Last week I was reading an article published in Human Molecular Genetics back in 2003. Besides other interesting things that the authors found, there is one figure in the paper, showing an example of intron loss/gain between human and mouse. I looked into the supplemental materials and found a few more cases of intron loss/gain between human and mouse that the authors claimed. I got excited. Sadly, when I checked more carefully at those genes' annotations (e.g. through Ensembl), I realized that the claim of intron loss/gain is simply due to incorrect assignments of orthology.
A bit disappointed. But this is indeed a good lesson to learn.