Mutations in the gene encoding beta-glucocerebrosidase are the main cause of Gaucher disease. The identification of some of these mutations in prenatal tests is a good complement to enzymatic assay and allows diagnosis and, in some cases, prognosis of the disease to be made. DNA analysis is particularly useful for carrier detection since the results of biochemical analyses are often ambiguous. The main drawback of mutation analysis for prenatal diagnosis and carrier detection in Gaucher disease is that rare mutations account for more than 30 per cent of the mutant alleles in most populations. The individual detection of these mutations is too expensive and time-consuming for routine use. Here we present a diagnostic protocol based on co-segregation analysis, using highly polymorphic markers, to be applied when at least one disease allele does not correspond to the most common mutations. Because of the frequency of the N370S mutation and its relevance for prognosis, an improved PCR detection method is included.