A Systematic p53 Mutation Library Links Differential Functional Impact to Cancer Mutation Pattern and Evolutionary Conservation

A quantitative model to predict pathogenicity of missense variants in the TP53 gene

“…As shown in Figure , missense variants within the DBD generally resulted in more severe abrogation of protein function, compared to missense variants affecting the NH 2 and COOH terminal domains (transactivation and tetramerization domains, respectively), although five missense variants affecting the DBD displayed activity comparable to the wild type form of the protein. Of the 15 DBD variants also assayed by Kotler et al (), functional results were completely consistent in direction (five retaining function, 10 disrupting function). The correlation between the performance of the clonogenic assay and systematic transactivation assays (Kato et al, ) was moderate, with an R 2 of 0.58.…”

“…Variants classified as pathogenic or likely pathogenic by the model were associated with earlier age of cancer onset (including a greater proportion of pediatric cancers), and were enriched in Classic LFS probands and depleted in a cohort of cancer‐free women over 70 years. When compared to the output of reported functional assays, the majority of variants classified as pathogenic or likely pathogenic were nonfunctional in transactivation yeast assays (Kato et al, ), showed disruption of antiproliferative function/LOF (Giacomelli et al, ; Kotler et al, ), or reduced activity in the mammalian clonogenic survival assays developed in this study. Variants classified as benign or likely benign variants showed the same trend in the opposite direction.…”

“…When compared to data from DBD region variants assayed for antiproliferative function using a high through‐put mammalian assay (Kotler et al, ), disrupted function was observed for a high proportion of DBD variants classified as P/LP (333/434, 76.7%), and retained function was observed for 159/165 (96.4%) DBD variants classified as B/LB by the quantitative model. When compared to other systematic screen data in human cells (Giacomelli et al, ), 311/418 (74.4%) of unique variants classified as P/LP exhibited LOF, as opposed to only 27/244 (11.1%) variants classified as B/LB.…”

“…The outputs of the model were also compared to three datasets from human cellular models. In a recent study (Kotler et al, ), a massively parallel mutational scan was used to measure the functional impact of ~10,000 TP53 variants in the DNA Binding Domain (DBD) in vitro, deriving a relative fitness score for each variant (RFS). High RFS scores corresponded to variants that strongly disrupted wildtype‐p53 antiproliferative functionality, whereas lower RFS scores represented variants that retained antiproliferative function.…”

“…In addition, the data includes extensive information on the functional characterization of TP53 sequence variants , including the comprehensive mutagenesis of all possible amino acid substitutions caused by single nucleotide substitutions and subsequent functional analysis of transactivation activity in yeast (Kato et al, ). Further, systematic approaches in human cells to study distinct mutant TP53 features have been published recently (Giacomelli et al, ; Kotler et al, ). Despite these resources, there is currently no systematic approach to interpretation of variants in TP53 that integrates these data sources in a quantitative and probabilistic model.…”