He's changed wild-type planarians to grow the heads of other species. It reverts after a few weeks, because the system has error-correcting mechanisms, but the DNA of these worms is unchanged.

He once compared tinkering with DNA as pulling out a soldering iron to fix a software bug.

In the case of morphology, DNA may not be the best level of abstraction. It's certainly possible, just as one can use chemistry for social problems, but for some problems, affecting cell-to-cell communication may be a more direct path.