Science Meltdown Part III?

Since the existence of modern science, the replicability (others get the same results) of experiments and studies was always a problem. And of course the increasing complexity of science does not help to prevent errors nor does the complexity of the human mind or body, when studied.

Still there seems to be an awful lot of studies that are not replicable, and not always can the reason be found in errors of the experiment (see also the Wikipedia article on replication crisis).

Now it seems that in the already hard to do cancer science there is a decade old problem that could render up to 33’000 studies more or less useless. (They may still be valid, but who wants to sieve through them to see if they are?)

The reason is the wide use of contaminated cell lines.

pic CC0

Most cells used in research are going back to only a few hundred “donors”. To do your research you have to have the right cells. It is easy to see that is important to know if your cells are from a heart or from a lung, if you want to study reasons for lung cancer.

But since nothing is perfect, some of those cell lines have become contaminated by other cells or have contaminated other cell lines themselves. Surprisingly this is not really new, but why does nobody know about this?

"Employees at [biomedical cell distribution] centres recognise the problem, but claim no one will listen to them," says Halffman.
"Sometimes it involves semi-private companies that refuse to disclose anything for fear of reputation or financial damage. The biggest factor by far is pride and fear of reputation damage."
Another contributing issue is pressure to publish, with researchers not having the time or money to verify their cell cultures adequately before they begin their research.
Of course, the tendency to skip that all-important verification is only something that worsens the broader reproducibility crisis plaguing science.

Still, it is important to keep in mind that, while this may be very bad, it does not make every research useless. What we need is a better error management: not blaming and bad-mouthing, but the honest try to make things better the next time. That is science.

And just keep in mind that even the best study may just have lost to a very unfriendly Random Number Generator that some people call life.

source

Click for the German version of this post at my German account.