Microhomology-mediated tandem duplication is a conserved mechanism of genomic variation with implications for human...

Tandem repeats are highly mutable genomic elements with significant functional consequences, yet the mechanisms underlying their evolutionary origin remain unclear. One proposed mechanism is microhomology-mediated tandem duplication (MTD), in which single-copy DNA segments flanked by microhomology undergo duplication and may subsequently expand. Although MTD was first described in Schizosaccharomyces pombe, its prevalence and evolutionary significance across life have not been systematically established. Using whole-genome deep sequencing and a unified analytical framework, we show that MTDs arise de novo across bacteria, archaea, fungi, and viruses. Analyses of 2,245 reference genomes, millions of genomes from 103 microbial species, and human datasets reveal that microhomology-mediated duplications constitute a major source of tandem duplication across domains of life. Genome-wide analyses show that most MTDs evolve under neutral or nearly neutral dynamics, while purifying selection preferentially depletes MTDs from coding regions. Mechanistically, deletion of the conserved flap endonuclease Rad27 specifically increases de novo MTD formation in budding yeast, implicating Okazaki fragment maturation in the generation of MTDs. In humans, microhomology signatures are pervasive among reference, polymorphic, and disease-associated tandem duplications and are enriched among pathogenic variants linked to genome stability and cancer. Together, these findings establish MTD as a conserved mechanism that shapes genomic variation, with implications for human disease.