Detect mosaic variants using 10x Genomics linked-read barcode information.
Mosaic variants can be confidently identified from sequence data using sequence read phasing (see Figure 1 of this article). mosaic-from-barcodes contains software to identify mosaic variants from VCF files with barcode information incorporated in BX tags, such as those produced by Long Ranger.
The identification of mosaic variants occurs according to the following steps:
- Variants are read from the input file and stored in a list. The barcodes of possible mosaic and germline variants are stored in a Python dictionary (hash table).
- Possible mosaic variants that occur some distance (default is 500,000bp) upstream of the current variant are phased to nearby germline variants.
- A sparse matrix of allele information from nearby germline variants (rows) and barcodes (columns) is constructed.
- The haplotype of each barcode is determined by majority vote. Low-quality barcodes are filtered.
- The candidate mosaic variant is phased to the identified haplotypes using the barcode information.
- If phasing indicates high-confidence mosaic status of the variant, a MOSAIC tag is added to the variant's INFO field.
- Variants twice as far upstream as the phasing position from the current variant are output and their barcode information is removed from the dictionary.
Clone the repository.
git clone git@github.com:DonFreed/mosaic-from-barcodes.git
Make a directory for storing test data and download input and annotation data.
mkdir mosaic-from-barcodes/test
cd mosaic-from-barcodes/test
# Download the input variants #
wget https://s3-us-west-2.amazonaws.com/10x.files/samples/genome/NA12878_WGS_210/NA12878_WGS_210_phased_variants.vcf.gz
# Download a dbSNP and 1000 Genomes VCF #
wget --user=gsapubftp-anonymous \
ftp://ftp.broadinstitute.org/bundle/2.8/b37/dbsnp_138.b37.vcf.gz \
ftp://ftp.broadinstitute.org/bundle/2.8/b37/1000G_phase3_v4_20130502.sites.vcf.gz
cd ..
Annotate the 10x VCF with information on variants in dbSNP and 1000 Genomes.
(zcat test/NA12878_WGS_210_phased_variants.vcf.gz | head -n 91; \
zcat test/NA12878_WGS_210_phased_variants.vcf.gz | tail -n +92 | sort -k1,1V -k2,2n) | \
python3 annotate_variants.py --ref_var_name ONEKG DBSNP \
--ref_var_file <(zcat test/1000G_phase3_v4_20130502.sites.vcf.gz) \
<(zcat test/dbsnp_138.b37.vcf.gz) > test/NA12878_10x_annotated.vcf
Identify mosaic variants from the annotated VCF.
python3 main.py --infile test/NA12878_10x_annotated.vcf --marks_germline ONEKG \
--absent_marks_possible_mosaic DBSNP --outfile test/NA12878_mosaic.vcf \
--ignore_filter 10X_PHASING_INCONSISTENT