That is an incredibly bad paper. They have gone on a fishing expedition across dozens of algorithmically detected differences, barely adjusted their statistical significance thresholds to make up for the fact that they're searching for something that might be "significant", and even then, the best they can do is come up with...50% of people have something. A coin toss.

This is the medical literature version of the xkcd green M&M comic:

https://xkcd.com/882/