摘要:SummaryDiscovering loci under balancing selection in humans can identify loci with alleles that affect response to the environment and disease. Genome variation data have identified the 5′ region of theDMBT1gene as undergoing balancing selection in humans.DMBT1encodes the pattern-recognition glycoprotein DMBT1, also known as SALSA, gp340, or salivary agglutinin. DMBT1 binds to a variety of pathogens through a tandemly arranged scavenger receptor cysteine-rich (SRCR) domain, with the number of domains polymorphic in humans. We show that the signal of balancing selection is driven by one haplotype usually carrying a shorter SRCR repeat and another usually carrying a longer SRCR repeat. DMBT1 encoded by a shorter SRCR repeat allele does not bind a cariogenic and invasiveStreptococcus mutansstrain, in contrast to the long SRCR allele that shows binding. Our results suggest that balancing selection atDMBT1is due to host-microbe interactions of encoded SRCR tandem repeat alleles.Graphical abstractDisplay OmittedHighlights•Clear evidence from many analyses show balancing selection at DMBT1•Scavenger-receptor cysteine-rich domain array associated with balancing selection•Genetic variation, not alternative splicing, responsible for protein isoforms•Long, but not short, DMBT1 isoforms bind a cariogenic strain ofStreptococcus mutansBiological sciences; Genetics; Evolutionary mechanisms