Hi Jerome,

Thanks for your interest! The first step will be to characterize what microbes are present in our samples using QIIME (we will be sequencing 16S rRNA gene amplicons), software we develop in the lab that facilitates microbial ecology studies and matches metadata to the samples. Non-metric dimensional scaling (NMDS) can be used to attempt to correlate the phylogenetic distance between samples (beta diversity) with particular metadata fields (such as concentrations of various compounds). If no set of individual metadata fields can effectively explain the differences between samples, then nonlinear regression can be used to test what combinations of factors may be responsible. Assuming we see some clustering of our data points, machine learning techniques can be used (e.g., random forests/decision trees) to determine which fields are the most powerful determinants of cluster membership according to the data, and techniques such as ANOSIM can be used to test the statistical significance.

Best,
Adam

Hi Markus,
Thanks for your questions. The samples we collected include examples where the end results were undesirable for biofuel production, e.g., culture crash or a preponderance of weedy algae species. Also, because we tracked some containers using a time course samples, we will try correlate increases and decreases in relative abundance of the various taxa we observe with overall health of the algae. Typically, an assumption is made that the health of the algae crop is positively correlated with the growth rate of the desired alge species. For now, we will be operating under this assumption as we try to identify bacterial species that are more frequently associated with healthy growth and good yield (or, stunted growth and poor yield). The future directions section of my poster includes looking at metatranscriptomics data, which might reveal not just which bacterial species are “good” and “bad,” but also why they are beneficial or detrimental: what are these bacteria expressing that leads to the changes we observe?

Antibiotics can be used in the production process. However, treating growth chambers with antibiotics is expensive, especially at an industrial-scale production facility. One of the things that Ron Parsons and I discussed is the difficulty that these companies face in terms of competition on the financial markets. Already, biofuel is substantially more expensive than traditional fossil fuels — just one of the hurdles curtailing wider adoption — and treatment with antibiotics would drive the price even higher, into an uncompetitive stratus. Moreover, it is not clear that eliminating all microbes is the most effective way to produce fuel. We hope (and hypothesize) that there exists some combination of bacteria that can actually help the algae grow (in the same way that the bacteria in our gut help us to metabolize the food we eat) while at the same time making the entire growth system more resilient to invasion by unwanted bacteria and other algae species.
Best,
Adam

Hi Govindarajan,
I appreciate your curiosity. In the current analysis, we will be looking only at community composition via 16S rRNA gene amplicon sequencing. Although we can map these taxa directly to full genomes in a handful of cases, typically whole genomes are not available for the taxa we identify. There are some tricks we can use to get approximations of these genomes (nearest neighbor mapping or, perhaps better, ancestral state reconstruction), but more often than not we are left with an incomplete picture of the functional capabilities of the bacteria in the samples we take. If our preliminary results show that the bacteria we observe to co-occur with the algae are not well characterized with full genome sequences, then a ver interesting future direction would be to look at the metagenomic data in our samples to identify what kinds of toxins or metabolites the bacteria might be contributing. So we will have to wait and see what kind of resolution we can achieve with respect to the metabolites (including endotoxins) that might be present. Although there is no soil, we anticipate that the different types of growth chambers (and different seasons of the year, and different nutrient and mineral concentrations, etc.) will contribute to different bacterial communities and, therefore, likely different toxins.
Best,
Adam

Hi Zhaomin,
Unfortunately, we have only gathered the samples. We are going to do our sequencing on the Illumina MiSeq next month. One of the challenges in amplifying 16S rRNA genes in our samples is that there are a lot of chloroplasts present because of the algae. So we have had to spend extra time thinking about alternative primers that will exclude chloroplasts, or blocking primers that will prevent amplification. We are currently testing the efficacy and specificity of a variety of primer sets, and once we have done that we will be able to move forward with sequencing.
Best,
Adam

Hi Kristin,
That’s a great question. Biofilms are a great hindrance to algal biofuel production because they coat the reaction containers, blocking sunlight and greatly reducing photosynthetic efficiency. Although it is not the focus of the project, I think this would be an excellent question to investigate. Although we do not have data directly related to this question, there is published data about other biofilms (guerrero negro, oral plaques, etc.), so we might have success looking either for taxa that are known to participate in biofilm formation or by looking at shared genomic features. We do not have the shotgun sequencing data that would best facilitate such an investigation, but that is a direction I would like to take this research in the future.
Best,
Adam

Thanks for the kind comment! I wish I could take credit for the photo, but I am not the photographer.

Thank you so much for the kind words!