The paper introduces a new functional impact score (FIS) for predicting the functional impact of amino acid residue changes in proteins, particularly relevant in the context of cancer genomics. The FIS is derived from evolutionary conservation patterns, using combinatorial entropy formalism to analyze multiple sequence alignments of homologous sequences within and between species. The method performs well in distinguishing disease-associated variants from common polymorphisms, with an area under the receiver operating characteristic curve of approximately 0.86. In cancer, the method effectively identifies more likely functional mutations ('drivers') among the mutations listed in the COSMIC database. The authors also report a list of about 1000 top human cancer genes frequently mutated in one or more cancer types, ranked by their likely functional impact, and an additional 1000 candidate cancer genes with rare but likely functional mutations. The study estimates that at least 5% of cancer-relevant mutations involve a switch of function, rather than simple gain or loss of function. The FIS is validated through various tests, including comparison with known disease-associated and polymorphic variants, experimental data on TP53 mutations, and analysis of mutations in the COSMIC database. The method's ability to predict functional impact is demonstrated through its performance in separating recurrent and single mutations, as well as mutations in multiply mutated genes and annotated tumor suppressor and oncogenes. The authors conclude that the FIS provides a robust tool for prioritizing mutations for experimental investigation and therapeutic development.The paper introduces a new functional impact score (FIS) for predicting the functional impact of amino acid residue changes in proteins, particularly relevant in the context of cancer genomics. The FIS is derived from evolutionary conservation patterns, using combinatorial entropy formalism to analyze multiple sequence alignments of homologous sequences within and between species. The method performs well in distinguishing disease-associated variants from common polymorphisms, with an area under the receiver operating characteristic curve of approximately 0.86. In cancer, the method effectively identifies more likely functional mutations ('drivers') among the mutations listed in the COSMIC database. The authors also report a list of about 1000 top human cancer genes frequently mutated in one or more cancer types, ranked by their likely functional impact, and an additional 1000 candidate cancer genes with rare but likely functional mutations. The study estimates that at least 5% of cancer-relevant mutations involve a switch of function, rather than simple gain or loss of function. The FIS is validated through various tests, including comparison with known disease-associated and polymorphic variants, experimental data on TP53 mutations, and analysis of mutations in the COSMIC database. The method's ability to predict functional impact is demonstrated through its performance in separating recurrent and single mutations, as well as mutations in multiply mutated genes and annotated tumor suppressor and oncogenes. The authors conclude that the FIS provides a robust tool for prioritizing mutations for experimental investigation and therapeutic development.