A Summary of Dr. C. Glenn Begley's Presentation

On November 14, 2017, the Office of the Vice President for Research and Office of the Provost hosted an important Research & Academic Development Seminar entitled, "10% of the time it works every time!" or Rigor & Reproducibility in the Laboratory with C. Glenn Begley, M.B., B.S., Ph.D., CEO BioCurate Pty, Ltd. from Victoria Australia.

Dr. Begley gave a fantastic presentation that spurred exciting discussion around the topic of quality research. Some of the key points from his lecture are as follows:

• Disease is a difficult system to address because 'biology sets the bar.' However, the system by which we study disease is much easier to address and should be done.
• Far too often, the scientific community accepts low quality research techniques and this has become a systemic problem. Individual investigators don't necessarily set out to distort data, but rather a lack of rigor in the lab create results that are often not reproducible.
• In an investigation conducted by Dr. Begley and his team of 52 high impact research articles published in top-tier scientific journals, (i.e. Science, Nature) they found that the results of 47 of these studies were not reproducible.
• Dr. Begley identified several tests to identify whether a study design's results are reliable. Some of the major tests are as follows:
  • Was blinding used?
  • Are all results shown?
  • Were reagents validated?
  • Were both positive and negative controls shown?
  • Was the analysis of the data appropriate? (i.e. cell growth shown on a log scale, not a linear scale)
• If a study fails one of these tests, it will almost always fail multiple tests, leading to a loss of reproducibility.
• These flawed studies have profound impacts including opportunity cost, clinical study cost, pointing public policy in the wrong direction, and a waste of taxpayer dollars.
• A large part of why these tests are sometimes not used is because there is a bias to publishing positive, groundbreaking results, often resulting in both investigators and journals to not follow all guidelines for a studies' design. There is a movement for researchers to be able to publish negative results as well.
• However, journals are not only to blame. Research institutions, as well as the individual investigator, also have the responsibility to demand good institutional practice.
• One way to handle these research errors is to host meetings for researchers akin to 'mortality meetings' held by doctors when a patient dies. Lab notes should be reviewed, proof of blinding shown, and guideline compliance should be ensured. There should be annual training, sharing of data, and consequences for bad research practices.

In closing, Dr. Begley stresses that these research errors are systemic problems and that the solution lies in offering researchers the proper incentive to change. And more than anything, the research community must always: Be Skeptical!