It is certainly annoying, and I mostly agree with what genomax said. However, in practice, I have yet to see a dataset for which starting from the raw reads or the trimmed reads changed much in the analysis. If we assume that the trimming was done reasonably (do you have the trimmomatic reports to make sure of that ?), the benefit of resequencing 30 samples is very small, compared to its cost in time and money + potential batch effect. I would rather invest the ressources into confirming the results using an orthogonal approach.

These are just some thought, I don't know what I would do if I was in your place. And I never want to be.

(Sadly) even sharing the trimmed data is actually sharing much more information than a great deal of the studies out there...