Genetic variation at transcription factor binding sites largely explains phenotypic heritability in maize

Abstract

Comprehensive maps of functional variation at transcription factor (TF) binding sites (cis-elements) are crucial for elucidating how genotype shapes phenotype. Here we report the construction of a pan-cistrome of the maize leaf under well-watered and drought conditions. We quantified haplotype-specific TF footprints across a pan-genome of 25 maize hybrids and mapped nearly two-hundred thousand genetic variants (termed binding-QTLs) linked tocis-element occupancy. The functional significance of binding-QTLs is supported by three lines of evidence: i) they coincide with known causative loci that regulate traits, including novel alleles ofUpright Plant Architecture2,Trehalase1, and the MITE transposon nearZmNAC111under drought; ii) their footprint bias is mirrored between inbred parents and by ChIP-seq; iii) partitioning genetic variation across genomic regions demonstrates that binding-QTLs capture the majority of heritable trait variation across ∼70% of 143 phenotypes. Our study provides a promising approach to make previously hiddencis-variation more accessible for genetic studies and multi-target engineering of complex traits.

Cross-species modeling of plant genomes at single nucleotide resolution using a pre-trained DNA language model

“…Unlike protein LMs that are limited to coding regions, DNA LMs enable a comprehensive understanding of the entire genome, offering deeper insights into gene regulation and evolution. Protein LMs have shown success in identifying pathogenic missense mutations in human genetics 8 , 11 , but increasing evidence shows that mutations in noncoding regions, including both intergenic and intronic regions, contribute significantly to both agronomic traits 12 and human diseases 13 , 14 . Additionally, training multi-species DNA LMs can capture evolutionary conservation at the DNA level, enhancing our understanding of genetic variation across different species.…”

“…Protein LMs have shown success in identifying pathogenic missense mutations in human genetics 8,11 , but increasing evidence shows that mutations in noncoding regions, including both intergenic and intronic regions, contribute significantly to both agronomic traits 12 and human diseases 13,14 .…”