Interesting question. I sincerely doubt you can foresee whether a molecule is mutagenic (I guess you are interested in human beings, because things are completely different in mice, sometimes).
There are classes of compounds which are known to induce mutations (such as polycyclic aromatic hydrocarbons), but I believe these properties were discovered through observations on patients.
Not to mention metabolites: when you design a drug, you carry out time-consuming trials to assay that.