exit\ncode: an invalid store command returned exit zero while printing an error in the\nreviewed CLI. TSV fields use `F:GT` for FORMAT\u002FGT and `I:DP` for INFO\u002FDP.\n\n## Cohort statistics and QC\n\n```python\nwith tiledbvcf.Dataset(uri, cfg=cfg) as ds:\n    stats = ds.read_variant_stats(regions=[\"chr1:1-100\"], drop_ref=True)\n    stats[\"vcf_pos\"] = stats[\"pos\"] + 1\n    print(stats[[\"contig\", \"vcf_pos\", \"alleles\", \"ac\", \"an\", \"af\"]])\n\nqc = tiledbvcf.sample_qc(uri, samples=[\"S1\"], config=config)\nprint(qc)\n```\n\nStatistics arrays are enabled by default at creation; older\u002Fdisabled arrays may\nnot support these operations. `read_variant_stats` has no `samples` argument.\nThe older `read_allele_frequency(dataset_uri, region)` wrapper accepts a single\nregion and delegates to the deprecated singular argument; prefer the Dataset\nmethod above. `sample_qc` uses `dataset_uri`, not `uri`, as its first argument.\n\nThe `ac`, `an`, and `af` values summarize the ingested cohort, not a selected\nancestry or query sample subset. A `read(samples=[...], set_af_filter=\">0.6\")`\nfilter still used cohort AF in the tested release. Compute subgroup frequencies\nfrom correctly selected calls with explicit missingness, ploidy, callable-region,\nand gVCF reference-block policies. Use an appropriate called-allele denominator,\nnot universally `2 * number_of_samples`; a missing VCF record is not proof of a\nhomozygous-reference call. QC metrics and sparse storage do not establish GWAS\nreadiness or scientific validity.\n\n## Object storage and TileDB Cloud\n\nDirect storage uses `s3:\u002F\u002Fbucket\u002Fpath`, `azure:\u002F\u002Fcontainer\u002Fpath`, or\n`gcs:\u002F\u002Fbucket\u002Fpath`, supported by the installed TileDB backend and its provider\ncredentials. This does not automatically distribute the computation. Use the\nprovider's credential chain or scoped configuration, and verify access to both\nthe dataset and source indexes. A TileDB Cloud token is distinct from bucket\ncredentials. Remote storage examples below are illustrative, source-verified,\nand not authenticated end-to-end tests.\n\nInstall `tiledb-cloud==0.14.4` in a compatible environment. Its `life-sciences`\nextra adds TileDB-SOMA, **not TileDB-VCF**. Supply `TILEDB_REST_TOKEN` through the\nexecution environment before importing the client; `TILEDB_REST_HOST` selects a\ncustom deployment if required. Do not embed tokens in code or print configuration.\n\n```python\n# Illustrative: requires an accessible registered cohort and a billing namespace.\nimport tiledb.cloud\nimport tiledb.cloud.vcf\n\ncloud_cfg = tiledb.cloud.Config()\nwith tiledbvcf.Dataset(\"tiledb:\u002F\u002Fmy-namespace\u002Fcohort\", tiledb_config=cloud_cfg) as ds:\n    sample_names = ds.samples()\n\nresult = tiledb.cloud.vcf.read(\n    dataset_uri=\"tiledb:\u002F\u002Fmy-namespace\u002Fcohort\",\n    config=cloud_cfg, attrs=[\"sample_name\", \"pos_start\", \"fmt_GT\"],\n    regions=[\"chr1:1-100\"], samples=sample_names[:2],\n    num_region_partitions=1, namespace=\"my-namespace\", max_workers=2,\n)\nframe = result.to_pandas()  # read returns an Arrow table, not a DataFrame.\n```\n\nDistributed reads assemble an Arrow result and may use significant worker\u002Fclient\nmemory. Bound the requested regions\u002Fsamples and worker count; partitioning does\nnot make the final concatenated result memory-free. This is SDK task execution,\nnot a paginated VCF REST-list API.\n\n```python\n# Illustrative, mutating cloud job: use only for an authorized ingestion task.\nsubmission = tiledb.cloud.vcf.ingest(\n    dataset_uri=\"s3:\u002F\u002Fmy-bucket\u002Fcohort\",\n    sample_list_uri=\"s3:\u002F\u002Fmy-bucket\u002Finputs\u002Fsample-uris.txt\",\n    namespace=\"my-namespace\", acn=\"registered-storage-role\",\n    register_name=\"cohort\", max_samples=2,\n    ingest_resources={\"cpu\": \"2\", \"memory\": \"4Gi\"},\n)\nprint(submission[\"graph_id\"])\n```\n\nUse exactly one of `search_uri`, `sample_list_uri`, or `metadata_uri`; file-search\npatterns apply to `search_uri`. Registration requires the access credential name\n(`acn`). `vcf.ingest` submits asynchronously and returns\n`{\"status\": \"started\", \"graph_id\": ...