Distributed Whole Genome Sequencing - Special Considerations¶
The Distributed Whole Genome Sequencing (dWGS) programme is a major transformation initiative led by NHS England and Genomics England to transition whole genome sequencing from a single centralised provider to a distributed network. The programme enables sequencing to be performed locally within commissioned NHS sequencing laboratories while maintaining a centralised bioinformatics, interpretation, quality assurance, and governance framework.
Identification of dWGS Cases¶
Cases that have been submitted to the centralised and distributed programmes can be differentiated by following sample identifiers.
| Source | Format | Example |
|---|---|---|
| Distributed | s_[a-z0-9]{26} |
s_03kjhfjkh23j23nmm34nedskp |
| Centralised | LP[0-9]{7}-DNA_[A01-H12] |
LP5000001-DNA_A01 |
Small variant artefacts¶
In the distributed sequencing programme, sequencing will be performed on the NovaSeq X Plus, moving from the NovaSeq 6000 used for the centralised programme. As with all sequencing technologies, we expect to see some technical artefacts occur. This effect is mitigated by the use of internal aggregate allele frequencies generated from the same sequencing technology and used to filter both variants with high presence in the population and technical artefacts.
We are aware of highly recurrent sites that may be prioritised in tiering when they are called:
- chrM:12329:T>G
- chrM:12335:T>G
These variants tend to occur on the same read and are accompanied by noise on the left-hand side and typically at a low allele fraction (<0.2). If these variants occur in your case and you believe them to be causative, we would encourage confirmatory testing to verify that they are true events.
We will continue to monitor cases for artefacts and would be grateful for any reports of possible artefacts.
STR calls with LowDepth VCF filter¶
In some scenarios DRAGEN/ExpansionHunter will call a STR with a LowDepth VCF filter (non-PASS). When this occurs for an STR locus that is green on an applied panel it is excluded from the report displayed in the DSS. While this behaviour occurs in both centralised and distributed sequencing cases, we have observed an increased but still low number of LowDepth STR calls in the dWGS cases.
We will be updating the DSS to highlight STR loci on applied panels that have been filtered as LowDepth in an upcoming release. As a precaution, Genomics England will also be reaching out to users to highlight any cases that are affected until this change is implemented.
Incorrect chrX ploidy used for CNV calling¶
In rare circumstances, participants with an XY karyotype may have an XX ploidy applied for CNV calling. We expect this to only occur where a participant has a somatic loss of chrY. If this issue occurs, the majority of chrX will be called as loss CNVs with a copy number of 1.
We will address the root cause in an upcoming release. As a precaution, Genomics England will be monitoring and communicating with users to highlight any cases that are affected until pipeline changes are implemented.
Note: this issue only affects samples that have been submitted through the decentralised model; samples from the centralised pathway are unaffected.