The ligase chain reaction (LCR) was evaluated as an amplification method for an in vivo mutation assay. Specifically, the ligase was tested for its ability to selectively amplify a DNA sequence mutated at a single base, in the presence of an excess of wild-type DNA. As a model template a 370-bp DNA fragment of the mouse Ha-ras protooncogene containing an A to T mutation at the second position of codon 61 was used. With the commercially available ligase Ampligase (Epicenter), 250 molecules of mutant fragments could be detected by an enzyme-linked immunoassay with digoxigenin marker (giving a theoretical detection limit of 1 target gene per 10(4) copies of genome). In the analysis of mixtures with corresponding wild-type DNA fragments, a 1:1 mixture resulted in a clearly stronger signal than control samples lacking wild-type and mutant DNA. However, the signal obtained from a 100-fold dilution of the mutant DNA with wild-type DNA could not be distinguished from the background noise. In this particular form, LCR lacks sufficient selectivity to be applied to an in vivo situation, where the ratio of mutant to wild-type DNA sequences might be expected to lie around 1:10(6).