Ag3 data downloads
Contents
Ag3 data downloads#
This notebook provides information about how to download data from the Anopheles gambiae 1000 Genomes project (Ag1000G) phase 3. This includes sample metadata, raw sequence reads, sequence read alignments, and single nucleotide polymorphism (SNP) calls. Data from other releases can be accessed by changing the release in the examples from v3 to the specific Ag release, e.g. v3.1.
Code examples that are intended to be run via a Linux command line are prefixed with an exclamation mark (!). If you are running these commands directly from a terminal, remove the exclamation mark.
Examples in this notebook assume you are downloading data to a local folder within your home directory at the path ~/vo_agam_release_master_us_central1/. Change this if you want to download to a different folder on the local file system.
Data hosting#
Ag3 data are hosted by several different services.
Raw sequence reads in FASTQ format and sequence read alignments in BAM format are hosted by the European Nucleotide Archive (ENA). This guide provides examples of downloading data from ENA via FTP using the wget command line tool, but please note that there are several other options for downloading data, see the ENA documentation on how to download data files for more information.
SNP calls in VCF and Zarr formats are hosted on S3-compatible object storage at the Sanger Institute. This guide provides examples of downloading thes data using wget.
Sample metadata in CSV format are hosted on Google Cloud Storage (GCS) in the vo_agam_release_master_us_central1 bucket, which is a multi-region bucket located in the United States. All data hosted on GCS are publicly accessible and do not require any authentication to access. This guide provides examples of downloading data from GCS to a local computer using the wget and gsutil command line tools. For more information about gsutil, see the gsutil tool documentation.
Sample sets#
Data in this release are organised into 28 sample sets. Each of these sample sets corresponds to a set of mosquito specimens contributed by a collaborating study. Depending on your objectives, you may want to download data from only specific sample sets, or all sample sets. For convenience there is a tab-delimited manifest file listing all sample sets in the release. Here is a direct download link for the sample set manifest:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/manifest.tsv
The sample set manifest can also be downloaded via gsutil to a directory on the local file system, e.g.:
!mkdir -pv ~/vo_agam_release/v3/
!gsutil cp gs://vo_agam_release_master_us_central1/v3/manifest.tsv ~/vo_agam_release/v3/
Copying gs://vo_agam_release/v3/manifest.tsv...
/ [1 files][ 453.0 B/ 453.0 B]
Operation completed over 1 objects/453.0 B.
Here are the file contents:
!cat ~/vo_agam_release/v3/manifest.tsv
sample_set sample_count
AG1000G-AO 81
AG1000G-BF-A 181
AG1000G-BF-B 102
AG1000G-BF-C 13
AG1000G-CD 76
AG1000G-CF 73
AG1000G-CI 80
AG1000G-CM-A 303
AG1000G-CM-B 97
AG1000G-CM-C 44
AG1000G-FR 23
AG1000G-GA-A 69
AG1000G-GH 100
AG1000G-GM-A 74
AG1000G-GM-B 31
AG1000G-GM-C 174
AG1000G-GN-A 45
AG1000G-GN-B 185
AG1000G-GQ 10
AG1000G-GW 101
AG1000G-KE 86
AG1000G-ML-A 60
AG1000G-ML-B 71
AG1000G-MW 41
AG1000G-MZ 74
AG1000G-TZ 300
AG1000G-UG 290
AG1000G-X 297
For more information about these sample sets, see the section on sample sets in the introduction to Ag1000G phase 3.
Sample metadata#
Data about the samples that were sequenced to generate this data resource are available, including the time and place of collection, the gender of the specimen, and our call regarding the species of the specimen.
Specimen collection metadata#
Specimen collection metadata can be downloaded from GCS. E.g., here is the download link for the sample metadata for sample set AG1000G-BF-A:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/metadata/general/AG1000G-BF-A/samples.meta.csv
Sample metadata for all sample sets can also be downloaded using gsutil:
!mkdir -pv ~/vo_agam_release/v3/metadata/
!gsutil -m rsync -r gs://vo_agam_release_master_us_central1/v3/metadata/ ~/vo_agam_release/v3/metadata/
Building synchronization state...
Starting synchronization...
