“…There have, however, been some estimates of the mutation frequencies in coding microsatellites of the genes contributing to tumor development (target genes) (Boland et al, 1998;Perucho, 1999a, b;Duval et al, 2001Duval et al, , 2002Zhang et al, 2001;Suzuki et al, 2002;Woerner et al, 2003), which appears to depend largely on the length of the repeat (Kunkel et al, 1994;Chen et al, 1995). According to statistical analyses describing the relationship between the mutation frequency and the relevance to tumor development, the frequency of the mutations in coding microsatellites in bystander genes (not related to tumorigenesis) are no higher than the frequency of mutations in non-coding microsatellites of the same length and type (Duval et al, 2002;Woerner et al, 2003). The frequency of mutations seen in many noncoding repeats that are less than nine mononucleotides, has been shown to be lower than 6% (Suzuki et al, 2002) and so the mutations seen in the RBM35A gene, for example, are less likely to be due to a bystander effect, as this gene was mutated in 25% of primary tumors with MSI (Table 2 and Figure 5).…”