A pipeline for phasing 200,031 sequenced genomes in the UK Biobank

This document contains instructions for running a pipeline that will filter, phase, and index 200,031 genomes in UK Biobank. The pipeline and its application to the first release of UK Biobank whole genome sequence data is described in "Statistical phasing of 150,119 sequenced genomes in the UK Biobank" [1, 2].

This pipeline is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY and without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Author: Brian Browning.
Last updated: November 18, 2022

Contents

• Before starting
• Upload resources to the cloud
• Phase the genomes
• References

Before starting

1. Obtain tier 3 access to the UK Biobank data.
2. Create a project on the UK Biobank Research Analysis Platform and dispense data to your project. It may take several hours before the data is fully dispensed.
3. Install the DNAnexus dx command line client on a unix or linux system. The dx command line client is part of the dx-toolkit.
4. If necessary, increase your DNAnexus spending limit. The cost of phasing chromosomes 1-22 and X with this pipeline is estimated to be between 3,200 and 5,000 British pounds (GBP), depending on how much of the data filtering can be performed on lower-cost "spot" compute instances. The cost of storing the compressed, phased data for chromosomes 1-22 and X is less than 15 British pounds per month.
5. In your DNAnexus user profile, set your job notification policy to "Only on Failure". This will reduce the number of e-mails that you receive. Access your user profile by clicking on the menu in the upper right corner of the Research Analysis Platform web interface. Then set your job notification policy in the "Email" section of the "User Account" tab.

Contents

Upload resources to the cloud

First, log into DNAnexus platform with the dx login command. Use the same username and password that you use for the Research Analysis Platform web interface.

Next, run the following three commands. When you run the third command, replace PROJECT with the name of your DNAnexus project.

git clone https://github.com/browning-lab/ukb-phasing.git
cd ukb-phasing
./upload.resources PROJECT

These commands will:

1. Download a copy of this git repository to your working directory
2. Change your working directory to the top-level "ukb-phasing" directory in the repository
3. Select the project that you substituted for "PROJECT" as your current DNAnexus project
4. Create three folders (apps/, maps/, and phased/) in your DNAnexus project.
5. Copy genetic maps for each chromosome to the maps/ folder and copy five DNAnexus applets to the apps/ folder.

Contents

Phase the genomes

From the top-level "ukb-phasing" directory in the downloaded repository, run the following command to phase a chromosome. When you run the command, replace PROJECT with the name of your DNAnexus project, and replace CHROM with 1, 2, ..., 22, or X.

./phase.ukb PROJECT CHROM

This command will filter the chromosome markers with bcftools [3], phase the filtered genotypes with Beagle [4], and index the output phased VCF file with tabix [5]. The run time for a chromosome is between 16 hours and three and a half days. Chromosomes can be processed in parallel.

The marker filter restricts the analysis to SNVs with AAScore > 0.95 and missing genotypes rate < 0.05. You can include structural variants by changing

filter='FILTER="PASS"&INFO/AAScore[*]>0.95&F_MISSING<0.05&TYPE="snp"'

to

filter='FILTER="PASS"&INFO/AAScore[*]>0.95&F_MISSING<0.05'

in the phase.ukb file. Changing the marker filter will affect both computational cost and phase accuracy [2].

The output phased VCF file is saved on the Research Analysis Platform as "PROJECT:/phased/chrCHROM.filt.phased.vcf.gz", where PROJECT is the name of your project, and CHROM is the chromosome number. A temporary folder, named "PROJECT:/tmp.chrCHROM/", is used to store interim files during the analysis. The temporary folder is deleted at the end of the analysis.

You can monitor the progress of your jobs with the Research Analysis Platform web interface. Click on the "PROJECTS" menu and choose "All Projects". Click on the name of your project in the list of projects, and then click on the "MONITOR" tab.

Contents

References

[1] B V Halldorsson, et al. The sequences of 150,119 genomes in the UK Biobank. Nature 2022 Jul; 607(7920):732-740. doi: 10.1038/s41586-022-04965-x.

[2] B L Browning, S R Browning. Statistical phasing of 150,119 sequenced genomes in the UK Biobank. doi: 10.1101/2022.10.03.510691.

[3] Danecek P, et al. Twelve years of SAMtools and BCFtools. Gigascience. 2021 Feb; 10(2):giab008. doi: 10.1093/gigascience/giab008.

[4] B L Browning, X Tian, Y Zhou, and S R Browning. Fast two-stage phasing of large-scale sequence data. Am J Hum Genet 2021 Oct; 108(10):1880-1890. doi: 10.1016/j.ajhg.2021.08.005.

[5] H Li. Tabix: fast retrieval of sequence features from generic TAB-delimited files. Bioinformatics, 2011 Mar; 27(5):718–719. doi: 10.1093/bioinformatics/btq671

Contents