Population Genomics: Scalable Ingestion

life sciences
genomics (vcf)
tutorials
ingestion
Learn about how to ingest VCF data into TileDB at scale with this tutorial.
How to run this tutorial

You can run this tutorial in two ways:

  1. Locally on your machine.
  2. On TileDB Cloud.

However, since TileDB Cloud has a free tier, we strongly recommend that you sign up and run everything there, as that requires no installations or deployment.

This tutorial demonstrates a powerful, scalable solution provided by TileDB Cloud to ingest VCF data into a TileDB-VCF dataset, handling dataset sizes from a few samples to biobank scale.

You will ingest a subset of the publicly available 1000 Genomes Phase 3 Reanalysis with DRAGEN dataset, which is managed by Illumina and hosted on AWS Data Exchange. The ingested TileDB-VCF dataset will be stored on Amazon S3 and registered on TileDB Cloud.

Programmatic ingestion

TileDB-VCF batch ingestion is initiated by calling a single Python method in the tiledb.cloud Python package and is recommended for the following use cases:

  • Running ingestion in larger bioinformatics or data processing pipelines.
  • Recording ingestion parameters in a script or notebook for documentation and reproducibility.
  • Accessing advanced options that are not available in one-click ingestion.

Import the necessary libraries, load the appropriate environment variables, set the URIs used throughout the tutorial, and delete any previously created VCF datasets with the same name. If you are running this from a local notebook, visit the Tutorials: Basic TileDB Cloud for more information on how to set your TileDB Cloud credentials in a configuration object (this step can be omitted inside a TileDB Cloud notebook).

import os

import tiledb
import tiledb.cloud
import tiledb.cloud.vcf
import tiledbvcf

# Print library versions
print("TileDB core version: {}".format(tiledb.libtiledb.version()))
print("TileDB-Py version: {}".format(tiledb.version()))
print("TileDB-VCF version: {}".format(tiledbvcf.version))
print("TileDB-Cloud-Py version: {}".format(tiledb.cloud.version.version))

# You should set the appropriate environment variables with your keys.
# Get the keys from the environment variables.

tiledb_token = os.environ["TILEDB_REST_TOKEN"]
# or use your username and password (not recommended)
# tiledb_username = os.environ["TILEDB_USERNAME"]
# tiledb_password = os.environ["TILEDB_PASSWORD"]

# Log into TileDB Cloud
tiledb.cloud.login(token=tiledb_token)
# or use your username and password (not recommended)
# tiledb.cloud.login(username=tiledb_username, password=tiledb_password)

# Set the TileDB-VCF dataset URI
vcf_name = "scalable-ingestion"
user_profile = tiledb.cloud.user_profile()
s3_bucket = user_profile.default_s3_path.rstrip("/")
tiledb_account = user_profile.username
vcf_uri = os.path.join("tiledb://", tiledb_account, s3_bucket, vcf_name)

# Delete the dataset if it exists
if tiledb.object_type(vcf_uri, ctx=tiledb.cloud.Ctx()):
    tiledb.cloud.asset.delete(vcf_uri, recursive=True)

Now, start the ingestion using the parameters defined above. The search_uri and pattern arguments are configured to recursively search for VCF files in the public 1000 Genomes bucket. The remaining arguments are configured to reduce the time of the ingestion for this tutorial.

# One-line scalable VCF ingestion
status = tiledb.cloud.vcf.ingest(
    # URI of the dataset to be created
    dataset_uri=vcf_uri,
    # TileDB Cloud account to charge and where the dataset is registered
    namespace=tiledb_account,
    # Recursively search for VCF files in the public 1000 Genomes bucket
    search_uri="s3://1000genomes-dragen-v3.7.6/data/individuals/hg38-graph-based/",
    # Ingest files that match the pattern
    pattern="*.hard-filtered.vcf.gz",
    # Limit the number of files to ingest
    max_files=100,
    # Only ingest chromosome 21
    contigs=["chr21"],
    # Ingest 10 VCF files per batch
    vcf_batch_size=10,
    # Do not sign request to public-read bucket storing source VCF files
    config={"vfs.s3.no_sign_request": True},
)

The ingestion progress can be tracked in the task graph log on TileDB Cloud by following the batch->ingest_vcf, vcf-filter-uris, vcf-populate-manifest, vcf-filter-samples, and vcf-ingest-samples task graphs.

After the TileDB-VCF dataset is ingested and registered on TileDB Cloud, test reading variants from all samples in the region chr21:10000000-12000000.

# Open the dataset
ds = tiledbvcf.Dataset(vcf_uri, tiledb_config=tiledb.cloud.Config())

# Read a region in all samples
df = ds.read(regions="chr21:10000000-12000000")
df
sample_name contig pos_start alleles fmt_GT
0 HG00245 chr21 10000023 [A, C] [0, 1]
1 HG00246 chr21 10000023 [A, C] [0, 1]
2 HG00245 chr21 10000480 [C, T] [0, 1]
3 HG00246 chr21 10000480 [C, T] [0, 1]
4 HG00243 chr21 10000509 [A, G] [0, 1]
... ... ... ... ... ...
682548 HG00103 chr21 10814535 [C, A] [0, 1]
682549 HG00105 chr21 10814535 [C, A] [0, 1]
682550 HG00106 chr21 10814535 [C, A] [0, 1]
682551 HG00107 chr21 10814535 [C, A] [0, 1]
682552 HG00097 chr21 10814554 [C, T] [0, 1]

682553 rows × 5 columns

Finally, clean up by deleting the dataset from TileDB Cloud.

# Delete the dataset if it exists
if tiledb.object_type(vcf_uri, ctx=tiledb.cloud.Ctx()):
    tiledb.cloud.asset.delete(vcf_uri, recursive=True)

Ingestion from the UI console

TileDB Cloud provides a method to ingest a batch of VCFs into a TileDB-VCF dataset directly from its UI console. Visit Catalog: Genomics (VCF) for more details.