Fixing Mistakes – Putting Action Behind My Big, Fat Mouth

Yesterday, I wrote a post about scientific fact checkers and how scientists interact with them, using Dr. David Sabatini as an example. I did NOT expect the attention it garnered, but I do hope that Dr. Sabatini (if he read it) took it in the spirit in which it was intended. Like I noted in the original post, I’ve been following and admiring his work on mTOR for years. Thank in part to his efforts, mTOR inhibitors are in the clinic and are helping cancer patients fight their disease.

As a scientist, I find that very intellectually stimulating and gratifying. As a cancer survivor, I’m eternally grateful. Should my disease recur, torkinibs may be a part of my treatment plan, or perhaps third or fourth generation inhibitors. There are more options now thanks to the efforts of laboratory and clinical investigators around the world, and that gives us all so much hope!

At the end of the previous post, I noted that I, too, had been flagged on PubPeer for an error in a 2012 publication. What happened? In figure 4C, we inadvertently duplicated photomicrographs in the top panels:

Yup – the sharp-eyed PubPeer reviewer caught what the graduate student, collaborators, myself, my co-investigator, peer reviewers, and the journal production editors missed. The “Parental” and “Vector” images are identical. Not good. Now, this is a relatively minor error, but if left alone, it could lead to the perception that our laboratory group is sloppy and not as rigorous as we should be. In science, like many other fields, reputation is everything.

How does this happen? It’s a product of long hours poring over data, trying to select the best representations of experimental results, building figures and revising them…and revising them…and switching out panels and photos until the student or postdoc putting together the figures eyes are crossing. Many of them are sleep deprived, overloaded, and after a while, really numb to looking at the same data over and over again. Is that an excuse? No, it’s a reason, which is why study PIs, collaborators, reviewers, and the journal have a responsibility to double, triple, and quadruple check our papers before they’re sent out into the cyberverse for other scientists to read.

I take my part of the responsibility for this one. At the time, I was a Research Assistant Professor. While not on tenure-track, I was quite senior in the laboratory and had a duty to co-mentor and support the graduate students and fellows in the lab. My name was on the paper. I missed this and dropped the ball. That’s on me.

Fortunately, thanks to PubPeer, I have the opportunity to fix it!

After sorting through electronic files (thank GOODNESS our former lab members were organized), I was able to locate original images and generate correct (distinct) panels for Parental and Vector controls:

Submitted to PLoS One as a correction

I contacted PLoS One today to request a correction. They may not re-issue the paper due to time and cost constraints, but I do hope they’ll add the corrected figure panels to an addendum. I’ll report back on this once I hear from them.

Bottom line: We ALL make mistakes from time to time. Yet, our culture discourages us from owning those mistakes, as if they’re a mark of shame or weakness. Certainly admitting mistakes has become uncomfortable and taboo. That is something that needs to change. Owning mistakes and fixing them are signs of integrity. When I’m entering into new collaborations, I pay attention to how my potential co-investigators and their team handle mistakes. Admission and ownership of mistakes and efforts to correct them are signs that new collaborators are trustworthy, rigorous, and have integrity, essential qualities in modern science. Very few scientific studies are performed by a single laboratory/person in that laboratory. Shared labor is the norm, so trust is key.

I WANT to be known as trustworthy, as someone with integrity whose work can be trusted and reproduced. As someone who recently withdrew a submitted manuscript when we (to our horror) found out that many of the data weren’t reproducible, I’ve become more vigilant. And I’m a better scientist for it. Science is better for vigilance. And that’s a good thing.

Being a Great Scientist – How to (and NOT to) Handle Mistakes

In the age of the Internet, trolls and trollish behaviors have multiplied exponentially. Science is not immune. Many of us were late coming to social media, but more and more laboratories, institutions, programs, scientific organizations, conferences, professional organizations, and journals are engaging in Twitter and other outlets. Overall, this is a great thing! It is so satisfying to peruse science Twitter and catch up on all of the latest findings, hot new studies and publications, and enjoy a community of peers that extends beyond the halls of individual institutions.

