[Paper Review] The pattern and distribution of deleterious mutations in maize

Sofiane Mezmouk, Jeffrey Ross‐Ibarra|arXiv (Cornell University)|Aug 2, 2013

Evolution and Genetic Dynamics42 references4 citations

TL;DR

This study conducts the first genome-wide scan of deleterious coding variants in maize using high-density genotyping data from inbred lines. It finds that most predicted deleterious SNPs are at low frequencies, and genes harboring them are significantly enriched for associations with complex traits—providing strong support for the dominance model of heterosis, where hybrid vigor arises from complementation of recessive deleterious alleles across inbred lines.

ABSTRACT

Most non-synonymous mutations are thought to be deleterious because of their effect on protein sequence. These polymorphisms are expected to be removed or kept at low frequency by the action of natural selection, and rare deleterious variants have been implicated as a possible explanation for the "missing heritability" seen in many studies of complex traits. Nonetheless, the effect of positive selection on linked sites or drift in small or inbred populations may also impact the evolution of deleterious alleles. Here, we made use of genome-wide genotyping data to characterize deleterious variants in a large panel of maize inbred lines. We show that, in spite of small effective population sizes and inbreeding, most putatively deleterious SNPs are indeed at low frequencies within individual genetic groups. We find that genes showing associations with a number of complex traits are enriched for deleterious variants. Together these data are consistent with the dominance model of heterosis, in which complementation of numerous low frequency, weak deleterious variants contribute to hybrid vigor.

Motivation & Objective

• To characterize the frequency and distribution of deleterious mutations across the maize genome using high-density genotyping data.
• To assess whether genes with deleterious variants are enriched for associations with complex agronomic traits in genome-wide association studies.
• To evaluate the contribution of rare, weakly deleterious alleles to heterosis in maize, particularly under the dominance model.
• To investigate how population structure, inbreeding, and selection shape the evolutionary dynamics of deleterious variants in maize.
• To explore the potential of in silico prediction of deleterious variants for improving breeding value and combining ability predictions in maize.

Proposed method

• Used whole-genome genotyping data from a large panel of maize inbred lines to identify single nucleotide polymorphisms (SNPs) across the genome.
• Applied in silico prediction tools to classify non-synonymous SNPs as deleterious based on evolutionary conservation and protein structure impact.
• Analyzed SNP frequency distributions within and across genetic groups to assess the effects of inbreeding and effective population size.
• Performed genome-wide association studies (GWAS) on 14 complex traits to test for enrichment of deleterious SNPs in associated loci.
• Conducted gene-level enrichment tests comparing genes with predicted deleterious SNPs to random gene sets, using statistical tests to assess significance.
• Used sign tests and correlation analyses to evaluate whether enrichment was due to chance, SNP density, or synthetic associations with rare deleterious variants.

Experimental results

Research questions

• RQ1 What is the frequency distribution of predicted deleterious non-synonymous SNPs in maize inbred lines, and how is it affected by inbreeding and small effective population size?
• RQ2 Are genes associated with complex agronomic traits enriched for deleterious variants compared to background levels?
• RQ3 To what extent do rare, low-frequency deleterious SNPs contribute to phenotypic variation in maize, particularly in the context of heterosis?
• RQ4 Is the observed enrichment of deleterious variants in trait-associated genes due to confounding factors such as SNP density or linkage disequilibrium?
• RQ5 Can in silico prediction of deleterious variants improve the accuracy of breeding value and combining ability predictions in maize?

Key findings

• Most putatively deleterious non-synonymous SNPs in maize inbred lines are at low frequencies, consistent with purifying selection despite small effective population sizes and inbreeding.
• Genes associated with complex traits such as plant height, yield, and days to tasseling show significant enrichment (5–45%) for deleterious SNPs, with statistical significance for whole plant yield and days to tasseling.
• At the gene level, there is strong and statistically significant enrichment (p-value = 3 × 10⁻⁵ in population A, p = 0.01 in population B) of genes with predicted deleterious SNPs among trait-associated genes.
• Enrichment is not observed for low-frequency synonymous SNPs (p ≈ 1), indicating that the signal is specific to deleterious coding variants and not an artifact of SNP density or linkage.
• The correlation between total number of SNPs per gene and number of significant associations is low (r ≤ 0.2), suggesting enrichment is not driven by gene size or SNP count.
• The results support the dominance model of heterosis, where complementation of recessive deleterious alleles across inbred lines contributes to hybrid vigor through the masking of deleterious effects in heterozygotes.