Data · dataset · 2026

<p>Recall and false discovery rates for variants at 5%, 10%, 15%, and 19% frequencies across Phred score scalar adjustments. Plasmid truth sets normalised to 2,000 × depth (n = 12 subsampling replicates per condition; df = 11, t-critical = 2.201). The Control row is a technical duplicate of the Uncalibrated condition, confirmed by identical VCF hash signatures. Phred score reduction removes all false positives at −8 on both chemistries. Variants at 15% AF and below are adversely affected by Phred reduction on R9.4.1 flow cells, while only 5% AF variants are affected on R10.4.1 flow cells. CI = 95% confidence interval; FDR = false discovery rate.</p>

Listed in NCL Data

<p>Recall and false discovery rates for variants at 5%, 10%, 15%, and 19% frequencies across Phred score scalar adjustments.