deletions | additions       diff --git a/Results & Discussion.tex b/Results & Discussion.tex  index f4b876a..d29d837 100644  --- a/Results & Discussion.tex  +++ b/Results & Discussion.tex 
...
Twenty fractions from a cesium chloride gradient fractionation for each treatment at each time point were sequenced (Fig. S1). Using NMDS analysis from weighted unifrac distances the relationship between all buoyant densities from all treatments and time points are plotted (Fig 1). \textsuperscript{13}C-labeled organisms are expected be to found in the higher buoyant density fractions. Each point on the NMDS represents the microbial community from a single fraction where the size of the point is representative of the density of that fraction and the colors represent the treatments (Fig1A) or days (Fig1B). The high density fractions (1.73-1.74...figure out exact densities) that are differentiating from the control along NMDS2 correspond to fractions that contain \textsuperscript{13}C-labeled OTUs (herein called 'responders'). The differential separation of high density fractions in the \textsuperscript{13}C-xylose treatment compared to the \textsuperscript{13}C-cellulose treatment is indicative of a difference in the responders for each of the substrates (Fig 1A). This data also presents an observable time signature of responders at days 1,3, and 7 for the xylose treatment and days 14 and 30 for the cellulose treatment (Fig1B). This demonstrates that different microbial community members are responsible for the consumption of these two different substrates and that xylose is consumed quickly, whereas, cellulose decomposition takes longer. This supports the hypothesis of a microbial community succession during the decomposition process. Furthermore, this demonstrates the sensitivity of this technique by being able to detect \textsuperscript{13}C-label incorporation in samples with low C additions (blah mg g\textsuperscript{-1} soil).   The first step towards determining responders is a biplot of OTUs using the average weighted abundance of each OTU for a single experimental treatment versus the control (FigSx). For each respective \textsuperscript{13}C-substrate biplot, the significance of each OTU's log\textsubscript{2} fold change in relative abundance in the treatment compared to the control was measured. control.  OTUs passing a threshold of \textit{p}-value = <0.05 were used to target potential responders. OTUs that exceed a a user defined 0.75 log\textsubscript{2} fold change threshold are identified as responders (Fig 2). Threshold was determined by controlling for false discovery rate based on sparsity.  their relative abundance in each fraction can be traced for an experimental treatment and compared to its relative abundance in the control fractions from the same time point (Fig Y). Using biplots we can tease out members that cause the greatest shift in experimental treatment versus control. We then generate C utilization charts to demonstration discrete OTUs in control versus treatment.   Using fractions from within a range of high density (1.7125-1.755 g/ml), relative abundances of phyla in the experimental treatments were compared to the respective relative abundances in the control treatment to calculate the log2-fold change (Fig2). Firmicutes show the strongest response on day 1 in the xylose treatment with a steady decline in subsequent time points. Proteobacteria demonstrate the second highest repsonse at day 1 and continue to increase in response up to day 7, followed by decline by days 14 and 30. Demonstrates the boom-bust of phyla with time. Bacteriodetes are the strongest responders on day 3 and proteobacteria are the strongest responders on day 7. Actinobacteria and planktomycetes fluctuate in responsiveness within the first 7 days then decline thereafter.