The role of missense changes in BRCA1 in breast cancer susceptibility has been difficult to establish. We used comparative evolutionary methods to identify potential functionally important amino acid sites in exon 11 and missense changes likely to disrupt gene function, aligning sequences from 57 eutherian mammals and categorizing amino acid sites by degree of conservation. We used Bayesian phylogenetic analyses to determine relationships among orthologs and identify codons evolving under positive selection. Most conserved residues occur in a region with the highest concentration of protein-interacting domains. Rapidly evolving residues are concentrated in the RAD51-interacting domain, suggesting that selection is acting most strongly on the role of BRCA1 in DNA repair. Investigation of the functional role of missense changes in breast-cancer susceptibility should focus on 38 missense changes in conserved and 3 in rapidly evolving regions of exon 11. breast cancer ͉ exon 11 ͉ gene evolution ͉ missense change ͉ BRCA1 P rotein-truncating mutations distributed across BRCA1 are associated with an increased cumulative lifetime risk of breast (60-80%) and ovarian (20-40%) cancer (reviewed in ref. 1). Known mutations in breast-cancer susceptibility genes have been collated in the Breast Cancer Information Core (BIC) database (2, 3). Nearly half the reported changes in BRCA1 are frameshift mutations and thus expected to be disease associated (2). Most of the rest are missense changes; 323 of these have been reported in 1,735 individual entries. Disease-association status is known for only a fraction of these: in the RING finger domain (4) and the C-terminal region of the protein (5, 6). Case-control and family studies are underpowered to draw conclusions regarding the remainder; these are not highly penetrant alleles (7-9).The BRCA1 gene encodes a 1,863-aa protein with a single large region, exon 11, encoding some 60%. The gene is highly polymorphic, with many common single-base exon changes. Regions interacting with other proteins have been identified, but structural and biochemical properties of the protein remain largely unknown, making it difficult to predict the consequences of any single missense change (10). Available functional assays are time-consuming, expensive, and applicable only to Cterminal mutations (6,11,12). Predictions regarding missense changes can be strengthened by comparative evolutionary analysis to establish whether mutations cluster in conserved regions (13)(14)(15). Such analyses may be particularly helpful in identifying low-penetrance missense changes in functionally important regions. Phylogenetic approaches can also determine whether certain residues have evolved more rapidly than predicted by neutral theory (the ratio of the rate of nonsynonymous to synonymous substitution, , Ͼ1), reflecting the action of positive (diversifying) selection (16,17).The ability to detect conserved (18-20) and rapidly evolving (21, 22) regions in BRCA1 has been limited by the small number of cloned sequences ...