VSClinical and Reporting
- VSClinical now has the ability to customize the number of reports used by a project and to define Custom Reports that are capable of calling APIs, creating HTML, JSON or XML outputs and augment the data available to the Word reports. See the new section of our manual Customizing VSClinical Reports for more details.
- VSClinical now supports adding ref/ref variants (where the reference allele may be clinically significant) for evaluation with both the AMP and ACMG Guidelines.
- Word-based clinical reports now support displaying coverage statistics that include custom depth threshold target outputs in the coverage summary table.
- GenomeAsia has been added as a frequency track that can be easily added as a population source within VSClinical.
- ClinGen Expert Curated and TP53 ACMG R20 can now be used in VSClinical to recommend and score ACMG criteria. See the ACMG Previously Interpreted Variant Sources tab of the VSClinical configuration dialog. They are series specific so recommended criteria will update when new versions of these sources become available.
- As an enhancement, the sidebar listing of currently scored criteria can now be clicked on to navigate to the relevant question for the ACMG and AMP workflows.
- In the Phenotypes tab clinical notes section, MONDO IDs and ORPHA IDs are now looked up in addition to OMIM IDs when present in the clinical notes.
- When changing the interpreted transcript for a variant added to an evaluation, a confirmation message now pops up and then then all variants and CNVs in the evaluation for that gene are switched and rescored.
- When scoring common CNVs, extra checks have been added so that you don’t get into a place where both section 2 and section 4 criteria are both scored, to say in accordance with the ClinGen CNV Guidelines.
- You are now able to save an empty interpretation for a CNV in the ACMG CNV workflow. This was previously not allowed and would result in an error if attempted.
- More exonic missense and synonymous variants will have novel splice site predictions applied. Previously exonic variants were not considered for splice site predictions. This may result in a change of recommended criteria in the “Computation Evidence” section. Specifically, BP7 will not be recommended for synonymous variants that have at least two novel splice prediction algorithms detecting a new acceptor splice site. Similarly, for missense variants, PP3 may be recommend instead of BP4 when at least 3 novel splice algorithms predict a novel acceptor splice site is introduced.
- In the rare case where the canonical splice site goes from not being predicted to existing with the mutation (i.e. the reference sequence did not contain the acceptor or donor site GT-AG canonical pairs), we do not recommend PVS1 and classify the variant as a splice region variant.
- When adding a variant previously saved with a PM4 classification, the criteria would show up twice in the ACMG interface. There are two scenarios in which a PM4 criteria can be applied: inframes and stop losses. PM4 was being applied for both scenarios. It now selects only the contextually relevant criteria for the added variant.
- When a variant overlaps two genes and the Transcript Dialog was used to switch the interpretation to a different gene, the recommendation text would show the previous gene name and HGVS. This has now been fixed to re-score the evaluation and update the recommended scoring criteria and text for the new gene.
- Missense Z-Score from gnomAD Gene Constraint was showing the default gene transcript Z-score, even when evaluation was switched to interpret a non-default transcript for a variant. The recommendation and display now stay in sync with the selected transcript.
- Externally imported CNVs would fail to be added to VSClinical if it had a “Flags” field and that field was not an array type. Detection of this edge case now allows these CNVs to be added.
- When switching which Overlapping Transcript a CNV was being scored on, the auto-recommended criteria and descriptions of the CNV location and impact was not being updated. The CNV is now re-scored when switching to a non-default transcript.
- The Cancer Ontologies track was updated to include more cancer diverse ontologies. In addition, biomarkers that impact multiple tumor types can now be added to AMP evaluations as “Solid Tumor” or “Hematologic Tumor Cancers”.
- VSClinical now supports broadened interpretation scopes; Activating mutations, Loss mutations and gene fusions regardless of fusion partner.
- Drug names that were being added to the clinical evidence section of the AMP workflow have been normalized and synchronized to account for differing capitalization.
- Clinical trials in the AMP workflow can now be searched by any of the provided fields alone, drug name, biomarker, or disease.
- For variants with allele frequency information brought in from the table, the display of the allele frequencies in the table and the report now more intuitively match the VAF calculation (AltReadCount/TotalReadCount).
- When adding biomarkers to the AMP workflow, the c. and single letter p. HGVS notations now display when hovering over the added variant.
