This workflow runs demuxlet to deconvolute sample identity when multiple samples are pooled by barcoded single-cell sequencing.
Create a sample sheet.
Please note that the columns in the tab separated file must be in the order shown below and does not contain a header line.
Column Description Name Sample name. BAM Location of the BAM file in the cloud (gs:// URL). Barcodes Location of the valid cellular barcodes file in the cloud (gs:// URL). VCF Location of the VCF file to use for this sample in the cloud (gs:// URL).
Upload your sample sheet to the workspace bucket.
gsutil cp /foo/bar/projects/sample_sheet.tsv gs://fc-e0000000/
Import demuxlet workflow to your workspace.
See the Terra documentation for adding a workflow. The workflow is under
Broad Methods Repositorywith the name “cumulus/demuxlet”.
Next, in the workflow page, click the
Export to Workspace...button, and select the workspace you want to export to in the drop-down menu.
In your workspace, open
Run workflow with inputs defined by file pathsas below
and click the
Please see the description of important inputs below.
|tsv_file||Four column tab-separated file without a header with name, coordinate sorted bam, barcodes, and vcf|
|min_MQ||Minimum mapping quality to consider (default 20)|
|alpha||Grid of alpha to search for (default [0.1, 0.2, 0.3, 0.4, 0.5]).|
|min_TD||Minimum distance to the tail (default 0)|
|tag_group||Tag representing readgroup or cell barcodes, in the case to partition the BAM file into multiple groups (default “CB”)|
|tag_UMI||Tag representing UMIs (default “UB”“)|
|field||FORMAT field to extract the genotype, likelihood, or posterior from (default “GT”)|
|geno_error||Offset of genotype error rate (default 0.1)|
The demuxlet output file contains the best guess of the sample identity, with detailed statistics to reach to the best guess.