“…A recent study on a series of CDKN2A missense variants, however, showed potential limits of these approaches: widely conflicting results, in fact, were obtained, with several specific variants, that were paradoxically predicted to be benign or pathogenic, depending on the software used (Kannengiesser et al, 2009). Undoubtedly, in silico approaches provide an excellent support for UVs classification, but their accuracy is apparently limited by the available evolutionary, mutational or structural databases and, in some cases, by intrinsic limitations of each individual method.…”