Version 13.15.0 of VarSome, VarSome API, and VarSome Clinical will be released on 14 March 2026.

Summary of Key New Features

• RNA fusion annotation from TSV and CSV files
• Integration of SpliceVarDB
• Structural Variant interpretation enhanced with overlap percentage against known SV datasets
• New OncoKB-based filters for oncogenicity, function, and levels of evidence
• Improved tumor type matching using OncoTree ontology
  
  

New Features

RNA Fusion Annotation from TSV/CSV Files

VarSome Clinical now supports RNA fusion annotation from TSV or CSV files generated by widely used fusion callers such as STAR-Fusion, FusionCatcher, Arriba, DRAGEN, and Archer.

Somatic samples can be defined from DNA, RNA, or combined DNA + RNA inputs, allowing laboratories to run flexible analysis workflows regardless of their sequencing setup.

Detected fusion events are automatically annotated and displayed in a dedicated Fusion table, where users can:

• Visualize fusions when breakpoints are available
• Inspect PFAM protein domains involved in the fusion event
• Review tier classification based on JAX-CKB
• Access consolidated biological and clinical relevance from JAX-CKB, OncoKB, and COSMIC
  

This automated fusion annotation pipeline reduces manual review effort, shortens interpretation time, and helps ensure that clinically actionable fusion events are not overlooked. By consolidating biological and therapeutic evidence into a single interface, it enables faster identification of relevant treatment options and strengthens VarSome Clinical’s capabilities for precision oncology workflows.

Note: RNA fusion annotation from TSV/CSV inputs is currently available for the somatic pipeline.

SpliceVarDB Integration

VarSome Premium and VarSome Clinical now integrate SpliceVarDB, bringing experimentally validated splicing evidence directly into variant analysis workflows.

This integration provides curated validation data for more than 50,000 human variants across hg19 and hg38, enabling users to evaluate splice-altering variants using experimental evidence rather than relying solely on in silico predictions.

When a variant is selected, users can review a structured summary of its splicing impact and explore detailed experimental evidence, including links to the primary literature. By presenting transparent and traceable data at the point of interpretation, this feature supports more confident clinical reporting and improved decision-making in both diagnostic and translational research workflows.

SpliceVarDB annotations are available exclusively to VarSome Premium and VarSome Clinical users and apply to newly created analyses.

Structural Variant Overlap with Known SV Databases

VarSome Clinical now displays the percentage overlap between detected structural variants and previously reported SVs in a new column within the SV table.

Known SVs are retrieved from multiple databases, including ClinVar, DECIPHER, DGV, dbVar, the Children’s Mercy Research Institute dataset, the Ira Hall Lab dataset, HPRC, and estd20.

For each variant, the system reports the highest overlap percentage with previously reported variants of the same CNV type (deletions or duplications), and enables interpretation based on the inferred pathogenicity derived from genomic overlap, including gene and exon-level overlap. Users can also navigate directly to the corresponding Known SV entry to explore supporting evidence and additional overlapping variants.

This functionality supports rapid interpretation of CNVs by providing quantitative overlap metrics and inferred pathogenicity directly within the analysis interface. Users can also filter SVs based on overlap thresholds and inferred pathogenicity, significantly reducing interpretation time and streamlining diagnostic workflows.

This feature is available for CNV analyses from WGS, WES, and tertiary SV analysis from VCF files.

OncoKB Dynamic Filters

Users can now filter variants using OncoKB-specific criteria, including oncogenicity classification, biological function, and levels of clinical evidence. These new filters enable faster identification of clinically relevant and actionable variants during somatic analysis. By reducing interpretation noise and prioritizing high-value findings, this improvement supports more efficient clinical review and evidence-based decision-making.

Minor Fixes and Additions

Improved Tumor Type Matching with OncoTree

Tumor type matching has been improved by adopting OncoTree as the unified reference ontology for cancer classification. The updated system reconciles synonymous tumor terms across multiple curated knowledge bases, ensuring that relevant clinical evidence is not missed due to terminology differences. The interface now guides users toward standardized tumor-type selections, improving consistency, increasing somatic classification accuracy, and strengthening confidence in clinically actionable results.

Updated CNV Annotation Threshold

The ACMG CNV annotation threshold used during CNV classification has been adjusted from 85% to greater than 70%, resulting in a slight improvement in classification accuracy without negatively affecting other performance metrics.

User Role Access Restriction

Users assigned the “No Role” designation are now restricted from accessing the User Management page. This prevents unauthorized role assignments within groups and safeguards against unintended changes to existing user permissions.

Delly CNV Caller Upgrade

The Delly CNV caller used for WGS single sample and tumor-normal analyses has been upgraded to its latest version. This update improves stability, reduces CNV analysis runtime, and maintains compatibility with the latest pipeline updates.

Support

We hope you find these improvements helpful. We would love to hear any feedback and suggestions you may have. Support is available as usual from support@varsome.com.

The VarSome Team.