}`; track that graph to terminal success\nand verify ingested samples before claiming completion. There is no released\n`ingest_vcf_dataset(source=..., output=...)` API. Pricing, availability, security\ncontrols, and deployment obligations require the current account\u002Fservice terms.\n\n## Official references\n\n- [TileDB-VCF 0.40.3 release](https:\u002F\u002Fgithub.com\u002FTileDB-Inc\u002FTileDB-VCF\u002Freleases\u002Ftag\u002F0.40.3)\n- [Dataset API](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Freference\u002FDataset.html)\n- [Ingestion](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fingest-samples.html)\n- [Large queries](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fhandle-large-queries.html)\n- [Cloud storage](https:\u002F\u002Ftiledb-inc.github.io\u002FTileDB-VCF\u002Fdocumentation\u002Fhow-to\u002Fwork-with-cloud-object-stores.html)\n- [TileDB Cloud client release](https:\u002F\u002Fpypi.org\u002Fproject\u002Ftiledb-cloud\u002F0.14.4\u002F)\n- [Verification details and known release issues](references\u002Fverification.md)\n",{"name":5,"license":43,"category":44,"metadata":45,"description":7,"compatibility":50},"MIT license","scientific-computing",{"version":46,"skill-author":47,"last-reviewed":48,"upstream-version":49},"1.3","Jeremy Leipzig","2026-10-01","TileDB-VCF 0.40.3; tiledb-cloud 0.14.4","Requires a native TileDB-VCF installation and Python 3.9-3.12 for the reviewed Conda builds. bcftools compresses\u002Findexes input fixtures. Installation and remote storage need network access; TileDB Cloud needs its Python client, account token, and storage permissions.","main","0da4485785ee599388900be50db1635489d53056",".agents\u002Fskills\u002Fkd-tiledbvcf","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6","https:\u002F\u002Fdisha6.6.thenitishkr.in\u002F","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh","www.thenitishkr","thenitishkr | AI • Cyber-security • Research\r\n","2026-11-08 03:33:25",{"source":8,"skillName":5,"commands":61,"prompt":71,"zipAvailable":72},[62,65,68],{"tool":63,"command":64},"npx","npx skills add Tashima-Tarsh\u002FDisha6.6 --skill tiledbvcf",{"tool":66,"command":67},"bunx","bunx skills add Tashima-Tarsh\u002FDisha6.6 --skill tiledbvcf",{"tool":69,"command":70},"pnpm","pnpm dlx skills add Tashima-Tarsh\u002FDisha6.6 --skill tiledbvcf","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkd-tiledbvcf\u002FSKILL.md",true,[74],{"path":75,"resourceType":76,"resourceTypeDesc":77,"sizeBytes":78,"sourceHtmlUrl":79},"references\u002Fverification.md","REFERENCE","参考文档",6119,"https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkd-tiledbvcf\u002Freferences\u002Fverification.md",[81,112,137,164,189,213],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":92,"publishedAt":38},"01M6QRD07D4TGNDK6VJ4EHF8BV","youtube-full","Use when the user needs YouTube transcripts, video search, channel browsing, playlist extraction, or content monitoring. Trigger phrases: 'get the transcript for', 'search YouTube for', 'what are the latest videos on', 'list this playlist', 'monitor this channel', or any request involving a YouTube URL, video ID, or @handle. Do NOT use for downloading video or audio files, YouTube engagement data (likes, comments), or private\u002Fage-restricted videos.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fyoutube-full\u002FSKILL.md",[87],{"categoryNo":88,"categoryName":89,"taxonomyType":90,"taxonomyTypeDesc":91},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",[93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,19,109,110,111],"youtube full","youtube","full","transcript","search","monitoring","data","Video","Audio","list","files","do","get","age","trigger","URL","browsing","request","youtube transcripts",{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":118,"publishedAt":38},"01M6QRCJ3112VYPA97YG7KGMC2","x-twitter-growth","X\u002FTwitter growth engine for building audience, crafting viral content, and analyzing engagement. Use when the user wants to grow on X\u002FTwitter, write tweets or threads, analyze their X profile, research competitors on X, plan a posting strategy, or optimize engagement. Complements social-content (generic multi-platform) with X-specific depth: algorithm mechanics, thread engineering, reply strategy, profile optimization, and competitive intelligence via web search.