Here are the first few rows of the sample metadata for sample set AG1000G-BF-A:
!head ~/vo_agam_release/v3/metadata/general/AG1000G-BF-A/samples.meta.csv
sample_id,partner_sample_id,contributor,country,location,year,month,latitude,longitude,sex_call
AB0085-Cx,BF2-4,Austin Burt,Burkina Faso,Pala,2012,7,11.151,-4.235,F
AB0086-Cx,BF2-6,Austin Burt,Burkina Faso,Pala,2012,7,11.151,-4.235,F
AB0087-C,BF3-3,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0088-C,BF3-5,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0089-Cx,BF3-8,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0090-C,BF3-10,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0091-C,BF3-12,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0092-C,BF3-13,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
AB0094-Cx,BF3-17,Austin Burt,Burkina Faso,Bana Village,2012,7,11.233,-4.472,F
The sample_id columns gives the sample identifier used throughout all Ag1000G analyses.
The country, location, latitude and longitude columns give the location where the specimen was collected.
The year and month columns give the approximate date when the specimen was collected.
The sex_call column gives the gender as determined from the sequence data.
Species calls#
We have made a preliminary call for each specimen as to which species it belongs to (Anopheles gambiae, Anopheles coluzzii, Anopheles arabiensis) based on the genotypes of the samples. These calls were made from the sequence data, and there are cases where the species is not easy to determine. We report species calls using ancestry informative markers (AIMs).
AIM species calls can be downloaded from GCS, e.g., for sample set AG1000G-BF-A:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/metadata/species_calls_aim_20220528/AG1000G-BF-A/samples.species_aim.csv
Alternatively if you ran the gsutil rsync command above to download sample metadata then this file will already be present on your local file system.
Here are the first few rows of the AIM species calls for sample set AG1000G-BF-A:
!head ~/vo_agam_release/v3/metadata/species_calls_aim_20220528/AG1000G-BF-A/samples.species_aim.csv
sample_id,aim_species_fraction_arab,aim_species_fraction_colu,aim_species_fraction_colu_no2L,aim_species_gambcolu_arabiensis,aim_species_gambiae_coluzzii,aim_species
AB0085-Cx,0.0007664303506418854,0.014316392269148175,0.016488046166529265,gambcolu,gambiae,gambiae
AB0086-Cx,0.0024904214559386974,0.023571428571428573,0.02631578947368421,gambcolu,gambiae,gambiae
AB0087-C,0.0017244682889442423,0.9247311827956989,0.9810074318744839,gambcolu,coluzzii,coluzzii
AB0088-C,0.001915341888527102,0.8686288585786073,0.9875930521091811,gambcolu,coluzzii,coluzzii
AB0089-Cx,0.0028735632183908046,0.8567335243553008,0.9735973597359736,gambcolu,coluzzii,coluzzii
AB0090-C,0.0011500862564692352,0.972083035075161,0.9777411376751854,gambcolu,coluzzii,coluzzii
AB0091-C,0.0015322735108216816,0.9157060518731989,0.9750830564784053,gambcolu,coluzzii,coluzzii
AB0092-C,0.0013417672992141077,0.870157819225251,0.9801652892561984,gambcolu,coluzzii,coluzzii
AB0094-Cx,0.002106069308826345,0.8658536585365854,0.984297520661157,gambcolu,coluzzii,coluzzii
The species_gambcolu_arabiensis column provides a call as to whether the specimen is arabiensis or not (gamb_colu).
The species_gambiae_coluzzii column applies to samples that are not arabiensis, and differentiates gambiae versus coluzzii.
Raw sequence reads (FASTQ format)#
The raw sequence reads used in this data release can be downloaded from ENA. Note that for most samples there were multiple sequencing runs, and hence there are usually multiple ENA run accessions per sample. For most samples there were 3 sequencing runs, but some samples have 4 and some have a single sequencing run.
To find the ENA run accessions for a given sample, first download the catalog of run accessions:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/metadata/ena_runs.csv
Alternatively if you ran the gsutil rsync command above to download sample metadata then this file will already be present on your local file system. Inspect the file:
!head ~/vo_agam_release/v3/metadata/ena_runs.csv
sample_id,ena_run
AR0001-C,ERR347035
AR0001-C,ERR347047
AR0001-C,ERR352136
AR0002-C,ERR328585
AR0002-C,ERR323844
AR0002-C,ERR328597
AR0004-C,ERR343648
AR0004-C,ERR343636
AR0004-C,ERR343468
For example, the sequence reads for sample AR0001-C are available from three ENA accessions: ERR347035, ERR347047 and ERR352136. To download the sequence reads, visit the ENA website and search for these accessions. E.g., links to download sequence reads for run ERR352136 are available from this web page: https://www.ebi.ac.uk/ena/browser/view/ERR352136.
