Mutations of short tandem repeats explain abundant trait heritability in Arabidopsis

Short tandem repeats (STRs) mutations are one of the most important fuels of evolution, but are often largely ignored despite their huge effects on phenotypic variations. Here, we leveraged seven Arabidopsis mutation accumulation lines to assess the STR mutation rate, which ranged from 1.93×10 ⁻² to 4.40×10 ⁻³ per STR locus per generation and was much higher than the common SNP mutation rate at the level of 10 ⁻⁹ -10 ⁻⁸ . Interspecific comparison revealed rapid STR turnover, with a large proportion (71.8%) only present in Arabidopsis. In addition, intraspecific comparison of the ten assembled Arabidopsis genomes revealed that 29.3% of STRs were presence/absence variation (PAV), 36.5% of STRs were length variation (LV), 13.0% of STRs were no variation (NV) and the 21.1% STR loci with both LV and PAV. Based on 413 RNA-seq datasets, we identified that length variation of 3,871 STR loci (eSTRs) was associated with gene expression and 651 STR loci (sSTRs) were associated with splicing variation, and over one thousand eSTRs or sSTRs were linked with published genome-wide association study signals. In addition, a direct association between STR length variation and phenotypic variation uncovered a series of STR-trait associations. More notably, using the expression levels of 24,175 genes, the splice strength values of 12,784 splice sites, and 16 phenotypes of natural Arabidopsis populations, we revealed that the STR heritability is 0.111, 0.143 and 0.101, respectively. Overall, our results not only revealed the evolutionary dynamics of STRs, but highlighted the importance of STR variations as an important contributor to missing heritability in the regulation of complex traits.