But OF COURSE there are trolls and trollish folks in the Twitterverse. Smart people aren’t immune to this. It’s human nature. Not our best quality as a species, but it’s there nonetheless and it extends to people across professions, educational backgrounds, political and religious ideologies…you get the idea.

No matter where you go, there’s always one (or more) assholes. That’s a fact of life, like death, taxes, and motherfucking reviewer number two.

A positive side-effect of the Internet Age is the rise of transparency and fact-checking. This is important in every field (side-eyes the news media), but especially in science. Scientists pore over the literature and build upon previous studies to move the field forward. If a study contains errors, inconsistencies, or (in the worst cases), fabricated or altered data, it undermines the integrity of the field and the scientific endeavor. That’s why sites like PubPeer work to spot inconsistencies in published scientific literature. Now, no one likes to have their mistakes posted in big bold font on the Internet, but setting aside ego and emotion, it’s actually a GOOD thing! It helps us be vigilant about our work and the work of our colleagues who contribute data to shared publications, it allows the authors of the studies in question to go back and examine the data and correct mistakes in collaboration with the journal in question. It makes science more rigorous. All good things.


Link to Meme

The scientist in question decides to get his or her TROLL on!

When a mistake is pointed out on PubPeer, there are basically two ways to handle it. #1 – Thank the fact checkers for finding the error, dig through the data and fix the mistake, contact the journal in which the mistake was published and (if possible) update the data presented or add a note with the updated data as an addendum, and promise to be more rigorous in the future. This is the best way to handle it. It preserves your integrity as a scientist (and the integrity of the scientific process), it shows your commitment to ethics and responsible conduct of research, and it shows that you value your reputation and quality of your work enough to protect it. Basically, it makes you look good, honest, trustworthy – all qualities you want to have as a human being and a scientist.

Link to Tweet

Then, there’s the second way to handle it…#2 – Insult the PubPeer fact checker (petty, failed scientists), block “steaming turd,” (grudgingly) contact the journal to see what you should do. See this post from For Better Science for the run down. The first two items are childish, unnecessary, and, well, trollish. Mistakes happen, even to the great and mighty Dr. David Sabatini (whose work I actually follow and admire, since I’m working on mTOR signaling in breast cancer in my own lab). He’s a leader in the field, and he could have used this as an opportunity to set a great example for his peers and colleagues by handling this situation, which happens to plenty of scientists, with grace, dignity, and integrity. Someone with his reputation and in his position has a great deal of influence and sets the tone for scientists at all levels.

This is not a great tone. This is not a good look. I’m 95% certain that, had he not gone on Twitter and got his troll on, no one would be talking about the errors that have been found in multiple papers. He could have quietly fixed them and moved on.

Instead, EVERYONE is talking about this. This isn’t the kind of publicity you want as a scientist. Here are some Twitter replies:

There are even more replies to the first post, some showing sympathy and support – while gently pointing out the importance of the work PubPeer is doing – others (who tend to get blocked) calling out Dr. Sabatini on the insults borne of ego. It didn’t have to go down like this. It’s sad and unfortunate that it did. Hopefully, it will serve as an example of what NOT to do when confronted with evidence of error in published work.

For a GREAT example of what to do when confronted with evidence of mistakes in a publication, check out how Nobel Laureate Dr. Frances Arnold retracted a published study when she and her colleagues found the results were not reproducible missing data from a lab notebook. This is a fantastic example of integrity, honesty, and ethics.

*Note – you might be wondering if I’m just armchair quarterbacking, safe in the knowledge that I’m not the one being called out by PubPeer. I am not. In fact, when I joined PubPeer and did a search on my name, I found a post about an error in a 2012 PLoS One paper for which I was a co-author – we accidentally duplicated a photomicrograph. These things happen, but I’m GLAD the PubPeer fact checker spotted it so we have the opportunity to correct this. We are working on it NOW!

(And yes, I did thank the fact checker, because my mama raised me to have manners)