You are correct. Without a cross section of natural plants from across the range of a species, with samples taken in situ, and without a comprehensive analysis, DNA studies cannot be done correctly. The results will always be skewed.Don't blame the tool Brucher, blame the workmen using it.
There is huge merit in DNA studies but they should be replicated using different parts of the genome and chloroplast DNA, not just one sequence from one area, especially in plants that can hybridise or are possibly recently diverged. Its results should also be interpreted alongside other evidence, not completely excluding it, imho.