Michael Shaffer

DRAM more than a genome annotation tool. While it performs this function, and does very well at it in comparison to similar software, what really sets DRAM apart is the ability to summarize the functions that are present in a genome/metagenome in an easy to understand way. It distills this information to generate three levels for users to examine. The raw contains all annotations for all genes from all databases. The distillate gives an overview of all the functional genes which are present across all genomes. The product gives an interative heatmap allowing the user to explore a high level overview of the functions a acollection of genomes are capable of. You can check out what the product looks like by clicking here.

In order to make DRAM available to a wider audience we have also developed a series of KBase apps. KBase is a cyberinfrastructure developed by the Department of Energy with the goal of making 'omics analysis available to everyone. You can read more about the project at www.kbase.us. The DRAM KBase apps allow users to access DRAM through a web interface and without needing access to powerful servers. The DRAM KBase apps also allow for integration into the larger KBase system of tools including using DRAM annotations in flux balance analysis simulations.

DRAM recieved over 20 citations in the first year it was published. You can check it out in Nucleic Acids Research. We also had the opportunity to present DRAM as part of the Ohio State University Center of Microbiome Science webinar series which you can watch online here. DRAM development continues on it's github page where there is also a wiki full of helpful documentation.

In 2020 I had the opportunity to act as an instructor for the 2020 DOE Environmental Molecular Sciences Laboratory summer school. This is a program put on each summer by EMSL with a different theme every year. The 2020 edition focused on microbial dynamics modeling. Ordinarily this program is run in person but this was not possible due to the COVID-19 pandemic. As a result it was taught online and the general lectures were opened up to anyone who wanted to attend while the small group projects were run online. I taught a module about genome annotation and how DRAM can be used to annotate gene function. The talk I gave is available on youtube. Additionally I build the KBase narratives that were used by the students and led a student group during the event. As a cap to this work I presented a talk at the 2020 AGU fall meeting about the discoveries we were able to make in collaborations with the students who participated.

An important part of my PhD thesis was developing methods for correlation network analysis with compositional data. Compositional data, such as 16S amplicon sequencing or untargeted LC/MS metabolomics, needs to be treated differently than count data. Methods to take compositionality into account have been developed but are not used in many of the most popular tools for building and finding modules in correlation networks. To address these issues I developed a tool called SCNIC. SCNIC builds correlation networks from compositional data, as well as other data types, and then finds modules of correlated observations, via user selected methods. The SCNIC manuscript is available as a preprint. To make SCNIC more useful for the 16S rRNA amplicon sequencing community SCNIC is also available as a QIIME2 plugin.