We have developed and clinically tested a rapid and largely automated procedure to detect mutations in the coding region of a gene of interest. Our method relies on the automated sequencing of the complete cDNA, followed by an advanced mutation search-and-verification routine using an integrated set of computer analysis tools. We have applied our automated procedure to the diagnosis of familial hypercholesterolemia (FH) in 52 unrelated FH families, by sequencing the whole cDNA coding region of the LDLR gene. Here we report the procedures and performance of our method in the identification of the most common types of LDLR mutations: short deletions or insertions and point mutations. Our method can provide a standard procedure for the 'overnight' unequivocal identification of mutations in those genetic diseases where several different mutations, none clearly prominent, may affect a given gene.