deletions | additions       diff --git a/Abstract.tex b/Abstract.tex  index 56ebde0..20524df 100644  --- a/Abstract.tex  +++ b/Abstract.tex 
...
\begin{abstract}  How the environment affects resource availability influences  which microbial plankton colonize new surfaces to form biofilms is poorly understood. Heterotrophic bacteria derive some to all  of their organic carbon (C) from photoautotrophs while simultaneously competing  for inorganic nutrients such as phosphorus (P). (P) or nitrogen (N).  Therefore, C inputs have the potential to shift the competitive balance of aquatic microbial communities by  increasing the resource space available to heterotrophic bacteria (more C)  while decreasing the resource space available to algae (less P mineral nutrients  due to increased competition from bacteria). osmotrophic heterotrophs).  To test how resource dynamics affect membership of planktonic communities and assembly of biofilm communities we amended a series  of flow-through mesocosms with C and P to achieve four target resource C:P  levels. alter the availability of C among treatments.  Each mesoscosm was fed with unfiltered seawater and incubated with sterile glass substrate for biofilm formation. We used 454 pyrosequencing of  bacterial 16S and 23S plastid genes to ask how resource driven shifts in the  pool size of each community affected community membership and structure. The 
...
including the highest C treatment. Our results suggest that even though  resource amendments affect community membership, microbial lifestyle (biofilm  or planktonic) places a significanlty stronger constraint on community  assembly. assembly and membership.  \tiny  \keyFont{ \section{Keywords:} microbial ecology, 16S, 23S, planktonic,  diff --git a/Discussion.tex b/Discussion.tex  index e846e86..bfb274d 100644  --- a/Discussion.tex  +++ b/Discussion.tex 
...
\section{Discussion}   \subsection{Biomass Pool Size} The goal of this study was  to evaluate how changes in resource stoichiometry available C  affected the biomass pool size, membership and structure of planktonic and biofilm communities. Our  results suggest that C subsidies increased bacterial biomass in both plankton  and biofilm communities as predicted. Carbon subsidies also resulted in 
...
resource treatments. The changes in the biomass pool size that did occur were  consistent with changing relationships (commensal to competitive) between the  autotrophic and heterotrophic components of the plankton communities but not  necessarily of the biofilm communities. While we recognize that other mechanisms may drive the shift in biomass pool size of these two components of the microbial community (e.g. increased grazing pressure on the algae with C additions, or producitons of secondary metabolites by the bacteria that inhibit algal growth) previous studies (e.g. Stets and Cotner 2008b, Cotner and Biddanda 2002) and the data reported here suggest that altered nutrient competition is the most parsimonious explanation for this shift in biomass pool size.  \subsection{Biofilm and Plankton Alpha and Beta Diversity} Beyond changes in  the biomass pool size of each community we explored how shifts in resource C:P C  affected a) the membership and structure of each community, and b) the  recruitment of plankton during biofilm community assembly. We highlight three  key results that we find important for understanding the assembly of aquatic 
...
the control, C:P = 10 and C:P = 100 resource treatments the membership and  structure of the bacterial biofilm and plankton communities were more similar  within a lifestyle (plankton versus biofilm) than within a resource treatment.  However, for the bacteria in the highest C:P C  treatment (C:P = 500) both membership and structure of biofilm and planktonic communities at day 17 were  more similar to each other than to communities from other treatments  (Figure~\ref{fig:pcoa}). Third, C subsides acted differently on the photoautotroph and bacterial communities. Specifically while the highest level of C subsidies (C:P = 500) resulted in a merging of membership in the  bacterioplankton and bacterial biofilm communities the same merging of  membership was not observed for the photoautotroph biofilm and plankton 
...
through incubators was dynamic in time. In this case the biofilm community  would represent a temporally integrated sample of the planktonic organisms  moving through the reactor resulting in higher apparent alpha diversity (i.e.  mass effects would be the dominant assembly mechanism). Second, the biofilm environment may disproportionately enrich for the least abundant members of the  of the planktonic community. In this case it is probable that the biofilm would  incorporate the most abundant members from the planktonic community (i.e. mass  effects) but also select and enrich (i.e. species sorting) the lease abundant  members of the planktonic community resulting in a higher level of detectable  alpha diversity. The second mechanism would result if the biofilm environment represented a more diverse habitat including sharply delineated oxygen,  nutrient and pH gradients that are not present in the planktonic environment.  In this case the more diverse habitat would be able to support a more diverse 
...
timepoints than any other bacterioplankton community (treatment or timepoint).  In addition, the control and two lowest C treatments (C:P=10 and C:P=100)  separated completely from biofilm communities in principle coordinate space  (Bray-Curtis distance metric). This suggests that the biofilm community was notonly  integrating variable bacterioplankton community membership, but rather was at least in part selecting for a community that was composed of distinct  populations when compared to the most abundant members of the plankton  community. As noted above, in the highest C treatment (C:P = 500) the bacterial 
...
