Missense base substitutions may or may not result in a functionally altered protein. Interpreting the biological effects of amino acid (AA) substitutions in cancer-related genes is critical in a variety of contexts. Clinical cancer geneticists often must decide whether a previously unknown allelic variant causes disease. Reliable functional assays usually are not available. The objectives of the proposed work are to develop and test a model for predicting the biological consequences and clinical relevance of missense mutations in the p16 gene by detailed study of evolutionary substitution patterns and protein structure. Aligned AA and nucleotide sequences are often compared to infer information about protein function. Detailed computational analyses are rarely performed on genes associated with human diseases, but preliminary data indicate they can improve upon simple sequence alignment in predicting function.
Specific aims are: 1) To collect p16 evolutionary and mutational data sufficient for refining a model that predicts the functional consequences of AA substitutions; to clone and sequence new p16 sequences in order to expand the database so that it is large enough for sufficiently detailed computations. 2) To test in vitro the cell cycle arrest and cdk binding of missense variants of p16. Logistic regression will be used to establish a computational model that predicts loss of function. Initial calculations of high conservation correctly predicted mutant function in 75-80 percent of tested codons, and low conservation correctly predicted wild type function in 85-90 percent. We predict that evolutionary parameters and structural features provide independent information for predicting functional changes. 3) To integrate computational and laboratory data in differentiating the biology of p16 and related proteins. p16 is used as a prototype for study because of its role in Familial Melanoma; missense mutations occur whose functional consequences are unknown; reliable functional assays exist; and a crystal structure is known, so data for mutational spectrum, evolution, structure, and function can be correlated. These studies should be generalizable to the interpretation of mutations in other cancer-related genes and to other single nucleotide polymorphisms (SNPs) found throughout the genome.
| Miller, Peter J; Duraisamy, Sekhar; Newell, Joan A et al. (2011) Classifying variants of CDKN2A using computational and laboratory studies. Hum Mutat 32:900-11 |
| Goldgar, David E; Easton, Douglas F; Byrnes, Graham B et al. (2008) Genetic evidence and integration of various data sources for classifying uncertain variants into a single model. Hum Mutat 29:1265-72 |
| Ollila, Saara; Dermadi Bebek, Denis; Greenblatt, Marc et al. (2008) Uncertain pathogenicity of MSH2 variants N127S and G322D challenges their classification. Int J Cancer 123:720-4 |
| Murphy, Joan A; Barrantes-Reynolds, Ramiro; Kocherlakota, Rama et al. (2004) The CDKN2A database: Integrating allelic variants with evolution, structure, function, and disease association. Hum Mutat 24:296-304 |