Note that FASTQ files are relatively large, several GB per sample, so they may take a long time to download, and may require a substantial amount of disk space on your local system.
Sequence read alignments (BAM format)#
Analysis-ready sequence read alignments are available in BAM format for all samples in the release and can be downloaded from ENA. A catalog file mapping sample identifiers to ENA accessions is available at this link:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/metadata/ena_alignments.csv
Alternatively if you ran the gsutil rsync command above to download sample metadata then this file will already be present on your local file system. Here are the first few rows:
!head ~/vo_agam_release/v3/metadata/ena_alignments.csv
sample_id,ena_analysis
AR0001-C,ERZ1695275
AR0002-C,ERZ1695276
AR0004-C,ERZ1695277
AR0006-C,ERZ1695278
AR0007-C,ERZ1695279
AR0008-C,ERZ1695280
AR0009-C,ERZ1695281
AR0010-Cx,ERZ1695282
AR0011-C,ERZ1695283
Each row in this file provides a mapping from Ag1000G sample identifiers to ENA analysis accessions. To find links for downloading the data, visit the ENA website and search for the corresponding analysis accession. E.g., links to download the analysis-ready BAM file for sample AR0001-C can be found from the web page at ERZ1695275.
Note that BAM files are relatively large, approximately 10G per sample, so they may take a long time to download, and may require a substantial amount of disk space on your local system.
SNP calls (VCF format)#
SNP genotypes#
SNP genotypes for individual mosquitoes in VCF format are available for download from Sanger S3-compatible object storage. A VCF file is available for each individual sample. To download a VCF file for a given sample, you will need the sample identifier and the sample set in which the sample belongs. Then inspect the data catalog in the metadata. E.g., for sample set AG1000G-AO:
!head ~/vo_agam_release/v3/metadata/general/AG1000G-AO/wgs_snp_data.csv | cut -f1,4 -d,
sample_id,snp_genotypes_vcf
AR0047-C,https://vo_agam_output.cog.sanger.ac.uk/AR0047-C.vcf.gz
AR0049-C,https://vo_agam_output.cog.sanger.ac.uk/AR0049-C.vcf.gz
AR0051-C,https://vo_agam_output.cog.sanger.ac.uk/AR0051-C.vcf.gz
AR0061-C,https://vo_agam_output.cog.sanger.ac.uk/AR0061-C.vcf.gz
AR0078-C,https://vo_agam_output.cog.sanger.ac.uk/AR0078-C.vcf.gz
AR0080-C,https://vo_agam_output.cog.sanger.ac.uk/AR0080-C.vcf.gz
AR0084-C,https://vo_agam_output.cog.sanger.ac.uk/AR0084-C.vcf.gz
AR0097-C,https://vo_agam_output.cog.sanger.ac.uk/AR0097-C.vcf.gz
AR0072-C,https://vo_agam_output.cog.sanger.ac.uk/AR0072-C.vcf.gz
A VCF file and associated tabix index can be downloaded via wget, e.g.:
!wget --no-clobber https://vo_agam_output.cog.sanger.ac.uk/AR0047-C.vcf.gz
!wget --no-clobber https://vo_agam_output.cog.sanger.ac.uk/AR0047-C.vcf.gz.tbi
--2022-12-01 13:36:09-- https://vo_agam_output.cog.sanger.ac.uk/AR0047-C.vcf.gz
Resolving vo_agam_output.cog.sanger.ac.uk (vo_agam_output.cog.sanger.ac.uk)... 193.62.203.62, 193.62.203.61, 193.62.203.63
Connecting to vo_agam_output.cog.sanger.ac.uk (vo_agam_output.cog.sanger.ac.uk)|193.62.203.62|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 3222999974 (3.0G) [application/gzip]
Saving to: ‘AR0047-C.vcf.gz’
AR0047-C.vcf.gz 1%[ ] 50.91M 8.62MB/s eta 5m 53s ^C
--2022-12-01 13:36:17-- https://vo_agam_output.cog.sanger.ac.uk/AR0047-C.vcf.gz.tbi
Resolving vo_agam_output.cog.sanger.ac.uk (vo_agam_output.cog.sanger.ac.uk)... 193.62.203.62, 193.62.203.61, 193.62.203.63
Connecting to vo_agam_output.cog.sanger.ac.uk (vo_agam_output.cog.sanger.ac.uk)|193.62.203.62|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 258551 (252K) [application/gzip]
Saving to: ‘AR0047-C.vcf.gz.tbi’
AR0047-C.vcf.gz.tbi 100%[===================>] 252.49K --.-KB/s in 0.04s
2022-12-01 13:36:17 (5.78 MB/s) - ‘AR0047-C.vcf.gz.tbi’ saved [258551/258551]
Note that each of these VCF files is around 3 Gb, so downloading may take some time, and sufficient local storage will be needed.
