Plasma Whole Genome Sequencing (pWGS) Reports

Djerba INI Configuration

To create a pWGS report, we run Djerba with two pWGS plugins enabled:

  • pwgs.sample performs sample identification and QC checks,

  • pwgs.analysis performs analysis of mrdetect workflow outputs

Information is obtained either from one of two Data Sources: the Requisition (Req) system or Dimsum.

The pwgs plugins require the following information to be populated into the .ini file:

Plugin

Parameter

Source

Description

Example

Core

author

bambooHR

Your name

Rosalind Franklin

pwgs_cardea_helper

requisition_id

Dimsum

The id of the requisition

PWGVAL_011418_Ct

pwgs.case_overview

primary_cancer

requisition system

Primary cancer in req system

Pancreatic Adenocarcinoma

wgs_report_id

requisition system

associated WG(T)S report

OCT-01-1328_Ut_P-v2

requisition approved

requisition system

date of approval for the req

2023/10/10

supplement.body

assay

Dimsum

assay short name

PWGS

pwgs_provenance_helper

No input parameters required

pwgs.sample

No input parameters required

pwgs.summary

No input parameters required

pwgs.analysis

No input parameters required

Interim Report Generation

The Djerba report html can be created with the command:

$ djerba.py report --ini config.ini --out-dir report

The report html will be of the name ${group_id)_report.clinical.html.

Review Sample Quality

The sample QC section has three metrics to be reviewed by CGI staff: median insert size, mean genome-wide coverage and primary SNVs, as shown in screenshot below:

Coverage

Mean deduplicate coverage should be ≥30X (40X target). Mean deduplicated coverage is automatically pulled into Djerba from QC-ETL. Samples under 30X deduplicated should be topped up by TGL.

Primary SNVs

The plasma whole genome sequencing assay assumes the WGS or WGTS (40X or 80X) assay has been completed (hereafter, this family of assays are referred to collectively as ‘WGS’).

The WGS report must be completed and the Djerba working directory be accessible on the cluster.

Assay calibration showed a minimum requirement of 4,000 candidate SNVs in WGS report for reliable and replicable results; the pWGS assay should not be run when the primary tumour has less than 4,000 SNVs as per the current validation report. Assay requisitioners have been informed of this requirement and that number is included in WGS reports; cases where the 4,000 SNV threshold is not met will be failed (see Failed pWGS Reports).

Additionally, cases will also be failed if tumor purity is below the 30% threshold, even if the SNV count exceeds 4,000.

Median Insert Size

We set an upper limit to the median insert size (see QM-24: Quality Control and Calibration for cutoff value). cfDNA samples have a distinct insert size distribution that can be used to distinguish them from either normal (buffy coat) or primary tumour tissue samples with DNA fragmented by sonication. Median insert size can therefore be used to detect a swap between (for example) cfDNA and buffy coat. Libraries where the median insert size is higher than the threshold should most commonly be failed, however, if the median insert size value for a case is just slightly above the threshold, the insert size distribution can be visualized to confirm how much higher the median is than 167 bp and the sample rescued manually. Insert size distribution can be found in insert_size_distribution.svg and the figure below is to be used to guide decisions:

Table 16 Insert Size distributions
../_images/isize1.png ../_images/isize2.png

PASS

FAIL

The distribution is very narrow and centered at 167 bp with a second very low peak at 334 (167 x 2).

The distribution is wide, centered near 250 bp and has some fragments larger than 500 bp.

Updating QCs

Sign off on the “Informatics Review” in Dimsum according to the QM-036. Quality Control Approval Procedure.

Generate Draft Report

If the report passes all QC metrics and all information is present, the report is converted from html to pdf, either using manually or using Djerba:

$ djerba.py render --json report/${group_id)report.json --out-dir report --pdf

Continue to Review the Draft Report ➡️

Change Log

Github commit log