habitat but readily became major constituents of the biofilm community.  % Fakesubsubsection:Very few studies have  Very few studies have previously simultaneously  evaluated the relationship among membership and/or diversity of the plankton and the biofilm community from complex  environmental microbial communities. One notable study looked at planktonic  community composition and biofilm formation on glass beads placed for three 
...
to soil communities in the watershed, as residence time of the system slows the  relative influence of species sorting increases. Thus, in headwater ecosystems  stream plankton communities can often be composed primarily of soil organisms  \citep{22378536}. In addition to the diverse source communities the \citet{22237539} study sampled the plankton community at multiple timepoints  and integrated the samples before sequencing, further increasing community  richness as compared to the current study where the plankton community was 
...
that sample-wise bacterial biofilm rarefaction curves may exceed the integrated  planktonic curve upon extrapolation and most exceed the integrated planktonic  curve at sampling depths where data is present for the biofilm and integrated  planktonic library (Figure~\ref{fig:integ_rarefaction}). This result is consistent with our conclusion that temporal heterogeneity in the plankton was  not sufficient to explain the higher diversity in the biofilm sample but would  explain the relative differences between planktonic and biofilm diversity found 
...
communities both microbial and otherwise. However, as with other experiments  with this result our experimental design did not allow us to tell whether  resources drove productivity that drove changes in diversity or whether  resources drove diversity which altered productivity. Rather Rather,  we note that, that  as diversity decreased in the highest C treatment, bacterioplankton and biofilm  membership became increasingly similar. This suggests that environments that  contained high amounts of labile C selected for fewer dominant taxa,  overwhelming the lifestyle species sorting mechanisms that appeared to dominate  biofilm community assembly in all other treatments. Similarly, while we did not measure extracellular polymeric substances (EPS), direct microscopy showed  that planktonic cells in the highest C treatment (C:P = 500) were surrounded by  what appeared to be EPS. Because biofilm EPS appeared also to increase moving 
...
as opposed to the plankton. Greater niche diversity should select for a more  diverse set of taxa but individual taxa would not be as numerically dominant as  in a more uniform environment such as the planktonic environment. At the Order  level, enriched bacterial OTUs tended to have membersin  that were enriched in both the plankton and the biofilm suggesting the phylogenetic coherence of  lifestyle is not captured at the level of Order. It should be noted however  that taxonomic annotations in reference databases and therefore environmental 
...
sufficient number of OTUs to evaluate coherence between taxonomic annotation  and lifestyle. Carbon amendments did not affect photoautotroph library  membership and structure to the same degree as it affected bacterial library  composition. As expected, bacterial OTUs enriched in the high C amended mesocosm (C:P = 500) include OTUs in classic copiotroph families such as  \textit{Altermonodales} and \textit{Pseudomonadaceae}. Interestingly, the most depleted OTU in the high C treatments is annotated as being in the HTCC2188  order of the \textit{Gammaproteobacteria} and shares 99\% sequence identity with another "HTCC" strain (accession AY386332). HTCC stands for 'high throughput  culture collection' and is a prefix for strains cultured  under low nutrient conditions \citep{Cho_2004, Connon_2002}.  \subsection{Conclusion} In summary this study shows mechanistic links between  large scale that changes in low resolution  community level dynamics and are concurrent with changes in  the underlying constituent populations that compose them. We found that autotrophic pools and heterotrophic pools responded differently to amendments of labile C as hypothesized. Notably while  C amendments altered both pool size and membership of the bacterial communities  we did not see similar dynamics within the photoautotroph communities.  Planktonic photoautotrophs decreased in response to C amendments presumably in  response to increased competition for mineral nutrients  from a larger bacterial community, however there was not a similar decrease in biofilm photoautotroph community. In  addition membership of the photoautotroph communities between the plankton and  biofilm lifestyles did not become more similar in the photoautotrophs as it did  for the bacteria in the highest C treatment. Consistent with a growing body of work our results suggest that complex environmental biofilms are a unique microbial community that form from taxa (both heterotrophs and autotrophs  alike) that are found in low abundance in the neighboring communities. This  membership was affected by resource amendments for heterotrophic but not  diff --git a/Introduction.tex b/Introduction.tex  index ebe8f80..53e9681 100644  --- a/Introduction.tex  +++ b/Introduction.tex 
...
competitors for limiting nutrients as has been observed  \citep[see][Figure~\ref{fig:conceptual}]{COTNER_1992}. These dynamics should  result in the increase in bacterial biomass relative to the photoautotroph biomass along  a gradient of increasing labile C inputs. We refer to this differential allocation of limiting resources among components of the microbial community as niche partitioning, in reference to the n-dimensional resource space available to members of the microbial community.  % Fakesubsubsection:While these gross level dynamics have been discussed  While these gross level dynamics have been discussed conceptually  diff --git a/Materials_and_Methods.tex b/Materials_and_Methods.tex  index 291ad57..df64217 100644  --- a/Materials_and_Methods.tex  +++ b/Materials_and_Methods.tex 
...
