Fig. 3From: Illumina error correction near highly repetitive DNA regions improves de novo genome assemblyThe final step of the BrownieCorrector pipeline: (1) de Bruijn Graph is built from the uncorrected reads in a cluster. Uncorrected reads contain sequencing errors which result in the appearance of erroneous k-mers and subsequently erroneous nodes/arcs in the graph; (2) erroneous nodes (colored in red) are detected and removed from the graph based on coverage and graph topology. Such erroneous nodes often appear as tips or bubbles; (3) reads are aligned individually to the corrected graph and mismatches and indels in the reads are detected and fixed with the correct path in the graphBack to article page