Fig. 1: Assembly strategy for high-quality microbiome references.

a Boxplots showing the number of MAGs obtained across metagenomic datasets using short-read and hybrid assemblies (n = 109). b Stacked barchart showing genus-specific breakdown of the number of MAGs obtained using short-read and hybrid assemblies (left) and boxplots for corresponding relative abundances of the genera (right). N represents the number of hybrid only MAGs for each genus. c Scatter-plot showing the relative abundance of Bifidobacterium genomes estimated using short-read or hybrid assemblies for a sample (y-axis) versus corresponding relative abundances obtained using the standard Kraken2 database (x-axis). Points found along the x-axis represent Bifidobacterium species found using the Kraken standard database but not found using either short-read or hybrid MAGs. d Violin plots showing the distribution of a contiguity metric (N50 – largest contig size where >50% of the genome is in larger contigs) for short-read and hybrid assembly based MAGs. e Stacked barcharts showing the relative proportion of MAGs satisfying different MIMAG quality standards with short-read and hybrid assemblies of SPMP datasets. f Violin plots showing the relative improvement in contiguity (N50) obtained using hybrid assembly MAGs from SPMP relative to matched genomes in the GTDB database. g Barcharts showing the number of GTDB reference genomes which were improved from medium to high MIMAG quality using SPMP MAGs. Center lines in the boxplots represent median values, box limits represent upper and lower quartile values, whiskers represent 1.5 times the interquartile range above the upper quartile and below the lower quartile, and all data points are represented as dots in the figures. Source data are provided as a source data file.