Skip to content

Latest commit

 

History

History
46 lines (36 loc) · 3.49 KB

array_discordance.md

File metadata and controls

46 lines (36 loc) · 3.49 KB

Array discordance pipeline for TOPMed

Prepare array data

The following assumes that the array data is dbGaP fingerprint files with 10,000 SNPs.

  1. Create directory for this study under /projects/topmed/qc/compare_to_arrays/array_data/. This is where we store results associated with the array data and not with any particular freeze.
  2. Convert tped to bed with tped2bed.sh <file>. bed file will be written in same directory as tped (under /projects/topmed/downloaded_data/prior_array_data/)
  3. If necessary, liftover the .bim file to match the sequence build with liftover_bim.R.
  4. Convert bed to gds with qsub -N bed2gds runRscript.sh bed2gds.R <config>. An AnnotatedDataFrame with sample annotation will be created also.

If we have an alternate array source, modify the above code so as to end up with a GDS file in the same build as the sequence data, and an AnnotatedDataFrame with columns "sample.id" and "subject.id". "sample.id" should be unique, and "subject.id" should match the "submitted_subject_id" in the TOPMed sample annotation.

Compare array to sequence

  1. Create directory for this study and the specific freeze to be checked under /projects/topmed/qc/compare_to_arrays/<freeze>/.

  2. Subset sequencing GDS to overlapping variants and merge chromosomes: array_subset.py <config>

    config parameter default value description
    out_prefix Prefix for files created by this script.
    array_gds_file Path to array GDS file.
    seq_gds_file Path to sequencing GDS file.
    seq_annot_file Path to sequencing sample annotation file.
    subset_gds_file Path to output file.
    study Study name
    maf_threshold 0.05 Minimum MAF for sequence variants to include
    missing_threshold 0.01 Maximum missing call rate for sequence variants to include

  3. If we have an alternate array source, check the overlap between fingerprints and the subset file with array_fingerprint_overlap.R. If the number of overlapping variants is much less than 10,000, the code will supplement with a randomly selected set of variants that overlap between the sequence and array data, then combine these variants and the fingerprint variants in a GRanges object.

  4. In the following steps, use a config file that sets granges_include_file to dbgap_fp_ranges.hg38.RData, or to the alternate include file defined in the previous step.

  5. Run duplicate discordance with array_disc.py -n N <config> where N is the number of sample blocks to run in parallel.

    config parameter default value description
    out_prefix Prefix for files created by this script.
    array_gds_file Path to array GDS file.
    array_annot_file Path to array sample annotation file.
    seq_gds_file Path to sequencing GDS file (same as subset_gds_file above).
    seq_annot_file Path to sequencing sample annotation file.
    study Study name
    granges_include_file NA RData file with GRanges defining subset of variants.
    sample_include_file NA RData file with sample.id to include, if different from all samples in study.
    variant_include_file NA RData file with variants to include (result will be the intersection with granges_include_file)

  6. If any samples were discordant, save an RData file with a vector of those sample ids. Check them against all other array samples with qsub -t 1-N -N match_samples runRscript.sh -s array_match_sample.R <config> where N is the number of samples to match.