Recently, two important advances have fostered a new era in biomedical research. First, we now recognize that humans, other animals, and plants are important ecosystems for microbial consortia, and that these consortia underpin their hosts' wellness. For example, we are just beginning to understand the role of human microbiota in mental health, immunity, and development. Second, advances in high-throughput sequencing technologies have provided cutting-edge experimental tools for observing the diversity and functions of microbial consortia. For the first time, researchers can grapple with the sheer diversity of microbial consortia associated with hosts, and also can begin to untangle how this diversity contributes to host wellness. Thus, a many biomedical researchers have generated immense, information-rich metagenomic datasets, hoping to realize the promise of these datasets to understand the intricate relationships between microbiota and host. Despite this promise, analyses of metagenomic data present a major challenge. Most biomedical researchers lack the computational, bioinformatic, and statistical training required for appropriate analysis, and also lack a working vocabulary to communicate their analysis needs to statisticians. This is especially concerning in human microbiome research because inaccurate or incomplete analyses can lead to erroneous interpretations that have implications for our approaches to preventative medicine and disease treatment. It also leads to generation of data that ultimately cannot be used to answer research questions because of inadequate statistical power or depth of sequencing in experimental design. We plan to address this need by offering an economical, one-week intensive course to train advanced graduate students, post-docs, and faculty in how to analyze microbial metagenomic data, from raw sequence handling to statistical analyses. Our integrated educational strategy addresses two related training needs. First, we offer general training in the fundamentals of effective computing so that participants will build computing skills needed to execute their analyses independently. We also offer specific training to overcome hurdles particular to microbial metagenome analyses. Participants will develop these skills via practical, hands-on tutorials motivated with real microbial metagenome datasets, and will enrich their learning by engaging in lectures and panel discussions with key leaders in the field. All of our course materials are continually improved and freely available on our course website (edamame-course.org) and disseminated on our GitHub repository. Participant learning will be assessed each year and materials iteratively adapted to best meet course objectives. We successfully ran this course in 2014 and received overwhelmingly positive feedback. Our course evaluation data shows that our educational strategy was effective at increasing skill level, confidence, and analysis sophistication among our participants.
To understand the implications of microbial consortia for host wellness, many biomedical researchers are generating high-throughput sequencing data of microbiomes, but are unsure as to the best approaches for analyzing these data. We propose to continue to teach an intensive, short course on how to analyze microbial metagenome data. Our integrated educational approach aims to help participants build a foundation in computing while learning state-of-the-science tools for metagenomics analyses, from raw sequence processing to statistical analysis and interpretation.
| Shade, Ashley; Teal, Tracy K (2015) Computing Workflows for Biologists: A Roadmap. PLoS Biol 13:e1002303 |
