Abstract

Classification of rare sequence variants observed during mutation screening of susceptibility genes in high-risk individuals presents an interesting and medically important challenge. A recently described method for analysis of unclassified variants in BRCA1 and BRCA2 provides an extensible framework within which several different types of analytic data can be integrated. Among the methods already integrated in this framework are a measure of sequence conservation at specific positions in BRCA1 and BRCA2, and a measure of the difference between wild-type and missense amino acid residues, the Grantham Matrix Score. Recently, we extended the idea of Grantham Matrix Scores to multiple sequence alignments by introducing two new measures, the Grantham variation and Grantham Deviation. We also created a measure of risk associated with sets of BRCA1 missense substitutions, the BRCA1-with-BRCA2 Ascertainment Ratio. Here, we complement these measures with a more powerful measure of risk associated with sets of missense substitutions, the Missense Enrichment Ratio. By combining these four measures, we demonstrate two points: (1) pooled evidence is completely in accord with a hypothesis that missense substitutions that fall at variable positions in the alignment of vertebrate BRCA1s and are within the range of variation observed at those positions are neutral, and (2) many of the missense substitutions falling at invariant positions in the alignment must be deleterious and the longer the period over which the position has been invariant, the stronger the evidence that this is so.