Attention:The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Thursday, August 13 until 12:00 AM ET on Friday, August 14 due to maintenance. We apologize for the inconvenience.


This content will become publicly available on December 1, 2026

Title: Large DNA and protein language models enhance discovery of deleterious mutations in maize
Background The maize inbred line Chang7-2 and derived lines are important male donors for hybrid production, contributing significantly to the development of high-yield and stress-tolerant hybrids. Additionally, Chang7-2 serves as a valuable model inbred line for genetic and genomic studies, facilitating the discovery of genes underlying hybrid vigor and other agronomic traits. Results Here, a reference genome assembly and a chemical-induced mutant population (N = 1,716) through ethyl methyl sulfonate (EMS) treatments are generated using Chang7-2. Each EMS line is whole genome sequenced and compared to the Chang7-2 genome, identifying 2,586,769 mutations with 4,939 mutations causing premature stop codons or altered splicing sites. The effect estimation of mutations using two large language artificial intelligence (AI) models, namely the protein language model ESM1b and the DNA language model PlantCaduceus, reveals 15,264 and 18,326 deleterious mutations, respectively. Mutation effects estimated with AI models accelerate revelation of four causal mutations underlying phenotypes of albino leaf, reduced cuticular wax, altered seed color, and male sterility. In addition, allelic expression quantification of genic mutations in 13 EMS M1 lines and their M2 heterozygous progeny, which contain both wildtype and mutant alleles, shows that mutant alleles are overall accumulated at a lower level compared to wildtype. Such allelic disparity is observed for some synonymous mutations, indicating they may not be biologically inconsequential. Conclusions AI-based estimation of mutation effects offers cross-species evidence for functional impacts of mutations. Our study demonstrates its application in revealing deleterious EMS mutations and identifying causal mutations responsible for mutant phenotypes.  more » « less
Award ID(s):
2311738 2011500
PAR ID:
10691858
Author(s) / Creator(s):
; ; ; ; ; ; ; ; ; ; ; ; ; ; ;
Publisher / Repository:
Springer Nature
Date Published:
Journal Name:
Genome Biology
Volume:
26
Issue:
1
ISSN:
1474-760X
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Ethyl methanesulfonate (EMS) mutagenesis offers important advantages for improving crops, such as cotton, with limited diversity in elite gene pools. EMS-induced point mutations are less frequently associated with deleterious traits than alleles from wild or exotic germplasm. From 157 mutant lines that have significantly improved fiber properties, we focused on nine mutant lines here. A total of eight populations were developed by crossing mutant lines in different combinations into GA230 (GA2004230) background. Multiple lines in each population were significantly improved for the fiber trait that distinguished the donor parent(s), demonstrating that an elite breeding line (GA230) could be improved for fiber qualities using the mutant lines. Genotypes improved for multiple fiber traits of interest suggesting that allele pyramiding is possible. Compared to midparent values, individual progeny in the population conferred fiber quality improvements of as much as 31.7% (in population O) for micronaire (MIC), 16.1% (in population P) for length, 22.4% (in population K) for strength, 4.1% (in population Q) for uniformity, 45.8% (in population N) for elongation, and 13.9% (in population O) for lint percentage (lint%). While further testing for stability of the phenotype and estimation of yield potential is necessary, mutation breeding shows promise as an approach to reduce the problem of the genetic bottleneck of upland cotton. The populations developed here may also contribute to identifying candidate genes and causal mutations for fiber quality improvement. 
    more » « less
  2. We provide a partial test of the mitonuclear sex hypothesis with the first controlled study of how male frequencies and rates of outcrossing evolve in response to mitonuclear mismatch by allowing replicate lineages of C. elegans nematodes containing either mitochondrial or nuclear mutations of electron transport chain (ETC) genes to evolve under three sexual systems: facultatively outcrossing (wildtype), obligately selfing, and obligately outcrossing. Among facultatively outcrossing lines, we found evolution of increased male frequency in at least one replicate line of all four ETC mutant backgrounds tested—nuclear isp-1 , mitochondrial cox-1 and ctb-1 , and an isp-1 IV; ctb-1M mitonuclear double mutant—and confirmed for a single line set ( cox-1 ) that increased male frequency also resulted in successful outcrossing. We previously found the same result for lines evolved from another nuclear ETC mutant, gas-1 . For several lines in the current experiment, however, male frequency declined to wildtype levels (near 0%) in later generations. Male frequency did not change in lines evolved from a wildtype control strain. Additional phenotypic assays of lines evolved from the mitochondrial cox-1 mutant indicated that evolution of high male frequency was accompanied by evolution of increased male sperm size and mating success with tester females, but that it did not translate into increased mating success with coevolved hermaphrodites. Rather, hermaphrodites’ self-crossed reproductive fitness increased, consistent with sexually antagonistic coevolution. In accordance with evolutionary theory, males and sexual outcrossing may be most beneficial to populations evolving from a state of low ancestral fitness ( gas-1 , as previously reported) and less beneficial or deleterious to those evolving from a state of higher ancestral fitness ( cox-1 ). In support of this idea, the obligately outcrossing fog-2 V; cox-1 M lines exhibited no fitness evolution compared to their ancestor, while facultatively outcrossing lines showed slight upward evolution of fitness, and all but one of the obligately selfing xol-1 X; cox-1 M lines evolved substantially increased fitness—even beyond wildtype levels. This work provides a foundation to directly test the effect of reproductive mode on the evolutionary dynamics of mitonuclear genomes, as well as whether compensatory mutations (nuclear or mitochondrial) can rescue populations from mitochondrial dysfunction. 
