Vgen23 is validated across germline, somatic, and ranking workflows. Performance claims are specific, reviewable, and tied to real evaluation cohorts.
You can only use Vgen23 if you can verify what it detects, how it ranks, and how it supports review. Vgen23 was built by the scientific team at Vgenomics, and our validation focused on detection accuracy, ranking performance, specificity, and review-readiness across real-world clinical categories.
Vgen23 was evaluated on a 120-patient cohort representing diverse clinical conditions and ethnic backgrounds, with 140 sites evaluated across true positive and true negative cases. The pipeline detected 122 of 123 true positive variants and correctly classified all 17 true negative sites, resulting in 99.2% sensitivity and 100% specificity. The cohort included neurological, ophthalmic, immunological, dermatological, skeletal, and congenital disorders.
120
Patients
140
Evaluated sites
99.2%
Sensitivity
100%
Specificity
122/123
True positives
0
False positives
In a somatic cohort of 96 lung cancer patients, Vgen23 evaluated 119,022 somatic variants and correctly identified 118,546 true positive variants, with 476 false negatives, resulting in 99.6% sensitivity.
96
Patients
119,022
Variants evaluated
118,546
True positives
99.6%
Sensitivity
This evaluation included 110 individuals with 118 mutations. Vgen23 ranked 98.3% of variants in the top 1 to 5.
110
Individuals
118
Mutations
98.3%
In Top 5
100%
In Top 10
Across 70 individuals and 73 variants, Vgen23 captured 91.8% of variants within the top 1 to 5 ranks, 97.3% within the top 1 to 10, and 100% within the top 1 to 20.
70
Individuals
73
Variants
91.8%
Top 1-5
97.3%
Top 1-10
100%
Top 1-20
Validation cohorts included samples from Indian, Chinese, Korean, Shanghai, Brazilian, Han, and Russian backgrounds, supporting strong performance across multiple genetic backgrounds.
Read the supporting documentation in our Trust Center, or upload your own solved or unsolved case to evaluate Vgen23 in your workflow.