Does this mean that of the two characteristics ("study in mice" and "is a preprint"), "is a preprint" is more predictive of the result failing to pan out in humans?
Note: this happened to me, and there is google maps for zebrafish genes, this is easy: https://zfin.org/search?category=&q=SHC
You never know what kind of questions they might ask! Papers do in fact get rejected for shoddy work, or get sent to revision. Often it's not a huge problem, maybe fix a figure, double check some other work, or refer you to someone else's study that contradicts your claim. If you're not a fraud, you generally bring your strongest case, so the only tuning needed is details.
That being said, peer review doesn't always work, as there are cases of publishing fraud.
However, for labs that get exceptional scoops you can bet people are going to try to replicate it and if it turns out it doesn't work you've got some explaining to do. (Your academic reputation is now on the line.)
Here's a short list of notes on retractions: https://en.wikipedia.org/wiki/Retractions_in_academic_publis...
In my career (10y), I personally witnessed two people immolate their careers over publishing papers that didn't work or stole results from someone else.
Fun retraction watch site here: https://retractionwatch.com
To directly answer your questions:
Mice (and to a greater extent, animal models) are the current standard for advancing studies in biology. Are there problems with mouse models? Sure. Is there currently a better alternative? No.
Are preprints bad? No. It used to be there was no such thing as a preprint, you submitted your work to a journal and that was that. Accept or reject. You might have sent the preprint version to people in your field you were collaborating with, or some deconstructed version to verify parts of your work. Now with arxiv style publishing, it's up to the community at large to figure out what's what. I'm not saying one or the other is better. Us biodorks are catching up to physics people. However, anyone can publish to arxiv, including kooks.
Can preprints be bad? Yes. This work has no scrutiny to its claims by anyone outside of the authors, so for laypeople not in the field we don't have a good insight as to make heads or tails of these claims. Could you look it all up? Of course! Contrast that to someone in the field that might say on sight: "Yeah, Nichols' work has been pretty out there" or "That gene doesn't work the same way in zebrafish." :D
Mice are uncontroversial. Preprints are uncontroversial. I don't think I'd put money on either. I would be more interested to see follow up work than change the animal model.
PS: Wow, I had good coffee this morning!
These first two make it possible to do repeatable studies at relatively low costs. The latter allows you to do things like knock out or replace a gene and see see what impacts it has biologically and you can find out in weeks instead of years. Research mice the way biologists reduce their build times.
There is a piece of truth in that "animals models don't always pan out", but that's the state of the art whether we like it or not.
My point is, if you start the criticism at "mice model, opinion rejected", you are rejecting nearly all basic bioscience.
"bioRxiv posts many COVID19-related papers. A reminder: they have not been formally peer-reviewed and should not guide health-related behavior or be reported in the press as conclusive."
You're free to check out the credentials of the people in the paper, which look great. A bunch of different people that have worked in COVID, virology, and vaccines.