Each of these VCF files is an “all sites” VCF file, meaning that genotypes have been called at all genomic positions where the reference nucleotide is not “N”, regardless of whether variation is observed in the given sample. This means that VCFs from multiple samples can be merged easily to create a multi-sample VCF, which may be required for certain analyses. For example, the code below merges VCFs for three samples for chromosome arm 3R up to 1 Mbp:
!bcftools merge --output-type z --regions 3R:1-1000000 --output merged.vcf.gz AR0001-C.vcf.gz AR0002-C.vcf.gz AR0004-C.vcf.gz
If you are just interested in analysing variants within a given set of samples, you might like to filter the merged VCF to remove non-variant sites and alleles, e.g., using bcftools view:
!bcftools view --output-type z --output-file merged_variant.vcf.gz --min-ac 1:nonmajor --trim-alt-alleles merged.vcf.gz
Site filters#
SNP calling is not always reliable, and we have created some site filters to allow excluding low quality SNPs. We have created some sites-only VCF files with site filter information in the FILTER column. These VCF files are hosted on GCS.
Because different species may have different genome accessibility issues, we have created three separate site filters:
The “gamb_colu” site filter is designed for working only with samples that are not An. arabiensis.
The “arab” filter is designed for when only working with samples that are An. arabiensis.
The “gamb_colu_arab” filter is suitable for when analysing samples of any species together.
Each filter is available as a set of VCF files, one per chromosome arm. E.g., here is the direct download link for the gamb_colu_arab filters on chromosome arm 3R:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/site_filters/dt_20200416/vcf/gamb_colu_arab/3R_sitefilters.vcf.gz
Alternatively, all site filters VCFs can be downloaded using gsutil, e.g.:
!mkdir -pv ~/vo_agam_release/v3/site_filters/dt_20200416/vcf/
!gsutil -m rsync -r \
gs://vo_agam_release_master_us_central1/v3/site_filters/dt_20200416/vcf/ \
~/vo_agam_release/v3/site_filters/dt_20200416/vcf/
SNP calls (Zarr format)#
SNP data are also available in Zarr format, which can be convenient and efficient to use for certain types of analysis. These data can be analysed directly in the cloud without downloading to the local system, see the Ag3 cloud data access guide for more information. The data can also be downloaded to your own system for local analysis if that is more convenient. Below are examples of how to download the Zarr data to your local system.
The data are organised into several Zarr hierarchies.
SNP sites and alleles#
Data on the genomic positions (sites) and reference and alternate alleles that were genotyped can be downloaded as follows:
!mkdir -pv ~/vo_agam_release/v3/snp_genotypes/all/sites/
!gsutil -m rsync -r \
gs://vo_agam_release_master_us_central1/v3/snp_genotypes/all/sites/ \
~/vo_agam_release/v3/snp_genotypes/all/sites/
Site filters#
SNP calling is not always reliable, and we have created some site filters to allow excluding low quality SNPs. To download site filters data in Zarr format, excluding some parts of the data that you probably won’t need:
!mkdir -pv ~/vo_agam_release/v3/site_filters/
!gsutil -m rsync -r \
-x '.*vcf.*|.*crosses_stats.*|.*[MG]Q10.*|.*[MG]Q30.*|.*[MG]Q_mean.*|.*[MG]Q_std.*|.*/lo_.*|.*/hi_.*|.*no_cov.*|.*allele_consistency.*|.*heterozygosity.*' \
gs://vo_agam_release_master_us_central1/v3/site_filters/ \
~/vo_agam_release/v3/site_filters/
SNP genotypes#
SNP genotypes are available for each sample set separately. E.g., to download SNP genotypes in Zarr format for sample set AG1000G-BF-A, excluding some data you probably won’t need:
!mkdir -pv ~/vo_agam_release/v3/snp_genotypes/all/AG1000G-BF-A/
!gsutil -m rsync -r \
-x '.*/calldata/(AD|GQ|MQ)/.*' \
gs://vo_agam_release_master_us_central1/v3/snp_genotypes/all/AG1000G-BF-A/ \
~/vo_agam_release/v3/snp_genotypes/all/AG1000G-BF-A/
Copy number variation (CNV) data#
Data on copy number variation within the Ag3 cohort are available as three separate data types:
HMM – Genome-wide inferences of copy number state within each individual mosquito in 300 bp non-overlapping windows.
