Skip to content

Latest commit

 

History

History
145 lines (104 loc) · 9.17 KB

data-access-request-guide.md

File metadata and controls

145 lines (104 loc) · 9.17 KB
Raw

Hartwig Medical Foundation Data Access Request Guide

This page provides practical information on how to access / work with the data that will be made available to you within the context of a Data Access Request (DR) from Hartwig Medical Foundation.

Note: more details on the methods used to generate both the genomic and clinical data can be found on a separate Methods page.

Contents

General Notes

  • Please be aware that the sample IDs provided to you are not publicly shareable. In case you want to publish something with respect to a specific sample from the Hartwig Medical Foundation database you need to use the Hartwig Sample ID for this. For newer DRs these have been made available; for older requests these can be requested for the samples you have been provided.
  • Internally at Hartwig, we load up all data into a MySQL database. The scheme and code to set this up yourself can be found on our resources page.
  • After one month of the creation of a bucket, files in the bucket will be removed. On request, we can leave or make the data available for an extended period. After receipt of such a request, we will start making the licensed data available again within one workday. You can submit a request by sending an email to [email protected], mentioning your reference number. Note that you still retain continuous access to the files that are shared via manifest.json for the duration of the license agreement.

Please use the unique ID given to your request (eg. "DR-XXX") in any communication with us about your data request.

Sample selection

By default, in addition to data-request specific criteria, samples for which one of the below applies are excluded:

  • Samples with PURPLE qcStatus != PASS: Samples with poor quality; Samples without any tumor evidence; Samples with less than 19.5% tumor cells.
  • Samples with insufficient coverage.

Primary tumor location and type via DOID ontology

The primary tumor location and/or type of the samples in the database are mapped to the DOID ontology (as detailed as possible on available data, for information of the DOID ontology system see: https://www.ebi.ac.uk/ols/ontologies/doid). Please find the tree of the doids in the Hartwig Medical Foundation database here.

Format of the data made available

Clinical Data (TSV format)

Clinical data will be made available in a metadata.tsv via GCP.

Some notes about the clinical data:

  • As much information as possible from the Electronic Case Report Form (eCRF) of the respective clinical studies is gathered. Please be aware that records are not guaranteed to be complete.
  • For patients who participated in the DRUP study we can not share any treatment related information.
  • The "purity" field in the metadata is the percentage tumor cells derived from WGS data by PURPLE.

Please find more details on the methods used to generate both the genomic and clinical data on a separate Methods page.

Somatic Data (VCF/TXT formats)

Somatic data will be made available in a somatics.tar.gz via GCP.

Per sample the following files are present:

In the purple folder --

  • circos.png (genome wide plot)
  • driver.catalog.tsv (affected driver genes)
  • purple.cnv.gene.tsv (somatic copy number per gene)
  • purple.cnv.germline.tsv (germline copy number regions) [only in case germline level is part of request]
  • purple.cnv.somatic.tsv (somatic copy number regions)
  • purple.purity.range.tsv (in depth information about purity measure)
  • purple.purity.tsv (implied percentage of tumor cells)
  • purple.qc (quality control outcome)
  • purple.somatic.vcf.gz + purple.somatic.vcf.gz.tbi (somatic SNVs and small INDELs)
  • purple.sv.vcf.gz + purple.sv.vcf.gz.tbi (somatic structural variants)
  • purple.version (purple version used)

For an explanation of the contents of the purple. files, see PURPLE.

In the linx folder --

  • driver.catalog.tsv (reproduction of the driver catalog produced by PURPLE with homozygous disruptions events appended)
  • linx.breakend.tsv (impact of each non PON filtered break junction on each overlapping gene)
  • linx.clusters.tsv (clustering of all non PON filtered SV events and their resolved classification)
  • linx.drivers.tsv (linkage of drivers from driver catalog to SV cluster which contributed to those drivers)
  • linx.fusion.tsv (inframe and out of frame fusions)
  • linx.links.tsv (chromosomal segments joining break junction pairs predicted to be linked and phased in cis on the derivative chromosome)
  • linx.svs.tsv (structural variant annotations)
  • linx.viral_inserts.tsv (viral insertions)
  • linx.vis_copy_number.tsv
  • linx.vis_fusion.tsv
  • linx.vis_gene_exon.tsv
  • linx.vis_protein_domain.tsv
  • linx.vis_segments.tsv
  • linx.vis_sv_data.tsv *the .vis_ files are visualisations.
  • linx.version (linx version used)

For an explanation of the contents of the linx. files, see LINX.

Germline Data (VCF/TXT formats)

Germline data will be made available via GCP.

We share the SNVs and small INDELs called from the reference sample using GATK haplotype caller.

Aligned readout data (CRAM format)

Aligned readout data will be made available per sample via GCP.

Some notes to keep in mind:

Example loading CRAM file in IGV:

It is possible to directly load CRAM files into IGV using the Google Cloud Storage URL. Please note that to do this, IGV requires your permission to access both Google Cloud Storage and Google Drive. It is at this time not possible to exclude Google Drive from these permissions. To load a CRAM file directly from Google Cloud Storage:

  • Open the IGV program and choose option 'Google' > 'Login' from the menu bar
  • Log in to your Google Account
  • Set the GCP project that will be used for billing the (small amount of) egress cost by choosing 'Google' > 'Enter Project ID'
  • Choose option 'File' > 'Load from URL'
  • Paste the CRAM file gs:// URL at field 'File URL'
  • Paste the CRAI file gs:// URL at field 'Index URL'

Please find more details on the methods used to generate both the genomic and clinical data on a separate Methods page.

RNA-seq data (FASTQ format)

RNA-seq data will be made available per sample via GCP.

Some notes to keep in mind:

Example data (COLO829v003T)

COLO829v003T is a melanoma cell line that can be used for testing. The COLO829v003T somatic tar file (different pipeline versions) can be downloaded from our resources page. We also have the COLO829v003T available on the Google Cloud Platform (somatic and germline tar file, and the cram files). To be added to the ACL of these files please send an email to [email protected] including the GCP account the data should be made available for (please note this should be a GCP account set up with an institutional email address, see Getting Started with Google Cloud Platform).

More information