Fig 1.
A diagram depicting the possible HTS-SIP analyses available in the HTSSIP R package.
The R functions to conduct each workflow step are italicized, and the figure references refer to example data produced by these workflow steps.
Fig 2.
Examples of analyses available in the HTSSIP R package for visualizing results from HTS-SIP experiments.
These figures were generated using data from an HTS-SIP experiment that consisted of 3 treatments where only the isotopic label was varied: an unlabled control (12C-Con), 13C-xylose (13C-Xyl), and 13C-cellulose (13C-Cel) [7]. DNA was extraced from each treatment 3 and 14 days after substrate addition, subjected to CsCl fractionation, and 16S rRNA sequencing was performed on ~24 fractions per CsCl gradient. Analyzing beta-diversity between gradient fractions can reveal changes in the buoyant density (BD) distribution of DNA in 13C-labeled treatments versus the unlabeled control. Panel A is a non-metric multidimensional scaling (NMDS) ordination comparing beta-diversity (weighted Unifrac) between gradient fractions from two unlabeled replicates, which indicates that sequence composition varies greatly with BD but not between experimental replicates. Panel B is an NMDS ordination similar to Panel A, but each facet contains a 13C treatment and the corresponding unlabeled control. Large relative distance between “heavy” control and labeled gradient fractions indicates a change in community composition caused by shift in DNA BD that resulted from 13C isotope incorporation. The NMDS stress values ranged from 0.06 to 0.07. Panels C and D depict the same data as in Panels A and B, respectively, but only depict beta-diversity among gradient fractions that correspond in BD between labeled treatment and unlabeled control fractions. In this visualization, increases in beta-diversity indicate a change in community composition caused by a shift in DNA BD resulting from 13C isotope incorporation. The line ranges represent 95% confidence intervals (CI) calculated by permuting OTU abundances and recalculating beta-diversity (100 bootstrap replicates), and actual beta-diversity values are represented by colored circles. "BD shift windows" (blue circles) indicate regions defined by ≥3 consecutive fractions with beta-diversities greater than expected by chance (i.e. exceeding the CI).
Fig 3.
Examples of using the HTSSIP R package for data processing, data exploration, incorporator identification, and quantification of BD shifts.
The SIPSim toolset was used to simulate a simple HTSSIP dataset consisting of two treatments: a 13C-treatment (“13C-Treat”) and a 12C-control (“12C-Con”). Each treatment has 3 replicates, with each consisting of 24 gradient fractions [7]. Half of the ten OTUs (OTU 1, 3, 4, 7, and 8) were randomly assigned an atom excess 13C between 30 and 100%. [3]. Panel A depicts the raw abundances (“Counts”), fractional relative abundance (“Rel. Abund.”), and relative abundances transformed by simulated qPCR data (“Rel. Abund. qPCR-trans.”). For clarity, only 1 of the 3 experimental replicates is shown. Panel B shows those OTUs identified as “incorporators” (incorporated 13C into genomic DNA) by HR-SIP, MW-HR-SIP, qSIP, or by using a “heavy-SIP” approach (the Mann Whitney U test). Panel C shows the BD shift of each OTU as determined by ΔBD or qSIP. The dashed line signifies a BD shift (Z) of 0.0 g ml-1, and the red bars show the true theoretical BD shift resulting from 13C isotope incorporation.