\section{Materials and Methods}  \subsubsection{Experimental Design}  We placed test tube racks in one smaller (185L, control) and 3 larger (370L)  flow-through mesocosms. All mesocosms were fed directly with marine water froma  an inflow source in Great Bay approximately 200 m from the shore. Each mesocosm had an adjustable flow rate that resulted in a residence time of approximately 12h. Irregular variation in inflow rate  meant that flow rate varied around that target throughout the day, however,  regular monitoring ensured that the entire volume of each system was flushed 
...
slide to the test tube racks using office-style binder clips. Twice daily 10 ml  of 37 mM KPO$_{4}$ and 1, 5 and 50 ml of 3.7M glucose were added to each of 3  mesocosms to achieve target C:P resource amendments of 10, 100 and 500  respectively. The goal of the resource ammendements were to create a gradient of labile carbon among treatments. The same amount of P was added to each treated mesocosom to ensure that response to additions of C were not inhibited by extreme P limitation. The  control mesocosm did not receive any C or P amendments. \subsubsection{DOC and Chlorophyll Measurements}  To assess the efficacy of the C additions we sampled each mesocosm twice  diff --git a/Results.tex b/Results.tex  index d14243a..d477e87 100644  --- a/Results.tex  +++ b/Results.tex 
...
\subsection{Bulk community characteristics}  We first assessed the effect of the resource treatments on the dissolved  chemistry and bulk community characteristics of the plankton and the biofilms.  Dissolved organic C (DOC) levels in the control and lowest C:P C  treatment (10) (C:P=10)  remained below 2 $\mu$moles C L$^{-1}$ throughout the course of the experiment.  Altered resource C:P in the The  higher two C treatments (C:P 100 and 500) resulted in changes in the DOC concentration of the water column. In the intermediate  treatment (C:P 100) DOC increased on the second day and then returned to the  same level as the lower two treatments for the remainder of the experiment. In 
...
sixth day of the experiment.   % Fakesubsubsection:This increase in DOC  This increase in DOC in the higher C:P C  treatments was associated with decreases in planktonic Chl \textit{a} in each treatment (Figure~\ref{fig:pool_size}a),  however there was no significant difference in biofilm Chl \textit{a} among  treatments (Figure~\ref{fig:pool_size}b). In combination with the decrease in  planktonic Chl \textit{a} on the 6th day of the experiment the highest C:P C  treatment had approximately 4-fold higher planktonic bacterial abundance than  the control and the 10 $\mu$M C treatment (Figure~\ref{fig:pool_size}d).  Similarly, biofilms had significantly higher total biomass in the high C  treatment compared to the other treatments (Figure~\ref{fig:pool_size}c).These  differences in biomass could also be clearly visualized among biofilms grown in  each treatment. Cell density, biofilm thickness and amount of apparent EPS all  increased visually with increasing C:P resource treatment  (Figure~\ref{fig:microscope}).  Thus the shift in resource C:P altered the pool size of both the photoautotroph and bacterial communities. Clear differences in bacterial and photoautotroph pool size among treatments allowed us to address how shifts in pool sizes were related to community membership and structure within and among plankton and biofilm communities. \subsection{Planktonic and biofilm community structure}  \subsubsection{Alpha diversity} 
...
bacterial and photoautotroph OTU richness was consistently higher in the biofilm  compared to the planktonic communities (Figure~\ref{fig:rarefaction}). For both  the photoautotroph and bacterial sequence datasets the biofilm and planktonic  communities had the fewest OTUs in the highest C:P C  treatment (C:P = 500) (Figure~\ref{fig:rarefaction}).  \subsubsection{Community membership biofilm versus plankton} Bacterial 
...
lifestyles only two bacterial OTU centroid sequences shared high sequence  identity ($>=$ 97\%) with cultured isolates (Table~\ref{Tab:01}).   \input{table1.tex}  % Fakesubsubsection:We similarly assessed membership  We similarly assessed membership among biofilm and plankton photoautotroph communities.  Photoautotroph 23S plastid rRNA gene sequence libraries also clustered strongly by  diff --git a/Sequence_Quality_Control_and_Analysis.tex b/Sequence_Quality_Control_and_Analysis.tex  index e46eea9..9772f4c 100644  --- a/Sequence_Quality_Control_and_Analysis.tex  +++ b/Sequence_Quality_Control_and_Analysis.tex 
...
