Fig. 3

Query neighborhoods in HuSB1 metagenome are large and contain additional marker genes. a Left panel: neighborhood sizes are larger in HuSB1 than in podarV. Here, we queried podarV and HuSB1 using each of 51 and 23 genomes fully present in the respective data sets and measured the relative size of its neighborhood—a size of 1 indicates that no additional sequence is present in the neighborhood, while a size of 2 indicates that the retrieved neighborhood is twice the size of the query genome. b Right panel: query neighborhoods are estimated to be more complete than the original genome bins. We queried HuSB1 using each of 23 genomes binned from SB1 and assembled the resulting neighborhoods using MEGAHIT and Plass. The blue points represent completeness estimates of MEGAHIT-assembled neighborhoods, while green and pink bars represent the additional or missing content in the Plass assemblies, respectively