Fig. 4
From: Merqury: reference-free quality, completeness, and phasing assessment for genome assemblies
Merqury hap-mer plots for evaluating haplotype phasing. a Example of phase blocks and switches. Blue and red bars are paternal or maternal hap-mers found in the assembly. A phase block is defined by at least two hap-mers (markers) from the same haplotype. Short-range switches are allowed in between markers, in defined ranges. Two consecutive red markers within a certain range are marked as short-range switches and counted for switch errors in block 1. As the red markers are consecutively found over a certain range, or in numbers above a certain threshold, a separate block is formed. Each switch between blocks is counted as a long-range switch. b Phase block statistics of the haploid assemblies with switch errors, allowing at most 100 switches within 20 kbp. c Hap-mer blob plot of the TrioCanu assembly. Red blobs represent Col haplotype contigs, while blue blobs are the Cvi haplotype. Blob size is proportional to contig size, and each blob/contig is plotted according to the number of contained Col (x values) and Cvi (y value) hap-mers. Col-specific k-mers are found in the Col assembly with almost no Cvi-specific k-mers, while Cvi k-mers are found in the Cvi assembly with almost no Col k-mers. d, e Blob plots for FALCON-Unzip and Canu assemblies show that most contigs are a mix of sequences from both haplotypes, but FALCON-Unzip preserves phase within its alternate contigs, as designed. f Phase block NG* plots of the haplotype resolved Col (left) and Cvi (right) assembly, sorted by size. X-axis is the percentage of the genome size (*) covered by phase blocks of this size or larger (Y-axis). Blocks from the wrong haplotype are very small and almost entirely absent. g, h Phase block NG* plot of the g FALCON-Unzip and h Canu assemblies. Col and Cvi phase blocks are distributed evenly, as is typical for pseudo-haplotype assemblies. i–k Phase block and contig NG* plots show the relative continuity of i TrioCanu, j FALCON-Unzip, and k Canu assemblies. Phase block sizes are similar to the contig sizes in i. Phase blocks are much shorter than the contigs in k, because of the frequent haplotype switches in the contigs