Fig. 1

Off-target amplifications dominate amplicon sequences from brain tissue. a Clean extraction increases off-target amplifications: The fraction of bacterial reads that could be attributed to off-target amplifications significantly increases in relation to the tissue handling method used (P-value = 8e−08, t-test, PK2 was more prone to contaminants, see “Methods”). Only samples with > 0.1% off-target relative abundance were included. b The relative fraction of off-target reads in each sample is significantly anticorrelated to the amount of 16S rRNA gene copies (determined via qPCR) in each sequencing well (uncorrected, P =3e−5, Pearson correlation, rho=− 0.52). c The fraction of mouse off-target reads is increased in the 8 mouse brain samples included in this study. No mouse off-target reads were found in human brain samples. The difference in off-target fraction between mice from germ-free facilities (Mouse GRF, N=3) was not significantly different to those from mice in specific pathogen-free facilities (Mouse SPF). d Classification of off-target amplicons, off-target classification of different databases for taxonomic assignment at the domain or phylum level. Most pipelines misclassify some off-target amplicons as bacteria and give taxonomic assignments at domain and phylum level (shown), in some cases down to the species level (not shown); H, human; M, mouse; Qiime1-s, Qiime 1 sortmerna, Qiime1-u, Qiime 1 uclust