applied to microbiome OTU count data see \citet{24699258}). We use the term  {\textquotedblleft}differential abundance{\textquotedblright} coined by  \citet{24699258} to denote OTUs that have different proportion means across  sample classes. We are were  particularly interested in two sample classes: 1) lifestyle (biofilm or planktonic) and, 2) high C (C:P = 500) versus  not high C (C:P = 10, C:P = 100 and C:P = control). A differentially  abundant OTU would have a proportion mean in one class that is 
...
used to summarize the proportion mean difference. Here we use log$_{2}$ of the  proportion mean ratio (means are derived from OTU proportions for all samples  in each given class) as our differential abundance metric. It is also important  to note that the DESeq2 R package we are using used  to calculate the differential abundance metric {\textquotedblleft}shrinks{\textquotedblright} the metric in  inverse proportion to the information content for each OTU. In this way the  magnitude of the differential abundance metric will be high only for OTUs which 
...
sequence 16S/plastid 23S library comparisons. The specific sparsity threshold  for plastid 23S and 16S libraries for biofilm versus plankton comparisons was  10\% (OTUs found in less than the sparsity threshold of samples were discarded  from the analysis). Cook's distance filtering was also disabled when calculating p-values with DESeq2. We used the Benjamini-Hochberg method to  adjust p-values for multiple testing \citep{citeulike:1042553}. Identical  DESeq2 methods were used to assess enriched OTUs from relative abundances  diff --git a/figures/conceptualmodel/caption.tex b/figures/conceptualmodel/caption.tex  index 1a36841..2f43b81 100644  --- a/figures/conceptualmodel/caption.tex  +++ b/figures/conceptualmodel/caption.tex 
...
Carbon subsidies in the form of glucose alleviate the dependence of  heterotrophic bacteria on photoautotroph derived C(C)  exudates. This should result in an increase in resource space and biomass for bacteria and a decrease in  resource space and biomass for photoautotrophs due to increased competition for  phosphorus (P). mineral nutrients (for simplicity we illustrate competition for P but this is equally applicable other elements that may limit primary production).  We hypothesized that this predicted change in biomass pool size of these two groups will result in changes in the plankton community  composition of both groups that will propagate to to the composition of biofilm  communities for both groups. We refer to shifts in the demand and availability of resources among components of the microbial community as 'partitioning.  diff --git a/figures/lightmicroscopy.002/biofilmsubsidiesFigs06252013.003.jpg b/figures/lightmicroscopy.002/biofilmsubsidiesFigs06252013.003.jpg  deleted file mode 100644  index 303eb77..0000000  Binary files a/figures/lightmicroscopy.002/biofilmsubsidiesFigs06252013.003.jpg and /dev/null differ diff --git a/layout.md b/layout.md  index fcc923c..f5c4a62 100644  --- a/layout.md  +++ b/layout.md 
...
Discussion.tex  figures/conceptualmodel/biofilmsubsidiesFigs06252013.001.jpg  figures/pool_size/biofilmsubsidiesFigs07252013.003.jpg  figures/lightmicroscopy.002/biofilmsubsidiesFigs06252013.003.jpg  figures/combined_rarefaction5/combined_rarefaction.png  figures/biplot_combined1/biplot_combined.png  figures/l2fc_pval_all1/l2fc.png  diff --git a/table1.tex b/table1.tex  index bcbc232..a0fdca6 100644  --- a/table1.tex  +++ b/table1.tex 
...
\begin{table}[htp]  \textbf{\refstepcounter{table}\label{Tab:01} Table \arabic{table}.}{ Results for BLAST search against Living Tree Project (top 25 lifestyle enriched bacterial OTUs) }  {\tiny  \begin{tabular}{llr>{\itshape}lrl}  \toprule \\  \textbf{OTU ID} & \textbf{Phylum} & $log_2(plankton:biofilm)$ $log_2(plank:biof)$  & \textbf{Species Name} & \textbf{BLAST percent identity} \%ID}  & \textbf{accession} \textbf{ACC}  \\ \midrule  \multirow{1}{*}{OTU.103} & \multirow{1}{*}{Bacteroidetes} & \multirow{1}{*}{7.78} & Zunongwangia profunda & 89.66 & DQ855467 \\ \midrule  \multirow{1}{*}{OTU.105} & \multirow{1}{*}{Proteobacteria} & \multirow{1}{*}{8.09} & Microbulbifer yueqingensis & 90.14 & GQ262813 \\ \midrule 
...
\bottomrule  \end{tabular}  }  \end{table}