Figure 1
From: eVIDENCE: a practical variant filtering for low-frequency variants detection in cell-free DNA
(a) Summary of the eVIDENCE pipeline. An input BAM file is converted to the BAM file with consensus alignment pairs using Connor. Candidate variants are called using the converted BAM file. eVIDENCE removes unique molecular tag (UMT) and stem sequences from a raw BAM file and creates new FASTQ files while retaining the UMT information. These FASTQ files are converted into a new BAM file and for each candidate variant, eVIDENCE performs filtering using the new BAM file. (b) Number of detected single nucleotide variants (SNVs) and insertions and deletions (indels) from cell-free DNA (cfDNA) sequencing data processed by Connor (left) and after applying eVIDENCE (right). The expected number of SNVs and indels are indicated by blue and red dotted line (70 and 5, respectively). (c) Number of detected variants among 6 cfDNA samples in which matched tumor sequencing data were available. An initial variant calling using the processed data by Connor detected 11806 variants containing 14 tumor variants (left). After applying eVIDENCE, a large number of candidate variants were discarded, but 12 tumor variants remained (right), showing that eVIDENCE efficiently filtered candidate variants.
