Materials and surface finish make a big difference, too. Torque is a very complicated issue, and specifications are best based on test data and past experience with similar joints.
If testing isn't practical, I'd go with 60-80% of the dry spec depending on the application and type of lubrication.
Even if there is no stretch on a bolt it can break from the twisting force trying to tighten it because even a bolt that is too long(bottomed out in a blind hole)can break from too much torque.