Computational protocol: A comparison of methods used to unveil the genetic and metabolic pool in the built environment

Similar protocols

Protocol publication

[…] Specific primers targeting the 16S and 18S rRNA genes were designed for each strain using primer-BLAST [] (Additional file : Table S1). The specificity of each primer set was verified to confirm no cross-amplification. Replicated samples and dilutions, as well as a no template, negative control were quantified on a iCycler Thermal Cycler and MyiQ™ Single-Color Real-Time PCR Detection System (BioRad, Hercules, CA, USA) (Additional file : Table S2). The qPCR standards and their cycle threshold (Ct) values were also used as a positive control, with standard deviation < 1.5 from the average Ct value. The Ct values of the no template, negative controls were, at minimum, 5 cycles higher than the detection limit (Ct value of the most diluted qPCR standard) [].The qPCR standards were constructed amplifying the rRNA genes from the pure strains using the designed primers and the same quantification protocol (Additional file : Table S1 and Additional file : Table S2) with no EvaGreen and no melting curve. After purification (QIAquick PCRPurification Kit, Qiagen, Hilden, Germany), the amplicons were cloned and transformed into Escherichia coli TOP10 using the Zero Blunt TOPO PCR Cloning Kit (Invitrogen, Waltham, MA, USA). The cloned sequences were confirmed by Sanger sequencing at UC Berkeley DNA Sequencing Facility. Plasmids were extracted and purified (QIAprep Spin Miniprep Kit, Qiagen, Hilden, Germany), then linearized (BamHI restriction enzyme, New England BioLabs Inc., Ipswich, MA, USA) followed by the quantification of the DNA concentration (Qubit® fluorometer and Qubit® dsDNA HS Assay Kit, Invitrogen, Waltham, MA, USA) and preservation at − 20 °C. [...] Swab and surface types were examined to evaluate their sampling performance. For surfaces, three physically diverse surface types commonly found in the BE (plastic, metal, and untreated wood) were evaluated. Also, three types of swabs (eSwab, BBL CultureSwab EZ, and BiSKit) were compared (Additional file : Table S3). The eSwab is a nylon fiber tipped swabs with inorganic buffer commonly used in indoor studies [, ], BBL CultureSwab EZ is a polyurethane-tipped fiber swabs that was found to have superior performance in human microbiome sampling [], and BiSKit is an sponge-based method with inorganic buffer, commonly used for sampling larger surfaces [].The mock community diluted in PBS buffer was spiked on a 30 cm2 of each type of surface previously washed and sterilized. Preliminary tests evaluating the surface sterilization were conducted, with no amplification detected for any of the primer sets. After the surface was completely dried, it was dry-swabbed in two perpendicular directions. 1 mL of PBS buffer was added to the BBL CultureSwab EZ, and the default buffers were used for eSwab and BiSKit sampling kits. The eSwab and the BBL CultureSwab EZ were then vortexed for 2 min, transferring just the buffer to the Lysis Matrix Y from the FastPrep co-extraction protocol (Additional file : Text S1). The manufacturer’s instructions were followed for BiSKit, centrifuging the buffer for 15 min. at 6800×g to pellet the sample, discarding the buffer, and leaving only 1 mL to resuspend the sample and proceed with the FastPrep co-extraction protocol along with the reference sample. All samples were duplicated. [...] A total of 1,337,415 bacterial 16S rRNA paired end sequences were analyzed using the Quantitative Insights Into Microbial Ecology (QIIME v. 1.9) pipeline []. The raw forward and reverse paired sequence reads were assembled and quality filtered with USEARCH (version 10.0.240) [], discarding the reads with a total expected error of greater than 1 and shorter than 280 bases. Following quality filtering, a total of 914,008 sequences were clustered into operational taxonomic units (OTUs) using the UPARSE [], with a clustering identity threshold of 97%. Taxonomy classifications were performed with the SILVA [] as reference database (version 128 release, 97% representative set file, total of 166,393 sequences). Chimeric OTUs were identified using UCHIME2 [] using the SILVA database. Negative controls of different sample groups (controls for each of DNA and RNA extraction) were included, and OTUs of taxonomic lineages present in more than 3% in the controls were removed from all samples. Following chimeric, contaminating, chloroplast, and mitochondrial OTU removal, OTUs present in less than 100 reads of the entire dataset were removed from the dataset to reduce the effect of noise on data analysis. Thus, a total of 569,372 reads were included for microbial community analyses. Community membership and composition were analyzed using unweighted and weighted UniFrac distances, respectively []. SParse InversE Covariance Estimation for Ecological Association Inference (SPIEC-EASI) was used to assess potential ecological associations between microbial taxa in the active and total populations, with a minimum lambda ratio of 0.01, and reiteration of 50 times []. Network structural properties, including degree distribution and natural connectivity in response to node removal, were examined using R []. Network visualization was constructed with Cytoscape (version 3.5.0) []. To look at the microbiome overlap between the viable bioaerosol population and viable populations of nearby surfaces, Bayesian source-tracking [] approach was performed in QIIME to estimate the contribution of potential sources of the viable component of the residential microbiome. The RNA-based community from different surface at various distances from the air sampling pumps were included in analysis. We performed source-tracking analysis based on two possible scenarios: (1) microbes be re-suspended into the air from surfaces (i.e., air as microbiome sink, and surfaces as sources), and (2) microbes be settled onto nearby surfaces from the air (i.e., air as source, and surfaces as sinks). [...] The results of the in vitro test are expressed as the proportion of the DNA (as 16S/18S rRNA gene copies) and the RNA (as 16S/18S rRNA copies) recovered from the spiked samples in comparison with the reference sample of each set of experiments. R software [] was used for the analyses, with ggplot2 package [] for generating the plots. Nonparametric Kruskal-Wallis (KW) and Mann-Whitney (MW) tests were employed and p values were adjusted for multiple comparisons using the false discovery rate (FDR).ANOSIM Global R and PERMANOVA pseudo-F statistics were calculated for the indoor microbiome samples using QIIME, based on the default setting of 999 permutations. To identify differentially abundant OTUs between genetic and metabolic pool, DeSeq2 was performed with an adjusted p < 0.05 considered statistically significant. Only OTUs with DeSeq2 log-fold changes of at least |2| were considered to be differentially abundant. Where indicated, p values were adjusted for multiple comparisons using the FDR, and Kendall’s τ ranked correlation was computed in R []. […]

Pipeline specifications

Software tools Biskit, QIIME, USEARCH, UPARSE, UCHIME, UniFrac, SPIEC-EASI, Ggplot2, DESeq2
Applications Miscellaneous, 16S rRNA-seq analysis
Organisms Homo sapiens