By: Marisa Boch, OUR Peer Research Ambassador
“I have not failed. I’ve just found 10,000 ways that won’t work.” – Thomas A. Edison
Full disclosure: the research process can be tedious. Progress may be slow, with setbacks seemingly more frequent than advances. Nevertheless, it is important not to get discouraged and, instead, to look at failure with a critical eye. Be open to what the results suggest. Troubleshoot. Use this failure to pave the way for discovery.
For me, my most notable experience with research setbacks occurred this past summer. I had received a Summer Undergraduate Research Fund (SURF) Award to work full-time on my independent research project, and I had high expectations for the progress I was going to make over the course of those 11 weeks. I wanted to achieve X, Y, and Z, and I had a clear-cut timeline for meeting those goals. In the end, however, I only achieved X (after failing several times over) and half of Y (hence why I am using this experience as an example of “setbacks in research”). Here is my story:
Glorified Expectations
The key to success in research is setting reasonable goals. If you are working on an independent project, discuss your project scope with your research advisor, a faculty member, or an upperclassman who has experience with independent research. In most cases, what you hope to achieve may be overreaching, and they will tell you that you should plan on meeting two-thirds, or maybe even half of your project goals. And that is okay!
When I was preparing my SURF proposal last December, I put together a timeline detailing my anticipated research activities for each week of my SURF summer. My faculty advisor and I knew that the timeline and project scope were ambitious, but, if everything went right, we still believed it could be attainable. Key phrase: “if everything went right.”
Attempt 1: Close Enough?
About five weeks into my SURF project, progress seemed to be going as planned. My first goal was to determine the concentration of the tumor necrosis factor alpha (TNF-α) necessary to inhibit the wound healing capacity of Caco-2 colorectal adenocarcinoma cells. Over the course of four weeks, I had lifted Caco-2 cells, seeded them on a 12-well plate, grown them to confluence, treated experimental groups with varying concentrations of TNF-α, and serum-starved the cells. This month’s work – a whole month! – culminated in a 12-hour wound healing procedure on July 14th. After scratching the cell surface, I imaged the wound every 2 hours for 8 hours. Based on the time-lapse images, it appeared that samples treated with the higher concentration of TNF-α displayed a slower wound closure rate, as desired. I was hopeful! When I quantitatively analyzed the images, I still had reason to hope that there would be a statistically significant difference between treated and untreated samples – the bars on the bar graph looked different. When statistical analysis outputted “P>0.05” and “Not Significant,” I wanted to say, “close enough,” and move forward with the next steps. I was reluctant to accept the results because, in all honesty, I dreaded repeating that treacherous procedure. Once I overcame this initial frustration, I accepted that there were flaws in the experimental procedure that were causing a large degree of variation within experimental groups. I knew I had to attempt to troubleshoot.
Attempt 2: All My Cells are Dead
Once again, I embarked on the weeks of prep work and was ready to repeat the 12-hour wound healing procedure by July 28th. This time, I was testing different methods of inducing the wound and different concentrations of TNF-α in an attempt to limit variability between experimental samples and ensure that the impaired wound healing effect would be observed. I struggled with the new method of inducing the scratch and also with imaging the wounds, causing the cells to be out of the incubator for a prolonged period of time. By about Hour 2, I came to the dreaded realization: my cells are dead. For nearly half of the samples, there was basically no wound to image, as all the cells surrounding the scratch had died. When I finally left for the night at 9pm, I knew that the results would be inconclusive. I was incredibly frustrated and decided to wait for feedback from my advisor. As suspected, we agreed that there were still flaws in the experimental procedure and that I should try again.
Attempt 3: Third Time is the Charm
By August 18th, I was ready to repeat the lengthy procedure again. The new semester was starting in a week, so this was my last chance to get results before the end of the summer. In an attempt to ensure that my cells did not die, I staggered the time frame for inducing the scratch and imaging the samples. I ran back and forth between the incubator and the microscope for 8 hours. By the end of the night, it appeared that my hard work had paid off – my cells were alive, and the 10 ng/mL concentration of TNF-α seemed to have an inhibitory effect on wound healing. The following week, statistical analysis finally confirmed this observation: P<0.05! I had finally developed an optimized, repeatable procedure that allowed for consistency between experimental samples, and I had determined the concentration of TNF-α necessary to simulate wound healing in the inflammatory state. After 3 months of work and two failed attempts, I am now ready to proceed with the next steps of my long-term research project.
Here are my major takeaways from my own experience with failure and troubleshooting:
- Set reasonable expectations. Be prepared for a procedure to require multiple attempts.
- Do not cut corners. While research protocols may be time-sensitive, one must balance speed with efficiency in order to achieve accurate, reproducible results.
- Do not be afraid of a procedure. What ultimately led to my success this summer was increasing the complexity of my research protocol. Some days in the laboratory may be more challenging than others, but feel free to treat yourself when you survive those 12-hour-procedure days!
- Be open to what the results are telling you. Hypotheses are not always supported by data, and it is important to draw conclusions from the results and not from what you expect the results to be.
- Be honest with yourself. Evaluate what aspects of the research protocol may be causing the setbacks. Is it your execution of the protocol or the protocol, itself? Think about how you can make improvements to the experimental design.
- Be confident in your abilities. A “failed” experiment is not a reflection on yourself as a researcher. Use your critical thinking skills, background knowledge, and available resources (including a research advisor/mentor!) to evaluate whether the results are accurate or the consequence of experimental error.
Marisa Boch is a senior majoring in Chemical Engineering & Molecular and Cell Biology. Click here to learn more about Marisa.