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fx-twitter-growth\u002FSKILL.md",[],[119,120,121,122,97,123,124,125,126,127,128,129,130,131,132,133,134,135,136,19],"twitter","x twitter growth","growth","research","content","engineering","plan","web search","competitors","analyze","engine","write","strategy","use","platform","grow","posting","competitive",{"skillNo":138,"name":139,"nameZh":6,"description":140,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":141,"categories":142,"keywords":146,"publishedAt":38},"01M6QRBN3EHQNJY42RERP7A149","write-a-skill","Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, build, or author a new skill.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fwrite-a-skill\u002FSKILL.md",[143],{"categoryNo":144,"categoryName":145,"taxonomyType":90,"taxonomyTypeDesc":91},"01JBSEEDCAT000000000000023","Skill与智能体开发",[147,130,148,149,150,151,152,153,132,154,155,51,156,157,158,159,160,161,162,163],"write a skill","github","create","https","productivity","matt pocock","direct","example","build","new skill","structure","new agent","resources","with","tree","progressive disclosure","author",{"skillNo":165,"name":166,"nameZh":6,"description":167,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":168,"categories":169,"keywords":170,"publishedAt":38},"01M6QRB6G8P33PNGAF9Y69EJ9T","wrap-up","Close out a launched Claude Managed Agent — recap every primitive the founder now owns, regenerate the single-file overview page, and suggest the next 1-2 upgrades. Use when the user says \"wrap up\", \"close this out\", \"what do I own now\", \"give me the summary\", \"recap the agent\", or when the orchestrator routes phase=wrap-up. primitives_inventory.py tables everything owned (agent, environment, session, memory, outcome, deployment); overview_page.py regenerates a self-contained .\u002Fmy-agent\u002Fagent-overview.html; upgrade_suggester.py ranks the next moves from recorded deferrals plus standing hardening steps. Companion to run-without-you; the last stop before phase=done.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fwrap-up\u002FSKILL.md",[],[171,172,173,174,175,176,177,178,179,132,180,104,181,182,183,184,185,186,187,188],"wrap up","up","wrap","gemini","deployment","run","file","hardening","closeout","stop","upgrade","founder","close","companion","you","tables","directions","phase",{"skillNo":190,"name":191,"nameZh":6,"description":192,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":193,"categories":194,"keywords":196,"publishedAt":38},"01M6QRA5AYC1RBFZ39GTWGS3N8","workflow-builder","Design and write deterministic multi-agent workflow scripts (.js files in .claude\u002Fworkflows\u002F) for Claude Code's Workflow tool. Use when a user wants to build, create, author, scaffold, or run a custom Claude Code workflow, orchestrate sub-agents (fan-out, pipeline, loop, judge-panel), or automate a repeatable multi-step task across fresh-context agents.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fworkflow-builder\u002FSKILL.md",[195],{"categoryNo":144,"categoryName":145,"taxonomyType":90,"taxonomyTypeDesc":91},[197,198,199,200,201,149,176,202,203,204,205,206,130,207,103,208,209,210,211,212],"workflow","workflow builder","builder","design","workflows","code","orchestrate","pipeline","loop","multi-agent","context","across","fresh","scripts","judge","claude code workflow",{"skillNo":214,"name":215,"nameZh":6,"description":216,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":217,"categories":218,"keywords":219,"publishedAt":38},"01M6QR7QMCZESKQA0ERW1E657M","weekly-review","Use when someone wants to run a weekly review, close open loops, audit stalled projects and commitments, get their system back to trusted, restart a lapsed review habit, or says \"\u002Fcs:weekly-review\". Walks David Allen's three-phase loop — GET CLEAR, GET CURRENT, GET CREATIVE — with deterministic scripts that inventory open loops, gate the checklist with named gaps, and score commitment health 0-100.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fweekly-review\u002FSKILL.md",[],[220,221,222,223,224,225,226,227,228,205,229,230,231,232,233,234,105,235,236,188],"review","weekly review","weekly","deterministic scripts","reflect","feeds","inventory","open","team","projects","project-management","scanner","conversation","back","project","sprint","cadence"]