Structural variants (SVs) are large insertions or deletions of DNA sequences. While less numerous than single nucleotide polymorphisms, SVs often account for a greater proportion of nucleotide differences between genomes. Their size and frequent association with repetitive sequences has historically hindered their detection, which has limited the ability to associate this variation with molecular and phenotypic trait variation. While some SVs have been linked to observable traits, it remains unclear whether such effects are rare or broadly distributed across the genome. To test for genome-wide relationships between SVs and gene expression, we analyzed genome assemblies and transcriptomic data from 10 tissues across 26 diverse maize inbred lines. We identified SVs amongst these lines and examined variants located within the 1kb promoter region upstream of genes. Thousands of genes showed expression differences associated with promoter SVs, often in a tissue-specific manner. One common feature of these SVs was the presence of transposable element sequences. LTR retrotransposons were enriched amongst promoter SVs associated with differential expression and often reduced expression of the nearby gene. Despite widespread expression changes, we found no enrichment for specific biological functions or pathways among affected genes. Our findings indicate that extant TE-mediated promoter SVs play a significant role in shaping gene expression patterns across the maize genome. However, their phenotypic effects appear limited or context-dependent, suggesting that many variants may have minimal impact outside specific developmental stages or environmental conditions.

Several datasets used in our analysis are publicly available. Please see the associated github repository for a full detailing of publicly available data used and where to download these datasets. Data in this Dryad repository derives from these publicly available data sets and represents combined, processed, or simplified versions of this data. We strongly recommend reading the preprint for the associated manuscript, as the methods section adds substantial clarity on why and how this data was generated. If additional assistance is needed, please contact mmunasin@umn.edu or cnhirsch@umn.edu.

canonical_gene_models.tar.gz - MaizeGDB hosts the genomic feature file (gff3) for the V5 B73 Reference Genome. We filtered out any gene located on a scaffold, and, for the remaining genes, we retained only the canonical transcript. This zipped folder contains several summary files for the canonical transcripts of chromosomal B73 genes. The file canonical_gene_models_README.txt housed within this folder contains specific information on each file.
Gene_Lists.tar.gz - Previously constructed lists of maize genes associated with different functions were downloaded. The file Gene_Lists_README.txt lists out the various gene lists used in our analysis and contains specific information on each file in this folder.
Linked_SVs.tar.gz - Structural variants were identified from summarized pairwise whole-genome alignments between the B73 reference genome and every other line. To determine whether an SV in one pairwise comparison corresponded to an SV in another, we defined two SVs as the same SV if all of the following criteria were met: 1) the start positions were within 50 base pairs of each other, 2) the end positions were within 50 base pairs of each other, 3) the size of the SV differed by less than 100 base pairs, and 4) the category (assigned according to Munasinghe et al. 2023a) was the same. This folder contains the linked SV calls. The folder './LinkedSVs_ByChromosome' contains 10 different files, one for each chromosome. The file Linked_SVs_README.md contains information about the structure of each file.
StructuralHaplotype_Info.tar.gz - We focused on annotated chromosomal B73 genes and, for each gene, defined three distinct features based on the canonical transcript of that gene - the gene promoter, exonic sequence, and intronic sequence. We defined the promoter as either 1kb or 2kb upstream of the transcription start site. Using bedtools (v.2.29.2), we intersected the coordinates (in B73) of every SV with these gene features. We generated structural haplotypes for each gene in each NAM line from these intersections. The file StructuralHaplotype_Info_README.txt contains additional information on how these structural haplotypes were generated and on the structure of all files stored within.
candidate_expression.tar.gz - Raw RNA-sequencing data released with Huffored et al. 2021: https://www.ebi.ac.uk/ena/browser/view/ERX3793507. For each inbred line, these raw reads were mapped to the B73 v5 reference genome and normalized counts for each chromosomal gene was obtained. For our specific analysis, we focused on a subset of candidate genes that had structural variation within either the 1kb or 2kb promoter and tested for differential expression between structural haplotypes using a mixed effect model. Final significance was ascertained using a permutation-based approach. This folder contains the gene-level counts for all chromosomal genes as well as the subsetted expression values for candidate genes used in our actual differential expression analysis. The file candidate_expression_README.txt details the structure of this folder and all files contained within it.
random_mimic.tar.gz - To contextualize the results of our analysis, we generated a Promoter-Random gene set that represents a randon mimic of our true data set. This folder contains the expression/structural haplotype information for this data set. random_mimic_README.txt provides additional details on how this random mimic was generated, on the structure of this folder, and the structure of all files containined within this folder.
SNPlotypes.tar.gz - To contextualize the results of our analysis, we also generated a Promoter-SNP gene set derived from Promoter SNP haplotypes (hence SNPlotypes). For the promoter, in place of structural haplotypes, SNP haplotypes were constructed by extracting all overlapping SNPs from the pairwise whole-genome alignments. These SNPs were concatenated and used in place of the promoter structural haplotype, while the exonic and intronic structural haplotypes were retained. Thus, the only difference between the structural and SNP haplotypes for a given gene was the definition of the promoter haplotype - based on SV or SNP content, respectively. This folder contains the expression/structural haplotype information for this data set. SNPlotypes_README.txt provides additional details on how this data set was generated, on the structure of this folder, and the structure of all files containined within this folder.
ExpressionVariability.tar.gz - To test whether expression variance across tissues differed between structural haplotypes the coefficient of variation (CoV) and the mean absolute deviation (MAD) were tested on the subset of candidate genes with two structural haplotypes differing by only a single SV. For both variance metrics, the mean value was calculated for each structural haplotype belonging to a given gene and the absolute difference between the values was computed for the two structural haplotypes. This folder contains the output of this analysis. ExpressionVariability_README.txt provides additional details on the structure of all files within this folder.
Epigenetic_Overlaps.tar.gz - To explore whether the presence of SV sequence potentially altered the accessibility of the promoter, a subset of Promoter-SV genes was identified that had only two tested structural haplotypes that differ by only a single TE-mediated SV agnostic of whether that gene is differentially expressed. In this set, one structural haplotype was the ‘B73-like’ state, while the other had either an insertion or deletion relative to B73 depending on the type of SV located within the promoter. We extracted regions flanking the SV and intersected them with the location of previously identified unmethylated regions and accessible chromatin regions. This folder contains the downloaded data used in this analysis, and several intermediary output files. Epigenetic_Overlaps_README.txt ccontains further information on this analysis, on the structure of this folder, and the structure of all filess contained within this folder.
significant_genes.tar.gz - This folder contains several files that list out the significant genes identified from all of our statistical analyses of differential expression. significant_genes_README.txt contains extended information on all of the statistical analyses performed and the structure of all contained files.

Data from: Promoter structural variants are drivers of genome-wide differential expression in Maize

Data files

Abstract

Data from: Promoter structural variants are drivers of genome-wide differential expression in Maize

Data files

Abstract

README: Promoter Structural Variants are Drivers of Genome-Wide Differential Expression in Maize.