Fig. 5
From: Benchmarking datasets for assembly-based variant calling using high-fidelity long reads
Performance of HiFi read- or assembly-based variant calling at varying sequencing depths. A–H Violin plots representing the detection rate of true-positive variants. We used the full-matched common founder variants in ALT1 and ALT2 described above (1804 in total) as the true-positive variant set to estimate the true-positive variant detection rate. Red represents the detection rate based on at least partial matching to the true-positive variant set and blue represents the detection rate based on perfect matching. I–K Violin plots representing the number of total detected variants. Red represents the number of variants at least partially matched to the common founder variants (1804 in total) and blue represents the detection rate based on perfect matching. Gray represents the remaining variants. Reliable variants were filtered from those obtained using read-based variant calling according to their quality metric (Q ≥ 5; Q5). Subsampling was repeated five times for each sequencing depth. Each horizontal solid line of the violin plot represents the median of five repeats