VarBin, a novel method for classifying true and false positive variants in NGS data.

Area of Science:

• Genomics
• Bioinformatics
• Genetic disease research

Background:

• Rare genetic disease variant discovery requires screening millions of variants.
• Sequencing errors generate false positive variants, complicating diagnosis.
• Existing methods for false positive reduction are often insufficient.

Purpose of the Study:

• To introduce VarBin, a novel method for prioritizing genetic variants.
• To predict and reduce false positive variants in rare disease diagnostics.
• To improve the accuracy of variant identification in whole exome and genome sequencing.

Main Methods:

• VarBin utilizes the Genome Analysis Toolkit to calculate Phred-scaled, likelihood-ratio by depth (PLRD) values.
• PLRD values are used to segregate variants into four bins (Bin 1-4) based on false positive likelihood.
• Variant PLRD values are compared against background samples to assign classifications.

Main Results:

• Sanger sequencing confirmed 97% of Bin 1 variants as true positives.
• Bin 3 and Bin 4 variants were exclusively false positives.
• Bin 2 contained a mix of true and false positive variants, indicating an area for further refinement.

Conclusions:

• VarBin effectively classifies the majority of true variants into Bin 1.
• Bins 3 and 4 accurately identify false positive variants.
• Further development is needed to resolve variants within the "uncertain" Bin 2.