Teaching microbiome data analysis through student involvement in actual faculty research projects (PPEM 497)
Report from 2017-18 Teaching Transformation and Innovation Grant
Terrence Bell, Assistant Professor, Plant Pathology and Environmental Microbiology, College of Agricultural Sciences
Traditional microbiology is a well-established field of science, but the study of microbiomes is not. Microbiomes are complex assemblages of microorganisms (bacteria, fungi, viruses, etc.) that occur in essentially every environment on Earth, including soil, volcanic hot springs, and the human gut. Microbiomes are involved in many processes that directly impact human and global health. Despite their importance, microbiome research has only been made possible in the last decade or so, thanks to recent advances in high-throughput DNA and RNA sequencing technologies. These technologies allow us to access genetic information from many microorganisms at once, including the vast majority that cannot be domesticated in a lab.
"This course taught me how to use…widely available software to research microbial communities, all in the context of microbial ecology and the latest research in this fast-growing field. Working with new and original data from faculty-provided samples made it realistic, and gave me the skills to use these tools and ideas in my own research." – Phillip Martin, PPEM 497 student
Each year, it becomes easier to acquire microbiome data, but the analysis and interpretation of these data is complicated. Because microbiome research is so new, there are few courses nationwide to prepare students for graduate work in this rapidly growing field. Most students enter graduate school lacking either the computational skills or background in biology that are needed to perform microbiome-based research projects.
The goals of PPEM 497 (Studying and Shaping Microbiomes of the Environment) are to introduce students with minimal computational skills, but a background in biological sciences, to the basic concepts of microbiome research, and to the freely available tools that are commonly used to analyze microbiome data. I came from a background in Ecology, and struggled to self-learn data analysis skills over the course of years. What I took away from that experience, and from mentoring many graduate and undergraduate students along the way, is that you need hands-on time (and lots of it) with data that matters.
"Everyone could explore the programs at their own pace, figuring out what worked and what didn't, and what resources were available to help when they didn't." – June Teichmann, PPEM 497 student
In this course, students are constantly working with data, with new assignments weekly. They progressively move from well-guided tutorials, to independent work with curated data, to open and creative work with newly generated raw data. These raw data come from large high-throughput sequencing datasets that are produced by the students, from samples provided by selected faculty in the College of Agriculture. Funds from the Schreyer Teaching Transformation and Innovation Grant supported these sequencing efforts.
This year, our two research projects were:
- Microbiome comparison of grapevine roots under different agricultural management regimes; Lead project faculty: Dr. Michela Centinari
- Variation in diseased and healthy mushroom cap microbiomes within a common production facility; Lead project faculty: Drs. Kevin Hockett and Carolee Bull (in collaboration with Drs. Fabricio Rocha, John Pecchia, and David Beyer)
The faculty leading these projects had little to no experience with microbiome analysis tools, and so student in-class work on these projects is essential for helping them generate data needed for upcoming grant applications. It also allows the research to go in new directions that it might not have otherwise. We are working now to turn these data into peer-reviewed publications that include some PPEM 497 students as co-authors.
"A really interesting part of this course was being able to analyze data generated from projects going on at Penn State. It made the programs less abstract, and gave me a better idea of how to use them to answer questions about my own data… It makes it easier to put hours of work into something if it's not just for a grade, or for the sake of learning." – June Teichmann, PPEM 497 student
Our format provides a real research experience, with real stakes. As an undergraduate, I remember performing the same experiments as everyone in my 200-person class, and wondering, "Why am I even doing this? They're about to give us the answer." In fact, they'd probably done the same experiments for the previous 20 or 30 years. In PPEM 497, students work with data that they invested in creating, and they have to get it right, because these data mean something. In addition, students may benefit further from their work by building their CVs, in the form of a publication authorship, or a formal oral presentation. I've asked two of the students to present on their projects at the Penn State Microbiome Center Seminar Series.
The course also allowed students to explore data analysis approaches creatively. Given the exact same data, students come up with a variety of ways to analyze and visualize those data, leading to different insights and understanding.
Collection of graphs produced by PPEM 497 students from the same dataset.
Student credits: Sarah Isbell, Phillip Martin, Pankaj Kuhar, Sofia Roitman, Siyi Ge, Amanda Mainello
In its pilot year, this course had 13 students, with plans to expand enrollment to 20 for each fall semester, ideally with a 50:50 split of upper-year undergraduates and beginning graduate students. The course has had great reception from faculty within the College of Agriculture, many of whom are eager to supply research projects and data to the class. Bell is in the process of applying to have this instated as a standing course (tentatively PPEM 440).
Going forward, Bell hopes to extend the impact of these research projects. By connecting them across years, they can potentially create large and unique datasets that are not possible to create within typical 2-4 year grants. In addition to serving students and selected faculty, such datasets could be of enormous benefit to the wider scientific community. All raw data from this and future projects, will be made available through free and publically accessible databases, allowing researchers from around the world to connect to the work of PPEM 497 students.