Whole genome DNA-DNA hybridization has been used to identify bacteria in periradicular lesions partly because there is no amplification of the bacteria, therefore, minor contaminants are not detected. There are, however, potential pitfalls with this technique, including inability to distinguish dead bacteria, cross-reactions of species within a genus, and inability to detect species present in low numbers because of loss of DNA during extraction and purification. Alternatively, inadequate extraction and purification of DNA could result in false positives. Therefore, controls are required to monitor DNA loss, DNA cross-reactions, and DNA of pure cultures mixed with bacteria-free tissue to monitor for false positives.
We determined that the quality of DNA extracted from histological sections of periradicular lesions is excellent for DNA-DNA hybridization. Although lesions contain large numbers of bacteria, histological sections through lesions barely contain sufficient quantity of bacteria for such analysis. This was confirmed by histological observation of sparsely distributed bacteria within lesions. Furthermore, we found that the bacteria are not distributed evenly throughout periradicular lesions, in numbers or species.