Rigor starts with lab metadata organization
This is an important consideration in our work. Rigor and reproducibility are well-known problems in biomedical research. We are interested in ways to improve rigor through data science and automation while maximizing experimental design. Over the years we have developed many lab approaches to enhance rigor. At the basis of these approaches is our metadata lab management system, which contains all our lab’s information. We have been running this system for many years, and we are sharing how to set up a basic system in this paper.
There are a few guiding principles of our approaches:
- All lab information should be digital (leave those trees alone).
- Any piece of information should be input only once (into a central database).
- The information must be readily accessible (with only a few clicks).
- Experiments should be executed with predefined protocols that are run by computers (automated), if possible.
- Analyses should be coded (scripted) from start to end (keep up that spaghetti programming), which makes them reproducible on demand.
- Compliance (IACUC, Drugs, etc) should be built into digital lab information systems to focus on the science, not the bureaucracy.
A central system for metadata management
A server, SQL databases, and a wiki are used to store and access all animal and specimen basic information, including surgeries, controlled drugs, chemicals, materials, computers, drives, data file links, scripts, how-to protocols, compliance, ordering, etc. This assures that any questions about “what was done” or “what is happening now” should be answerable with just a few clicks. Data files are accessible with a single click. You can take a peek at a simplified system here. If you want to set this up for your lab, follow the instructions in our Bio-Protocol paper.