Mikaela Cashman — University of Nebraska, Lincoln
Our research team—-consisting of computer scientists, computer networkers, biochemists, and mathematicians–has found KBase to be an essential tool to our work studying the metabolism of B. theta and M. smithii. Our group chose to use KBase due to its unique capabilities for metabolic modelling and because the genomes of our organisms have already been integrated into KBase’s data resources. We run high-throughput experiments using the metabolic modeling apps Build Metabolic Model, Gapfill Metabolic Model, and Flux Balance Analysis to generate growth terms so we can better understand how different media compounds affect an organism’s metabolism. An example of our work in KBase is the BioSIMP (Biological Sampling, Inference, Modeling, and Prediction) Narrative in which we apply BioSIMP to an organism to see how changes in the media affect the objective value and reaction fluxes.
As a computer scientist, I have made use of KBase’s Software Development Kit (SDK) to develop and deploy my own apps into KBase. The SDK allows me to take new tools and integrate them back into KBase so other researchers can utilize them. Specifically, I have integrated packages from a machine learning software suite called Weka into KBase apps, in order to apply these computer science principles to solve biological problems. Due to the flexibility of KBase, the group is in the process of setting up larger scale analyses to more broadly study the relation between a microbe’s environment and its phenotypic response.
In each adventure we have taken, the KBase team has shown incredible effort and interest in pushing our research through. We plan to continue to use and help in developing new apps for KBase.
Juan C. Villada — City University of Hong Kong
In the Lee Research Group (the Environmental Microbiology and Biotechnology Laboratory at the City University of Hong Kong), we investigate bacteria for energy, environmental and biotechnology applications. We apply different methods at the interface of evolution, metabolism and ecology to understand the microbial conversion of carbons. We find in KBase a comprehensive set of data and tools—some of which are not available anywhere else—that has assisted us to answer many of our research questions. For example, we have applied the metabolic model reconstruction methods to study the metabolism of a single methanotrophic bacterium, and its interactions with different species.
We are particularly amazed by the fast processing time of KBase. We are able to annotate genomes, compare them, find phylogenies, reconstruct metabolic models, and integrate transcriptomics data in just a matter of hours. Doing all of these steps without KBase would have required employing multiple different software tools and websites, probably requiring days of processing time.
Additionally, using KBase Narratives, we can work in a collaborative fashion, organizing different parts of our project in different Narratives. Once we complete our project, we will be able to share our work with the public so other researchers can follow and interact with our research. KBase is of enormous help to solve our research questions and we are sure it will also be useful for other research groups.
Image: Biochemical network including genes, proteins, reactions and metabolites of a methane-oxidizing bacterium. The image was generated after manual curation of a draft genome-scale metabolic model reconstructed using KBase. Villada et al., 2017
Collin Timm — Oak Ridge National Laboratory
KBase is a great environment with many tools that are easy to use. Using the metabolic modeling (FBA) tools we have studied differences in metabolic properties of Pseudomonas fluorescens isolates from Populus trees. As a first-time user of metabolic modeling tools I found it very fast (~15 minutes) and easy (~5 steps) to go from genome sequence to draft models that could be used for biological discovery. The interface allows for easy organization of steps and data for easy documentation of modeling steps. Using KBase tools we found that endosphere (internal root compartment) have significantly different metabolic properties than rhizosphere isolates (external root compartment) that is consistent with what is known about these compartments. This work helps us rapidly characterize new bacterial isolates and will help identify processes important for plant-microbe interactions. We are preparing this work for publication in a paper describing phenotypic variation between endosphere and rhizosphere bacterial isolates.
Matt Scarborough — Great Lakes Bioenergy Research Center
My research interests include constructing and using metabolic models of numerous microbial species. As part of this research, our lab at the Great Lakes Bioenergy Research Center developed a metabolic and regulatory model of Rhodobacter sphaeroides 2.4.1. We worked with KBase collaborators at Argonne National Laboratory to load this model, iRsp1140, into the KBase system. Since that time, iRsp1140 has become a central example in KBase. Furthermore, the publications created with this model have been classified as “highly accessed,” a designation reserved for only a few papers and one that demonstrates the scientific community’s interest in this research.
Since then, I have used KBase to propagate our published Rhodobacter iRsp1140 model to new genomes, dramatically speeding the rate at which I’m able to produce new high-quality models. I’ve extensively used KBase’s metabolic modeling, genome annotation, and proteome comparison tools to help determine how other purple non-sulfur bacteria species vary from the R. sphaeroides 2.4.1 strain.
Others in my group and university are using KBase to annotate and build metabolic models for various microbes, including novel species involved in nitrogen cycling in the environment. We also hope to use KBase to construct metabolic models of multi-organism systems to better understand syntrophic interactions among microorganisms.
Steve Lindemann — Pacific Northwest National Laboratory / Purdue
We have used KBase extensively in our work on the Principles of Microbial Community Design Science Focus Area. We applied KBase to annotate and construct genome-scale metabolic models for the 22 species comprising our unicyanobacterial consortia. We manually reviewed many of these annotations, finding good agreement with our own annotations performed outside of KBase. This improved our confidence in using the KBase tools. We then applied KBase to place our genomes into a phylogenetic context and compare our closely related species in detail to identify subtle differences between strains. This provided a valuable new perspective to versions of this analysis that we performed outside of KBase. Additionally, we applied the phenotype simulation tool in KBase to predict numerous growth conditions for each of our species, which we used to identify media formulations that might enrich for specific species that we wish to isolate. This was a capability that was unique to KBase, and it provided valuable support for new experiment design. We plan to expand this approach to use KBase to study interactions between pairs of species.
We are now working with the KBase to construct a community metabolic model of our Hot Lake microbial mat, integrating transcriptomic and metabolomic data using new prototype tools developed by the science engagement team of KBase. Overall, we have been impressed at how easy the Narrative Interface is to use. Members of our team who are not computational biologists, such as myself, were able to perform complex workflows using the Narrative interface. The Narrative is also valuable as a collaborative tool, as it shows every detail of the analyses performed by our collaborators using our data This makes it possible for us to not only understand the analysis, but also to tweak it to explore new scientific questions.