    more » « less
  3. Abstract Sweet orange originated from the introgressive hybridizations of pummelo and mandarin resulting in a highly heterozygous genome. How alleles from the two species cooperate in shaping sweet orange phenotypes under distinct circumstances is unknown. Here, we assembled a chromosome-level phased diploid Valencia sweet orange (DVS) genome with over 99.999% base accuracy and 99.2% gene annotation BUSCO completeness. DVS enables allele-level studies for sweet orange and other hybrids between pummelo and mandarin. We first configured an allele-aware transcriptomic profiling pipeline and applied it to 740 sweet orange transcriptomes. On average, 32.5% of genes have a significantly biased allelic expression in the transcriptomes. Different cultivars, transgenic lineages, tissues, development stages, and disease status all impacted allelic expressions and resulted in diversified allelic expression patterns in sweet orange, but particularly citrus Huanglongbing (HLB) shifted the allelic expression of hundreds of genes in leaves and calyx abscission zones. In addition, we detected allelic structural mutations in an HLB-tolerant mutant (T19) and a more sensitive mutant (T78) through long-read sequencing. The irradiation-induced structural mutations mostly involved double-strand breaks, while most spontaneous structural mutations were transposon insertions. In the mutants, most genes with significant allelic expression ratio alterations (≥1.5-fold) were directly affected by those structural mutations. In T19, alleles located at a translocated segment terminal were upregulated, including CsDnaJ, CsHSP17.4B, and CsCEBPZ. Their upregulation is inferred to keep phloem protein homeostasis under the stress from HLB and enable subsequent stress responses observed in T19. DVS will advance allelic level studies in citrus. 
    more » « less
  4. Seed mutagenesis using alkylating chemical agents such as ethyl methanesulfonate (EMS) can generate somatic and germinal mutations in many plant species. In monoecious plants like maize, the sperm- and egg-producing reproductive germlines are derived from distinct cell lineages in the embryo. This separation results in independent mutations inherited via the egg and sperm lineages and prevents the recovery of recessive mutant phenotypes in diploid progeny after the first round of self-pollination. Thus, two generations of self-pollination are required to screen for recessive mutations when conducting seed mutagenesis. The additional time and manual self-pollination make this approach laborious. However, a high mutation rate and the ability to screen for somatic sectors in heterozygous mutant plants and other defined genetic backgrounds make seed mutagenesis an effective but underutilized mutagenesis tool for maize research. This protocol provides the directions and optimization steps to perform effective seed mutagenesis in maize. A high frequency of somatic mutations from seed mutagenesis can be achieved, but comes at the expense of poor and disordered growth, failure to form reproductive structures, and low or no seed production at high EMS concentrations or long contact times. In experiments where germinal mutations are a goal, an optimum dose of EMS is required in the first generation. Maize genetic backgrounds vary in their sensitivity to EMS, requiring some pilot testing in new genetic backgrounds. Researchers using this protocol can carry out seed mutagenesis safely and effectively to develop libraries of mutants or alleles for various experiments. 
    more » « less
  5. Creating mutations in maize has provided key foundational information for our mechanistic understanding of genetics, evolution, and even the role of chromosomes as units of inheritance. Chemical mutagenesis is used in biological research to create novel genetic variation. Ethyl methanesulfonate (EMS) is an alkylating agent and a highly potent and frequently used mutagen. EMS mutagenesis can be used to identify genes based on phenotypes induced by mutagenesis (forward genetics) and to validate the functions of genes by independently creating multiple mutant alleles in known genes (reverse genetics). Due to our ability to collect huge quantities of maize pollen and to easily apply pollen to the silks of maize ears to conduct pollination and achieve hundreds of fertilization events, pollen EMS mutagenesis is uniquely facile in maize. While pollen EMS mutagenesis is commonly performed, treatment of maize seeds with EMS is also highly effective, and can be used for certain research objectives that are difficult to achieve with pollen mutagenesis, such as recovering mutant sectors. The alkylation of guanine residues by EMS primarily results in G > A or C > T transitions in the DNA, making the molecular profiling of mutations caused by EMS easy, with an extremely low false positive rate. EMS is hydrophilic, has a moderate half-life in water, and is sensitive to light and high temperatures. With appropriate precautions in research settings, EMS can be relatively safe to handle. Here, we provide an introduction to chemical mutagenesis via EMS, including some history on its use in maize and the considerations for the effective and safe design of mutagenesis experiments with EMS in maize. 
    more » « less