- Re-scoring an empty AMP evaluation was causing the VSClinical UI to hang. This has now been fixed.
- When opening the “Review and Save” dialog for one of the Clinical Evidence sections of a Biomarker, previously saved interpretations did not pull up in the history view (right-hand side of the save dialog).
- The description of CNVs in AMP workflow report data would always describe the CNV as an amplification even when the CNV was a deletion.
- A multi-exon CNV added to the evaluation would have its location described by only its first exon, but has now been fixed to describe the range of exons it overlaps
VarSeq Annotations and Algorithms
- The new Gene Panels manager can be accessed from Tools > Manage Gene Panels… that supports creating and updating gene panels based on a custom gene list, a list of phenotypes or from one of the curated UK Panel App predefined gene lists. See the Manage Gene Panels Dialog section of the manual for more details.
- The PhoRank Gene Ranking feature has a new default mode called PhoRank Clinical. This algorithm uses semantic similarity for phenotypic prioritization and demonstrates a direct clinical relationship between phenotypes and genes. PhoRank Research is the previously available version of PhoRank and uses ontology propagation for phenotypic prioritization and is still available for use in VarSeq projects. Existing projects with PhoRank will continue to use PhoRank Research, but can be switched by right-clicking on the source and choosing “Edit” to change the algorithm options.
- When inputting terms into the PhoRank dialog as HPO terms (such as HP:0001637), you can now see the detected HPO term name by hovering over the input before running the algorithm.
- The PhoRank Gene Ranking algorithm is now available at a per sample level for CNV analysis. When multiple samples are added to a single project, individual sample phenotypes can be assigned in the CNV PhoRank Gene Ranking algorithm dialog. Previously CNV PhoRank Gene Ranking was only available at the project/cohort level.
- A Match Panels (Per Sample) and Match Panels algorithms has been added to incorporate gene panels from the Manage Gene Panels dialog.
- A Match String List algorithm has been added to VarSeq to match any field (not specifically gene names) to a list of specific values.
- A Match Genes Linked to Disorders algorithm has been added to VarSeq to list genes associated with user specified disorders.
- The System Gene Preferences file has been updated to update default transcripts based on the latest ClinVar submission counts. Any gene with a greater than 400 submitted assessments from ClinVar will have a default transcript set in the System Gene Preferences. It also incorporates the latest OMIM gene inheritance information to update genes as Dominant or Recessive. See the Gene Preferences manual section for more details.
- The Mendel Error algorithm output fields now are grouped with the imported sample fields when added to a family project.
- An option for the RefSeq Overlapping Genes algorithm has been added to use inexact breakpoints for the c. HGVS for CNVs.
- The CNV Caller can now optionally run at all sensitivity and precisions levels at once in the Advanced Options. This creates an additional Precision Level field that reports for every CNV what minimum sensitivity level it would have been detected at. This feature can be used to help establish the right default value to use in a test production environment.
- Editing the CNV Targeted Caller algorithm on older projects would not support adding the GC correction option, because the required supporting files were not originally available. This type of edit is now supported and the required reference sequence will be detected.
- The LOH caller has improved handling of non-standard variants such as multi-allelic variants. To avoid double-counting and not influence the calling process, multi-allelic variants are now ignored when calling Loss of Heterozygosity.
- The CNV ACMG Auto-Classifier has improved performance when using a VSWarehouse-based gene dosage catalog by batching queries of many CNVs to the warehouse for annotation. This can result in a significant speedup of the algorithm when a table includes many CNVs or large CNVs.
- Importing external CNVs with hemizygous males with a copy number of 1 were incorrectly being identified as heterozygous deletions on import. We now factor in the genotype to detect the correct Diploid state.
VarSeq Projects and General
- The built-in downloader for VarSeq has been updated to support more firewall and proxy settings, improve resuming of downloading when network connections are lost, and better detection of in-progress downloads started by another VarSeq instance on the same computer or accessing the same the same shared network drive for annotations.
- From the filter chain, you can “Add New Gene Panel Filter” from the right-click menu that enables filtering Variants that are in one or more Gene Panels defined in the new Manage Gene Panels dialog.
