Figure 1. Distribution of p53 mutations. Mutation frequencies collected in the IARC TP53 R10 database (18 145 nonsynonymous mutations) are plotted against the protein domains. The DNA-binding (p53DB) domain contains six residues considered mutational hotspots in cancer. Reproduced from Arbiza et al. (2006), Copyright Elsevier (2006).

Figure 2. Mapping of selective constraints in the p53DB domain. (a) The three-dimensional structure of the p53DB domain showing residues coloured according to different selective pressures. (b) Primary amino acid sequence and secondary structure elements of the p53DB domain. Residues in red, orange, yellow and green show the gradual distribution of the selective constraints represented by _SLR values. Residues in red (0<0.1) and orange (0.1<0.2) are generally associated to DNA contact sites (blue circles), Zn²⁺ contact (white circles) and sites where mutants are known to be denaturizing (red circles) among others. A few of the sites seem to be below the limit considered for selective constraints (yellow, 0.2<0.3). Residues where selective constraints were predicted to be low (green, >0.3) are distributed along the external regions of the core domain, and most of them are interspersed between sheets and helices. Arg²⁴⁸ binds in the minor groove of the DNA. Ser¹⁸⁵ was conserved in the cluster of primates and rodents, but was discarded in the analysis due to gap insertions in the basal species. *, phylogenetically conserved residue; +, SLR detected PFS at 95% confidence after correcting for multiple testing and !, as in +, but at 99% confidence. See text for a detailed explanation. Reproduced from Arbiza et al. (2006), Copyright Elsevier (2006).

Figure 3. Mutation frequency and distribution in p53. (a) The distribution of p53 residues in the -frequency space describe an ‘L-shaped’ curve where sites under selective constraints (low values) are preferentially associated to high mutational frequencies associated to cancer. Conversely, residues above the limit of the effects of purifying selection (high ) are preferentially associated to low mutational frequencies. (b) Mutational hotspots show high evolutionary constraints imposed by natural selection and the highest mutational frequencies. The cutoff value represents the maximum value for which residues were deduced to be under the influence of purifying selection at 95% or 99% confidence using the SLR method. This threshold represents the a priori hypothesis used to detect a statistical significance between mutation frequency and values using a large set of human disease genes. Reproduced from Arbiza et al. (2006), Copyright Elsevier (2006).

Figure 4. ·distribution, disease and polymorphism. More than 8000 amino acid mutations defined as disease and polymorphic variation in the Swiss-Prot database are clearly differentiated by selective constraints. The boxplot shows the median (horizontal bold line), the upper and lower quartiles (box) and the interquartile range (dashed vertical lines). For visual clarity a horizontal dotted line indicates =0.1.

Figure 5. Analysis of the selective constraints of cSNPs in the human genome. The PupaSuite web server provides a complete set of tools for SNP characterization to assist users interested in genotyping experiments. The selection strength acting on all the cSNPs of the human genes can be reported according to different thresholds. By default the functional analysis of PupaSuite reports cSNPs showing <0.1.