Coverage calls – Genome-wide copy number variant calls, derived from the HMM outputs by analysing contiguous regions of elevated copy number state then clustering of variants across individuals based on breakpoint proximity.
Discordant read calls – Copy number variant calls at selected insecticide resistance loci, based on analysis of read alignments at CNV breakpoints.
For more information on the methods used to generate these data, see the variant-calling methods page.
For each of these data types, data can be downloaded from Google Cloud Storage, and are available in either VCF or Zarr format.
CNV HMM#
The HMM inferences of copy number state are available in VCF, Zarr and text formats, and are organised by sample set.
For example, the VCF file for sample set AG1000G-AO can be downloaded from the following URL:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/cnv/AG1000G-AO/hmm/vcf/AG1000G-AO_cnv_hmm.vcf.gz
VCF files for all samples sets can be downloaded via gsutil as follows:
# create a local directory to hold downloaded CNV data
!mkdir -pv ~/vo_agam_release/v3/cnv/
# download the HMM data in VCF format for all sample sets
!gsutil -m rsync -r \
-x '.*/coverage_calls/.*|.*/discordant_read_calls/.*|.*/hmm/zarr/.*|.*/hmm/per_sample/.*' \
gs://vo_agam_release_master_us_central1/v3/cnv/ ~/vo_agam_release/v3/cnv/
Zarr files for all sample sets can be downloaded as follows:
# download HMM data in Zarr format for all sample sets
!gsutil -m rsync -r \
-x '.*/coverage_calls/.*|.*/discordant_read_calls/.*|.*/hmm/vcf/.*|.*/hmm/per_sample/.*' \
gs://vo_agam_release_master_us_central1/v3/cnv/ ~/vo_agam_release/v3/cnv/
CNV coverage calls#
Coverage-based CNV calls are available in VCF and Zarr formats, and are organised by sample set. Additionally the coverage calls were performed separately for An. arabiensis and (An. gambiae + An. coluzzii), and so are subdivided into “arab” and “gamb_colu” datasets, depending on which species are present in a given sample set.
Note that some samples were excluded from coverage calling because of high coverage variance.
For example, the VCF file for sample set AG1000G-AO and the gamb_colu callset can be downloaded from:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/cnv/AG1000G-AO/coverage_calls/gamb_colu/vcf/AG1000G-AO_gamb_colu_cnv_coverage_calls.vcf.gz
VCF files for all sample sets can be downloaded with:
# download coverage calls in VCF format for all sample sets
!gsutil -m rsync -r \
-x '.*/hmm/.*|.*/discordant_read_calls/.*|.*/coverage_calls/.*/zarr/.*' \
gs://vo_agam_release_master_us_central1/v3/cnv/ ~/vo_agam_release/v3/cnv/
Zarr files for all sample sets can be downloaded with:
# download coverage calls in Zarr format for all sample sets
!gsutil -m rsync -r \
-x '.*/hmm/.*|.*/discordant_read_calls/.*|.*/coverage_calls/.*/vcf/.*' \
gs://vo_agam_release_master_us_central1/v3/cnv/ ~/vo_agam_release/v3/cnv/
CNV discordant read calls#
CNV calls for selected insecticide resistance loci are available in VCF and Zarr formats, and are organised by sample set.
For example, the VCF file for sample set AG1000G-AO can be downloaded from:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/cnv/AG1000G-AO/discordant_read_calls/vcf/AG1000G-AO_cnv_discordant_read_calls.vcf.gz
VCF and Zarr files for all sample sets can be downloaded with:
# download discordant read calls for all sample sets
!gsutil -m rsync -r \
-x '.*/hmm/.*|.*/coverage_calls/.*' \
gs://vo_agam_release_master_us_central1/v3/cnv/ ~/vo_agam_release/v3/cnv/
Haplotypes#
The Ag3 data resource also includes haplotype reference panels, which were obtained by phasing the SNP calls. To allow for a range of different analyses, three different haplotype reference panels were constructed, each using a different subset of samples and applying different site filters:
“gamb_colu_arab” - This haplotype reference panel includes all wild-caught samples, and phases biallelic SNPs passing the “gamb_colu_arab” site filters.
