TABLE 1.

Results of 16S sequencing analyses by using QIIME2, Bioconductor, UPARSE, or mothur.

	QIIME2		Bioconductor		UPARSE		Mothur		p-Valuec
	Linux	Mac	Linux	Mac	Linux	Mac	Linux	Mac
Software version	2018.08		R version 3.5.1 dada2 v 1.8.0 phyloseq v 1.24.2 phangorn v 2.4.0		v11.0.667_i86 linux32	v11.0.667_i86 osx32	v.1.43.0	v.1.43.0
Approximate analysis time	4 h	3 h	9 h	8 h	45 min	45 min	9 h	9 h	–
# input reads	4715000
# reads after filtering/denoising	3,391,670		3,736,927		3,173,733		3,244,489		–
# reads assigned at phylum levela (tot n, mean ± SD)	2,941,772 (87%) 73,544 ± 40,280		3,143,413 (84%) 78,585 ± 37,730		3,123,028 (98%) 78,076 ± 46,872	3,131,211 (99%) 78,280 ± 46,973	2,812,333 (89%) 70,308 ± 35,353	2,812,470 (87%) 70,312 ± 35,356	<0.001
# unclassified reads at phylum level (tot n, mean ± SD)	3,567 (< 1%) 89 ± 210		163 (< 1%) 4 ± 7		77,395 (2%) 1,935 ± 5,119	79,031 (2%) 1,976 ± 5,130	47,883 (2%) 1,161 ± 4,046	48,770 (2%) 1,219 ± 4,048	<0.001
# reads assigned at genus levelb (tot n, mean ± SD)	2,770,029 (82%) 69,251 ± 37,324		2,798,953 (75%) 69,974 ± 32,881		3,123,028 (98%) 78,076 ± 46,283	3,131,211 (99%) 78,280 ± 46,382	2,812,333 (89%) 70,310 ± 34,911	2,812,470 (87%) 75,717 ± 12,956	<0.001
# unclassified reads at genus level (tot n, mean ± SD)	171,743 b (5%) 4,294 ± 4,139		Taxonomic filtering during processing		801,052 (25%) 20,026 ± 16,033	923,329 (29%) 23,083 ± 19,208	763,315 (24%) 19,083 ± 13,009	751,717 (23%) 18,793 ± 12,956	<0.001

QIIME2 and Bioconductor gave identical outcomes on both operating systems. Mothur failed to complete the pipeline on Mac. Percentage are calculated using the # of reads after filtering/denoising as denominator. ^aReads assigned any bacteria identity (including “unclassified”) but excluding Eukaryota and Archaea. ^bFor QIIME2 we considered as unclassified sequences those with incomplete taxonomies (terminating at any level with ‘__’). ^cp-Values using Friedman test followed by Dunn’s multiple comparisons test.