Genome-wide screen of genomic imprinting in endosperm and population-level analysis reveal allelic variation for imprinting in flax

Genomic imprinting is an epigenetic phenomenon caused by the biased expression of maternally and paternally inherited alleles. In flowering plants, genomic imprinting predominantly occurs in triploid endosperm and plays a vital role in seed development. In this study, we identified 241 candidate imprinted genes including 143 maternally expressed imprinted genes (MEGs) and 98 paternally expressed imprinted genes (PEGs) in flax (Linum usitatissimum L.) endosperm using deep RNA sequencing. The conservation of imprinting in plants is very limited and imprinting clustering is not a general feature. MEGs tends to be endosperm expression specific, while PEGs are non-tissue specific. Imprinted SNPs differentiated 200 flax cultivars into oil flax, oil-fiber dual purpose flax (OF) and fiber flax subgroups, suggesting that genomic imprinting contributes to intraspecific variation in flax. The nucleotide diversity ({pi}) of imprinted genes in oil flax subgroup is significantly higher than that in fiber flax subgroup, indicating that some imprinted genes undergo positive selection during flax domestication from oil flax to fiber flax. Imprinted genes undergo positive selection is related to the functions. Eleven imprinted genes related to seed size and weight were identified using the candidate gene-based association study. Our study provides information for further exploring the function and genomic variation of imprinted genes in flax population.