VarSome Clinical offers two types of filtering:  algorithmic filters, which run during analysis to prioritize variants, and dynamic filters, which give you flexible control over how results are viewed and explored. Together, they help reduce noise and deliver relevant variants in clinical or research workflows. This article provides only brief descriptions of the filters and features in VarSome Clinical. For more detailed information, please visit our Help Center.

Algorithmic Filters: Prebuilt, Purpose-Driven Tools

VarSome Clinical has a range of algorithmic filters designed to tackle common but complex genetic analyses. They streamline variant prioritization by applying logic tailored to specific clinical scenarios. Our algorithmic filters, combined with our deep variant database, allow us to address many of the most challenging analyses directly. If you don’t see your specific use case here, our team can design a custom filter to match your needs. For more detailed information on all our algorithmic filters, click here. 

VarSome Picks

VarSome Picks, or phenotype-driven variant prioritization, uses AI to rank variants in a disease-specific context using phenotype, gene, and variant data. It prioritizes variants in the top candidate genes linked to user-provided phenotypes and matches the inheritance mode. Variants marked as artifacts are excluded. VarSome Picks helps spotlight likely causative variants quickly in single-sample, trio, or gene-list workflows. For more information on VarSome Picks, click here.

Trio Recessive

This filter finds rare recessive variants in a child when both parents are unaffected. It removes variants homozygous in either parent or located on mitochondrial or Y chromosomes. It offers filters to include strong VUS only, variants missing from one parent, or set a population frequency cutoff.

Family Trio Recessive (coding & rare)

This filter focuses on rare coding or splice-site variants in a trio. It removes variants with low allelic balance or low coverage, excludes those common in the population or homozygous in parents, excludes mitochondrial/Y-linked variants, and keeps homozygous child variants or compound heterozygous candidates. Thresholds for frequency, balance, and coverage are adjustable.

Compound Heterozygous Candidates

Finds genes with multiple heterozygous variants that may be causative. Supports filtering by ACMG rule strength and phasing, and excludes genes with homozygous pathogenic variants. 

Compound Heterozygous for Trios (n=3)

Designed for trio analysis. Identifies compound heterozygous variants in the child while excluding genes where parents carry the same variant pairs. Includes options for phasing and filtering by strong VUS. 

De novo (strict)

Identifies variants in the child that are absent in both parents. To be included, neither parent can have any reads supporting the variant, and coverage must be at least 8 at that position. 

De novo candidates (naive)

Includes variants likely to have arisen in the child from unaffected parents. A variant qualifies if the child is homozygous and only one parent carries it, or if it’s present in the child but not detected in either parent. We recommend you further filter for pathogenicity, coverage, frequency, and mode of inheritance.

ACMG Actionable Genes

Identifies pathogenic or likely pathogenic variants in ACMG SF v3.2 genes. For autosomal recessive genes, it includes only homozygous or double heterozygous variants. Useful when screening for secondary findings with established clinical relevance.

Exonic & Splicing algorithmic filter

Focuses on coding regions, UTRs, non-coding exons, and splicing variants within 10 base pairs of a splice site. This filter is applied automatically to very large analyses to help keep results manageable.

Carrier Risk for Couples

Highlights shared heterozygous pathogenic or VUS genomic variants in healthy couples, including X-linked and compound heterozygous combinations.

Genes in Common / Variants in Common

Used in cohort analysis. “Genes in Common” returns all variants found in shared genes across samples, while “Variants in Common” returns specific shared variants. Useful for identifying potential common drivers in a group.

Max Other Samples

Excludes variants seen in too many samples or with a population frequency over 5%. For example, helps remove likely benign variants in large datasets.

PharmGKB / GWAS Catalog / Imprinted Genes

Filters variants with known links to drug response (PharmGKB), GWAS traits, or imprinted genes. Relevant in pharmacogenomics, complex traits, or when considering parent-of-origin effects in rare diseases.

Segregating Variants

Finds dominant, recessive, or compound heterozygous variants that segregate in trios or cohorts. Configurable to suit the study design. For example, used to highlight variants consistent with expected inheritance patterns.

These filters offer a balance of speed and specificity, reducing the need for manual curation in routine workflows.

Dynamic Filters: Custom Control for Specific Questions

Dynamic filters give you full control to define, save, and apply your filtering logic. They adapt to changing needs in research or diagnostics. For more detailed information on dynamic filters, click here.

Creating Custom Filter Sets

Combine multiple filtering criteria into reusable sets that can be applied across analyses. Filters can be reordered, duplicated, or edited to match changing project needs.

VCF-Based Filtering

Filters variants using the values in the VCF’s FILTER field, such as “PASS”, “FAIL”, or caller-specific tags like those from Mutect2. Useful for including or excluding variants based on external quality flags.

Call Status Filtering in Trio Analyses

Handles missing genotype data in trios, such as uncalled positions marked “./.”. Helps retain potentially valid variants that might otherwise be discarded due to incomplete calls.

Shared Filters

Filter sets can be shared across an organization, allowing teams to standardize analysis and reduce duplicated work. Once a filter is shared, it appears alongside personal filters in both the Filters screen and the analysis side panel, with the creator's name clearly displayed. 

Any group member can apply a shared filter directly to an analysis. Only the original creator can edit the shared version, but others can duplicate it and make their own changes. This gives teams a balance between consistency and flexibility.

It’s important to note that shared filter sets are composed of multiple filter instances (for example, filtering only pathogenic variants). If the creator changes the status of one of these instances by disabling it, for example, the results seen by all other uses of that filter set will change accordingly. This means filters remain dynamic and dependent on the creator’s choices. To avoid unexpected changes in your results, users can duplicate the shared filter set so they have full control over which filter instances are active.

Integration with Gene Lists

Filters can be combined with gene lists, created manually, modified from existing ones, or generated automatically from phenotype data. Phenotype-based lists include both directly associated and disease-linked genes.

Together, algorithmic and dynamic filters give you the flexibility and precision to focus on the variants that matter most. Whether you’re screening a trio or analysing a large cohort, these tools help streamline your workflow without sacrificing control. For more information on applying filters to your analyses, please visit our Help Center.