Enthusiasm for understanding how changes in the human microbiome affect health has led to an influx of researchers who have little experience in documenting and performing bioinfomatic analyses. Compounding this problem is that over the past 10 years microbiome datasets have grown from hundreds to millions of sequences. With the increased size of the datasets the complexity of the analyses has also grown. Researchers that once used spreadsheets to analyze their data now struggle to use command line tools. Traditional training programs have not been able to meet the needs of these researchers and so it is essential that instructional materials be developed to train these researchers on the best practices for documenting and disseminating their analyses so that they can be reproduced by others. The objective of this proposal is to develop a training module that microbiome researchers can use to improve the reproducibility and overall quality of their research. Aside from the general importance of insuring that all research is reproducible, the significant growth of the community makes it urgent that such instructional materials are developed now.
The specific aim of the proposed effort will develop a set of autotutorials to teach microbiome researchers habits for engaging in reproducible research.
This aim will be achieved through an iterative process of development, evaluation, and refinement using a wide network of microbiome researchers to assess the materials prior to broader dissemination. The expected outcomes of the proposed modules are the improvement of the reproducibility of research within the microbiome research community, increased accessibility to raw original data, and a greater use of literate programming tools for constructing manuscripts and oral presentations. Furthermore, if it is possible to improve the reproducibility of the original research, then it wil be more likely that other researchers will use those data, results, and methods to perform additional analyses resulting in a greater understanding of the microbiome. Such a process is rare within the microbiome literature given the vast size of many of these datasets. The long-term goal of this project is to establish a broader community- supported resource devoted to disseminating best practices in performing reproducible microbiome research. Given the significant role of the microbiome in human health and the significant growth in our understanding of how it shapes health and disease, improving the reliability of the results from these studies will have a meaningful positive impact. This project will yield a significant vertica step in the field because it will put tools into the hands of researchers performing microbiome-focused studies empowering them to perform sophisticated and reproducible analyses. The approach taken in the proposed research is innovative because it represents the first concentrated effort to develop formal, public, and open training modules directed at the microbiome research community. Finally, we anticipate that the materials we develop for the microbiome research community will be easily disseminated across other bioinformatics research disciplines.
The proposed research is relevant to public health because it supports researchers within the domain of microbiome research, who have shown that it is impossible to separate human health from the structure and function of the human microbiome. Thus the research is relevant to the part of NIH's mission that pertains to the development, maintenance, and renewal of scientific resources that will assure our ability to perform robust and reproducible research in order to prevent disease.
| Schloss, Patrick D (2018) Identifying and Overcoming Threats to Reproducibility, Replicability, Robustness, and Generalizability in Microbiome Research. MBio 9: |
