As a spin‐off from fundamental molecular biological research, there has been a remarkable increase in new methods for diagnosis (i.e. detection and identification) in recent years. Because of their origin, these new methods all have in common that they use structural elements of the target organisms such as nucleic acids, lipids, fatty acids, proteins, polyamines and polysaccharides as a basis. These structural elements are either used as templates for development of so‐called probes for detection (and identification) or they are placed into man‐made patterns and used for identification/classification. The article presents the advantages and perspectives of the new methods compared with conventional ones. It may be noted that, in many studies, the specificity and reproducibility of the new methods has not been adequately treated or has even been only assumed. These features, which are closely linked with experimental and sampling error, lead to the principal pitfalls of the new methods, which are also rev