Smart Variant Filtering
The variant filtering process consists of selecting highly confident variants and removing the ones that are falsely called. This process used to be mostly left out from deeper testing in secondary genomic DNA analyses, even though it can boost precision of variant calls significantly. After developing graph genome alignment tool and reassembly variant caller, we are now launching the Smart Variant Filtering tool used for filtering germline variants.
Smart Variant Filtering (SVF) uses machine learning algorithms trained on features from the existing Genome In A Bottle (GIAB) variant-called samples (HG001-HG005) to perform variant filtering (classification). The comparison results obtained during deep, three-stage testing demonstrate that it outperforms the solutions currently used within most secondary DNA analyses. Smart Variant Filtering increases the precision of called SNVs (removes false positives) by up to 0.2% points while keeping the overall f-score higher by 0.12-0.27% points than in existing solutions.
Learn more about the Smart Variant Filtering public project from our documentation.