- In the filter chain, toggling the check-mark “enabled” state for containers no longer propagates that state to individually enabled or disabled the contained filter cards and containers. This allows for disabling a container, but not resetting the enabled state of a nested containers filters. The previous behavior can still be accessed by using the new right-click menus “Enable all Children/Descendants” and “Disable all Children/Descendants” on containers.
- The novel splice prediction predictions run during transcript annotation now predict and report exonic novel acceptor splice sites, but with an “Unknown Effect” as it is unknown whether the wild-type or novel splice site would be preferred without functional studies. Previously, variants in this context would not receive the novel splice annotations.
- VSWarehouse-based assessment catalogs can now be exported from the assessment catalog view in VarSeq. This is helpful when migrating catalogs between warehouse instances or creating a copy or backup of a catalog.
- If the length of files created under VarSeq project’s data folder is a concern (such as when the 260 windows character limit has been reached), “shortPathMode”: true can be added to the hosts.json system defaults file and VarSeq will generate data files with incrementing numbers as file names.
- The VarSeq start screen news URL can be customized from the hosts.json system defaults file as well.
- The Help > About VarSeq screen has a new Device Info button to display manufacturing and regulatory details as well as updated list of third-party libraries used under Credits.
- The variant table on Linux now has alternating row colors.
- The Secure Annotations data repo was failing to load for some users behind strict firewall and network proxy rules due to client-side SSL settings. Updates were made to use SNI and only secure protocols, which now supports loading annotations in these environments.
- The global Tools-> Open Folder menu has been expanded to have the location of other commonly accessed folders, such as Liftover Chain Files and Report Templates. The VarSeq Program Folder was renamed to the VarSeq Installation for clarity.
- VarSeq projects now remember the locked and unlocked state of specific annotations from the template used to create it, making it easier to update project templates.
- When the source coordinates from the VCF cannot be detected, an updated message informs the user to check the VCF file coordinates match the genome assembly of the project.
- The ACMG Gene Panel and ACMG Trio templates have been updated to include the ACMG secondary findings v3.0, and to use the ClinGen Expert Curated Interpretations for new VSClinical evaluations
- The Compute Expression feature used on import, export and through the Compute Fields algorithm has added compliment, union and intersection functions to aid in comparing values in array fields (fields containing a list of values for each record).
- When an update of the software is initiated, but the user does not have write permissions to the software install location, a informative error message immediately lets the user know the update requires write permissions to proceed.
- VarSeq 2.2.1 or 2.2.2 based project templates that included an ACMG classifier algorithm were not correctly scoring PVS1 or BS2 when used to create projects in VarSeq 2.2.3. See VarSeq 2.2.3 Known Issues and Work Arounds for more details.
- Opening VarSeq on windows when it is installed on a network drive would result in VSClinical views being blank due to a web engine security setting change. The workaround was previously to run from a local computer, but VarSeq can now run fully from a network drive location.
- The variant import algorithm has improved handling of normalizing multi-allelic insertions/deletions to correctly track inheritance. In trio projects, the inheritance of variants that were in a multi-allelic state in one parent and present in a child with a different variant representation in the child was not being detected. The variant normalization and genotypes are now filled in correctly, so the Mendel inheritance algorithm reports variant inheritance correctly for these edge cases.
- For specific input data and platforms, the ACMG CNV Classifier algorithm would crash as it was cleaning up on completion, this has now been fixed.
- When VarSeq was running on a network drive, updates to QtWebEngine security was causing the internal web browser and VSClinical to have issues loading. Now, we support running from network drives.
- Custom track titles set for project-based annotation sources plotted in GenomeBrowse are now preserved when saved into VarSeq project templates.
- BAM file coverage for the current sample was not plotting when switching to a new sample. This has now been fixed.
- VSPipeline was not able to import VCF files if the VCF file did not specify a genome assembly. This has been fixed and the VCF file is imported based on the genome assembly defined by the project template.
- The VSPipeline import command now has a “liftover=true” option that auto-detects and applies a liftover from the detected VCF file assembly to the project assembly.
- VSPipeline multi-sample CNV VCF import caused a segmentation fault. This issue has been fixed.
- When importing CNVs using VSPipeline, when none of samples in the VCF matched the project samples, the error message was uninformative and would cause VSPipeline to hang. The error message has been updated and VSPipeline no longer hangs.