“gamb_colu” - This haplotype reference panel includes wild-caught samples assigned as either An. gambiae, An. coluzzii or intermediate An. gambiae/An. coluzzii via the AIM species calling method, and phases biallelic SNPs passing the “gamb_colu” site filters.
“arab” - This haplotype reference panel includes wild-caught samples assigned as An. arabiensis via the AIM species calling method, and phases biallelic SNPs passing the “arab” site filters.
Haplotype data can be downloaded in either VCF or Zarr format. See the subsections below for further details
Haplotype reference panels (VCF format)#
These are the VCFs created by the phasing pipeline, containing all samples included each of the phasing runs. There is one VCF per phasing analysis per chromosome arm. The URL for each file has the following structure:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/snp_haplotypes/panel/{analysis}/ag3_{analysis}_{contig}_phased.vcf.gz
…where {analysis} is one of “gamb_colu_arab”, “gamb_colu” or “arab”; and {contig} is one of “2R”, “2L”, “3R”, “3L”, “X”.
E.g., the panel VCF for the gamb_colu_arab analysis for chromosome arm 3L can be downloaded here:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/snp_haplotypes/panel/gamb_colu_arab/ag3_gamb_colu_arab_3L_phased.vcf.gz
Note that these files can be large, up to ~50 GB.
If you’d like to download all of the panel files, you could also use gsutil, e.g.:
# create a local directory to store the data
!mkdir -pv ~/vo_agam_release/v3/snp_haplotypes/panel/
# copy files from cloud to local file system
!gsutil -m rsync -r \
-x '.*/.*zarr.zip' \
gs://vvo_agam_release_master_us_central1/v3/snp_haplotypes/panel/ \
~/vo_agam_release/v3/snp_haplotypes/panel/
Sample set haplotypes (VCF format)#
These VCFs are subsets of the panel VCFs, containing only samples in a given sample set. There is one VCF per sample set, per phasing analysis, per chromosome arm. The URL for each file has the following structure:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/snp_haplotypes/{sample_set}/{analysis}/vcf/AG1000G-AO_{analysis}_{contig}_phased.vcf.gz
…where {analysis} is one of “gamb_colu_arab”, “gamb_colu” or “arab”; {contig} is one of “2R”, “2L”, “3R”, “3L”, “X”; and {sample_set} is one of the Ag1000G phase 3 sample sets.
E.g., the VCF for sample set AG1000G-AO, for the gamb_colu analysis, for chromosome arm 2L, can be downloaded here:
https://storage.googleapis.com/vo_agam_release_master_us_central1/v3/snp_haplotypes/AG1000G-AO/gamb_colu/vcf/AG1000G-AO_gamb_colu_2L_phased.vcf.gz
If you’d like to download all of the VCF files for a given sample set, you could also use gsutil, e.g.:
# create a local directory to store the data
!mkdir -pv ~/vo_agam_release/v3/snp_haplotypes/AG1000G-AO/
# copy files from cloud to local file system
!gsutil -m rsync -r \
-x '.*/zarr/.*' \
gs://vo_agam_release_master_us_central1/v3/snp_haplotypes/AG1000G-AO/ \
~/vo_agam_release/v3/snp_haplotypes/AG1000G-AO/
Sample set haplotypes (Zarr format)#
These contain the haplotype data in Zarr format, with one Zarr hierarchy per sample set. The root zarr path for a given hierarchy has the following structure:
gs://vo_agam_release_master_us_central1/v3/snp_haplotypes/{sample_set}/{analysis}/zarr
Data can be downloaded with gsutil. E.g., download the Zarr data for sample set AG1000G. Note that the sites are stored in a separate hierarchy:
# create local directories to store the data
!mkdir -pv ~/vo_agam_release/v3/snp_haplotypes/sites
!mkdir -pv ~/vo_agam_release/v3/snp_haplotypes/AG1000G-AO
# copy haplotype data from cloud to local file system
!gsutil -m rsync -r \
-x '.*/vcf/.*' \
gs://vo_agam_release_master_us_central1/v3/snp_haplotypes/AG1000G-AO/ \
~/vo_agam_release/v3/snp_haplotypes/AG1000G-AO/
# copy phased sites data from cloud to local file system
!gsutil -m rsync -rn \
gs://vo_agam_release_master_us_central1/v3/snp_haplotypes/sites/ \
~/vo_agam_release/v3/snp_haplotypes/sites/
Feedback and suggestions#
If there are particular analyses you would like to run, or if you have other suggestions for useful documentation we could add to this site, we would love to know, please get in touch via the malariagen/vector-data GitHub discussion board.