Sugar structure sleuth Sébastien Vidal is on a mission to fix errors in the scientific literature

If Sébastien Vidal sees a mistake in a published chemical structure—a missing double bond or a hydroxyl group drawn in the wrong configuration—it “hurts my eyes immediately,” he says. Vidal, an organic chemist at the French National Center for Scientific Research, researches the syntheses of novel carbohydrate syntheses and prevention of bacterial infections. About once a week, he says, as he reads new papers in his field, he comes across an error in a published structure.

Vidal routinely flags the errors to help prevent people from copying and propagating incorrect structures. He points out the mistakes to the authors and journal editors so they can be corrected. Since joining Twitter in 2018, he has also tweeted the errors, using the hashtag #vidalized to bring attention to them. Louisa Dalton talked with Vidal about why he’s so passionate about correcting sugar structures. This interview was edited for length and clarity.

When did you start pointing out errors?

I started by chance in 2013. I was confused by the configuration of a sugar in a paper. The authors drew the structure as an α anomer. In fact, the molecule took the form of the β anomer, according to the experimental NMR data. I talked to the author, and they changed it.

Maybe from then on, I was paying more attention. I scan tables of contents by looking at abstracts and titles to keep up to date. In 8 years, I’ve found errors in 350 papers just by chance. That’s one per week, basically. But I’m sure this is the tip of the iceberg.

Where did the Twitter hashtag #vidalized come from?

#Vidalized came from Fraser Stoddart, actually. I did a postdoc with him, and he taught me that accuracy in figures is important; the publication stays forever. More recently, he told me about a retired professor at Emory University, Fredric Menger, who similarly pointed out errors in published data a few decades ago. People called it “mengerizing” papers. That’s where I said, “I should use #vidalized on Twitter.”

A lot of the errors relate to stereochemistry. Do you see those mistakes immediately?

It’s like if you see the word ambulance written the other way around. It’s very easy to me. I’ve been working on sugars for 20 years now. Working with something every day makes it natural to you.

For me, it takes 1 s to see a mistake and 1 min to check it. If I have a doubt, I take my model of glucose that I have on my desk and move it around and change it, because sometimes you have to play with conformation.

How do errors usually creep in?

The main mistake is flipping a structure horizontally or vertically in PowerPoint or ChemDraw. It’s easy to flip a molecule instead of dragging and rotating it, but then it becomes the enantiomer. That has consequences. Chirality is very important.

Also, very often, when you draw oligosaccharides, the component sugars bump into each other because of geometry. People sometimes switch the configuration of the alcohol groups on the sugars from equatorial to axial because they want to make a nicer drawing. They don’t notice that if you switch from equatorial to axial, you’re making a different sugar.

Usually, it’s a simple mistake. A student can make a mistake preparing the figures. The main author, the referee, and the editor all might not catch it. In the process, everybody’s responsible.

And sometimes—this happens very often—one mistake in an original paper is wrong, and then the review or the next papers use the same mistake. To me, a review is like the Holy Bible. If you write a review, and you take a paper where there is a mistake, you should fix the mistake in the review. You shouldn’t copy and paste the same figure. Then the mistake is transported to the next paper and so on.

How do you think these mistakes could be reduced?

I think journals should be more proactive. In my opinion, each journal dealing a lot with carbohydrates or natural products or anything involving a lot of stereochemistry should have a structure checker. That could be one thing. Or there might be an automated way to check with a machine-learning system. It would take time and effort to develop but could be useful.

In terms of papers and journals, it’s a simple fix: it’s calling for a correction. I flag it, and more than 90% of authors then make the correction immediately and are very happy to do it.

But errors in other places could have more implications. How clean are the databases? How are journals transferring the data from the papers to the databases for online structure searches? Also, I don’t read patents. Is a patent still valid from a legal point of view if you draw the compounds the wrong way?

It’s our job in science to be very rigorous and accurate because otherwise, it doesn’t mean anything.

Louisa Dalton is a freelance writer based in Virginia who covers chemistry. A version of this story first appeared in ACS Central Science: cenm.ag/vidal.