1. Home
  2. VarSeq
  3. VarSeq 2.2.5 Known Issues and Work Arounds
  1. Home
  2. VarSeq Known Issues
  3. VarSeq 2.2.5 Known Issues and Work Arounds

VarSeq 2.2.5 Known Issues and Work Arounds

1. VarSeq Sometimes Crashes on “Project Open” using MobaXterm to Connect to Linux

  • Fix: VarSeq 2.2.5 SP1 and newer will protect against this bug in the underlying platform code that caused a crash when opening projects. Contact support if you would like access to this version.

VarSeq sometimes crashes when opening a project in one reported environment. Specifically, the crash occurs when using MobaXterm to connect to Linux from a specific desktop client. The nature of the crash is dependent on rare timing of events related to the displaying of the “splash” window.

Work around: Contact Golden Helix customer support at support@goldenhelix.com.

2. Importing Sample Data With a Sample Manifest Files Is Limited to 200 Lines

When using a sample manifest file to import additional sample data, VarSeq was limited to reading only 200 lines (samples). VarSeq would show the error below even though the vcf for the sample was in the manifest file. However, VarSeq was not able to find the sample name in the manifest file when the sample was beyond 200 lines.

Work around: Limit a single sample manifest to under 200 samples. VarSeq can optionally attach more than one sample manifest file on import.

3. When importing DRAGEN haploid CNV calls into VarSeq, duplication events in the male X chromosome were not imported correctly

The DRAGEN CNV caller calls duplication in X chromosome in males as haploid. In this context, a copy number “2” should have a CNV type of “Gain”. However, when these CNVs are imported into VarSeq, it was assigned the “normal” state usually associated with copy number 2 and not as a Gain. This logic was updated to have a different mapping of the numeric copy number state for haploid calls.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive VarSeq 2.2.5 SP1.

4. The Cancer Report Template V2 shipped with VarSeq does not support reporting variants of uncertain significance

The Variants of Unknown Significance section of the shipped Cancer Report Template V2 does not populate into the rendered report even when the Report As field in VSClinical is set to “VUS”.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive a fixed report template.

5. VSClinical displays errors when a CNV overlaps a gene that is not saved in the system or user gene preferences

The gene preferences file saves transcript, inheritance mode, disorder, and wild type settings for tumors (amp workflow) for genes. This information is set by default for some genes in the system genes preferences. Users can override and add new gene preferences to the file over time as new variants are interpreted in VSClinical. The errors below occur when a CNV overlaps a gene that is not saved in the system genes preferences or the user gene preferences, particularly when the inheritance mode is set to undefined.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive VarSeq 2.2.5 SP1.

6. The VarSeq annotation convert wizard crashes when transforming a field type to a listed field type

The VarSeq convert wizard in the “Manage Data Sources” dialog is crashing when changing a category field type to and array field type (i.e. string array or categorical array) from a unit array type.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive VarSeq 2.2.5 SP1.

7. Annotating two specific genes with the Ensembl gene track caused VarSeq to crash

There are two genes (TRAJ1 and GHD3-22) that have bad transcripts that cause VarSeq to crash when annotating with the transcripts using the Ensembl gene tracks for both GRCh37 and 38.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive a filtered version of this track which removes these genes from the Ensembl gene track preventing VarSeq from crashing when annotating with this source.

8. A regression introduced in 2.2.5 dropped small variants that had END= field set by bcftools

The bcftools merge command adds a “END=” field in the INFO field, even when it is not required. This field is generally used for CNV and for encoding other non-variant records into VCF files. Multi-base deletion variants were being dropped due to be being mistaken for CNV or gapped records. This issue is specific to this tool output, as we have not seen the END field used for small variants in any major variant caller output. The logic for detecting these features has been updated to keep these on variant import.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive VarSeq 2.2.5 SP1.

9. The autofill function for CNV findings in VSReports is causing VarSeq to crash

This issue only occurs for when using VSReports and is not an issue for users using the VSClinical word-based reporting system. The shipped CNV Report Template was crashing when attempting to detect primary CNV findings and autofill in the interpretations.

Work around: Contact Golden Helix customer support at support@goldenhelix.com to receive a fixed templateUtils.js file which fixes the issue causing the crash.

10. The ACMG Classifier algorithm forces the download and use of the latest version of ClinVar Transcript Counts as a dependency despite having a different version of ClinVar Variants already as a dependency

Normally, the ACMG Classifier algorithm checks the variant table for annotation sources that are being used in the project and uses the same sources as dependencies to run the algorithm. The ClinVar Transcript Counts source is a dependency of the ACMG Classifier but the algorithm is not checking the table for this source prior to computation. Thus, the newest version of ClinVar Transcript Counts is being used as the dependency regardless of the version of ClinVar Variants that is being used as another dependency of the ACMG Classifier. The ACMG Classifier algorithm should be matching the version used for these sources (i.e if the 03-03-2022 version of ClinVar Variants was used in the project, the ACMG Classifier algorithm should also use the 03-03-2022 version of ClinVar Transcript Counts even if a newer version is available).

Work around: From the Manage Data Sources dialog, download the matching version of ClinVar Transcript Counts source to the version ClinVar Variants that is in the project. Then, the Public Annotation sever can be temporarily deleted so the ACMG Classifier algorithm does not detect that a newer version of ClinVar Transcript Counts is available. To delete the Public Annotation server, right click on the Public Annotations server in the Manage Data Sources dialog and select delete. To add access back to the Public Server, click on the + icon in the top left hand corner, select Remote as the Location Type and choose Public Annotations as the Location.

Updated on May 13, 2022

Was this article helpful?

Related Articles

Leave a Comment