Set of 140 in silico microbial communities used to evaluate the performance of metagenomic classifiers. These mock communities have varying species diversity, intraspecific diversity, and genomic similarity to a standardized reference genome database. Each mock was simulated to a depth of 2.1 Gb using 2×150 bp paired-end reads with strain abundances following a log-normal distribution. Each mock community has the format ani<ani>_c<complexity>_st<strain_diversity>_r<replicate> where: ani: indicates the ANI between genomes in the mock and standardized reference database. This field is 95, 97, or 99 indicating ANI ranges of [95%, 97%), [97, 99%), and [99, 99.75%), respectively. Mock communities at 100% ANI are also provide as a baseline and were constructed from genomes within the standardized reference database. complexity: indicates the species diversity of the mock. This field is either LOW or HIGH indicating the mock is comprised of 100 +/- 25 or 500 +/- 100 species, respectively. strain_diversity: indicates the strain diversity within each species. This field is False or True indicating the mock is either comprised of a single strain (genome) from each species or up to 10 randomly selected strains (genomes) from each species. replicate: indicates the replicate number of the mock. Ten replicates were generated for each parameter combination. Mock communities were simulated under all combinations of these parameters, with the exception of mocks with HIGH species diversity and [95%, 97) ANI similarity as there were insufficient species with available genomes within this ANI range. Note: provided files contain both the low and high diversity mocks for the indicated parameters in the filename.