Fig. 4

Improvement of taxonomic and functional classification of sequencing reads by HRGM. a Proportion of taxonomically classified sequencing reads of WMS data from four different populations. The significance of the improvement was calculated by Wilcoxon signed-rank test. Brown–Forsythe test was used to evaluate the decrease of variance. b, c Percent improvement of the read classification proportion in HRGM-based database compared with the standard database (b) and the UHGG-based database (c). d The number of reads aligned to the UHGP-95 and HRGM-95 protein catalogs. Statistical significance was calculated by using Wilcoxon signed-rank test. e The number of sequencing reads that are classified by UHGG and HRGM-based kraken2 database. f Top six most aligned families (left), genera (center), and species (right) by 831.6 million reads that are classified by HRGM-based database but not by UHGG-based database