Time and time again poorly conducted studies have linked autism to vaccines. The studies are harmful. They encourage people to stop their children from getting potentially life-saving vaccines, putting them at risk of contracting entirely preventable illnesses unnecessarily.

Enter a new study which claims that autism symptoms in mice are linked to aluminum in vaccines. The study, published on September 5, stated that aluminum caused symptoms "consistent with those in autism" in mice. 

After the study was published, articles claiming that there's a link between autism and aluminum in vaccines began circulating the internet almost immediately, influencing an unknown number of people into not giving life-saving vaccines to their children.

_{As always with these stories, the damage may already have been done. The story has already been shared countless time by ant-vaxxers. A retraction won't reach the same people.}

But yet again the link is proving to be false, and the study extremely questionable, after one of its own co-authors claimed that figures in the paper were deliberately altered before publication. The data had been tampered with.

As with any bold claim like this, the study came under close scientific scrutiny after it was published. David Gorski, a professor and surgeon at Wayne State University, called the paper "the torture of mice in the name of antivax pseudoscience".

But of particular note was the PubPeer community, who spotted something odd about some of the images used in the study. The community, which allows users to analyze and comment on scientific papers, spotted that certain images appeared to have been manipulated. 

Some of the images appeared to have been flipped. I.e. they weren't showing what they were supposed to be showing.

After seeing the allegations on PubPeer, Dr. Chris Shaw – one of the paper's co-authors – told CBC News that his lab ran its own analysis on the figures. After doing this, he asked for a retraction of the paper from the Journal of Inorganic Biochemistry, where it was published.

"It appears as if some of the images in mostly what were non-significant results had been flipped," Shaw told CBC. "We don't know why, we don't know how... but there was a screw-up, there's no question about that."

Shaw told CBC that his lab doesn't know how the figures were allegedly changed as they no longer have the original data, and their analysis was performed on compiled rather than raw data. He stated that based on viewing the data years ago and reviewing his subsequent analysis, he had believed that everything was fine with the data and that it was a "good question" how the dodgy figures weren't spotted before the paper was published.

But it's a question that might not get answered. To put icing on this terrible cake, the author also claims the original data is now "inaccessable", and is "stuck" in China with an analyst who worked on the paper. 

This is the second retraction for the co-authors Christopher Shaw and Lucija Tomljenovic, according to Retraction Watch. Another paper, published in 2014 and listing both authors, was also retracted. The University of British Columbia told The Globe and Mail that they will investigate the allegations of misconduct in the 2017 study.

Dr Shaw said he thinks his conclusions in the study still stand, but because the researchers don't know, they felt it was best to withdraw the paper. As for the future, he told CBC News:

"I'm honestly not sure at this point that I want to dabble in [vaccines] anymore."