Evolution of Large Polymorphic Inversions in a Panmictic Songbird

Chromosomal inversions have long been appreciated as an important source of genetic diversity, local adaptation, and speciation. However, selection pressures maintaining ancestral and derived alleles at high frequency over extended periods of time remain poorly characterized. Using genome-wide single-nucleotide polymorphism markers and shared barcodes of linked-read sequences from 20 wild and 7 captive zebra finches Taeniopygia guttata, we systematically scanned a high-quality zebra finch reference genome and identified all large polymorphic inversions that segregate at high minor allele frequencies. Apart from the known polymorphic inversions on chromosomes Tgu5, Tug11, Tgu13, and TguZ, we characterized two inversions on microchromosomes Tgu26 and Tgu27 and identified another eight putative inversions, located mostly on microchromosomes and ranging in size from 0.42 to 65.22 Mb. Population genomic analyses show that most of the six bona fide inversions are complex, containing short nested inversions. The early inversions emerged an estimated 0.6 to 2.2 million years ago and segregate at relatively high frequencies in the wild (minor haplotype frequency range: 0.289 to 0.429). Based on fitness-related measures of about 5,000 captive zebra finches, we conclude that three of the inversion polymorphisms (Tgu11, Tgu27, and TguZ) may be maintained by net heterosis. In the youngest of the six inversions (Tgu13), the derived haplotype showed weak positive additive effects on various fitness components. In combination with previous discoveries, we provide a comprehensive overview of the genomic distribution and evolutionary dynamics of large polymorphic inversions in the panmictic zebra finch. Our findings highlight (i) that microchromosomes may harbor quite a few additional inversion polymorphisms, (ii) that most of the inversions contain smaller nested or overlapping inversions, and (iii) that inversions were most likely maintained by weak heterosis with small fitness effects requiring large sample sizes to be detected.