bioRxiv Science⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2021.08.23.457367

Comparison of Single-Trait and Multi-Trait Genome-Wide Association Models and Inclusion of Correlated Traits in the Dissection of the Genetic Architecture of a Complex Trait in a Breeding Program

Merrick, L. F.·Burke, A. B.·Zhang, Z.·Carter, A. H.
View original source ↗Export metadata ↓

Abstract

Traits with an unknown genetic architecture make it difficult to create a useful bi-parental mapping population to characterize the genetic basis of the trait due to a combination of complex and pleiotropic effects. Seedling emergence of wheat (Triticum aestivum L.) from deep planting is a vital factor affecting stand establishment and grain yield, has a poorly understood genetic architecture, and is historically correlated with coleoptile length. The creation of bi-parental mapping populations can be overcome by using genome-wide association studies (GWAS). This study aimed to dissect the genetic architecture of seedling emergence while accounting for correlated traits using one multi-trait GWAS model (MT-GWAS) and three single-trait GWAS models (ST-GWAS) with the inclusion of covariates for correlated traits. The ST-GWAS models included one single locus model (MLM), and two multiple loci models (FarmCPU and BLINK). We conducted the GWAS using two populations, the first consisting of 473 varieties from a diverse association mapping panel (DP) phenotyped from 2015-2019, and the other population used as a validation population consisting of 279 breeding lines (BL) phenotyped in 2015 in Lind, WA, with 40,368 markers. We also compared the inclusion of coleoptile length and markers associated with reduced height as covariates in our ST-GWAS models for the DP. ST-GWAS found 107 significant markers across 19 chromosomes, while MT-GWAS found 82 significant markers across 14 chromosomes. MT-GWAS models were able to identify large-effect markers on chromosome 5A. FarmCPU and BLINK models were able to identify many small effect markers, and the inclusion of covariates helped to identify the large effect markers on chromosome 5A. Therefore, by using multi-locus models combined with pleiotropic covariates, breeding programs can uncover the complex nature of traits to help identify candidate genes and the underlying architecture of a trait, such as seedling emergence of deep-sown winter wheat.

Source connections

Explore related subjects

geneticsPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Merrick, L. F., Burke, A. B., Zhang, Z., Carter, A. H.. 2021-08-24. Comparison of Single-Trait and Multi-Trait Genome-Wide Association Models and Inclusion of Correlated Traits in the Dissection of the Genetic Architecture of a Complex Trait in a Breeding Program. https://doi.org/10.1101/2021.08.23.457367

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Pangenome Discoveries from Clinical Long-Read Genomes in Underrepresented Populations with Neurodevelopmental Disorders

Genetic diagnoses remain elusive for many individuals with neurodevelopmental disorders (NDDs), and the gap is wider in populations poorly represented in genomic reference resources. To demonstrate the equitable advantage of long-read whole genome sequencing (LR-WGS) and pangenomes in terms of clinical utility, we analyzed 211 NDD cases and 107 unaffected relatives from 24 populations, integrating their assemblies, resulting in a total of 1204 haplotypes with the Human Pangenome Reference Consortium Release 2 (HPRC_r2) and the Arab Pangenome Reference (APR) datasets. Clinical interpretation of small variants and structural variants (SVs) revealed pathogenic or likely pathogenic findings in 38.9% of the cases (82/211), and our workflow further detected pathogenic repeat expansions in 3.8% of the cases and runs of homozygosity (ROHs) exceeding 300 Mb in 4.7% of the cases, revealing substantial autozygosity suggestive of strong inbreeding. Beyond these clinical assessments, pangenome analysis revealed 3.9 million population-level novel small variants, 39,782 SVs and an average of 2.17 Mb of novel sequence per individual absent from linear references and global pangenomes. We identified 85.9 kb of case-unique nonreference sequences in exonic segments overlapping with curated NDD gene sets. Graph-based tertiary analysis further identified rare exonic complex homozygous rearrangements in the NDD associated genes for 2.4% of the cases. Together, these findings emphasize how diverse genome assemblies extend variant discovery beyond established linear reference-based clinical interpretation.

genetics↗
bioRxiv2026

Integrated exome-transcriptome analysis of a schizophrenia case with an exceptionally low polygenic risk score

Schizophrenia liability reflects both common polygenic variation and rarer variants of larger effect. We investigated a post-mortem schizophrenia case (SZ07) with a schizophrenia polygenic risk score below the first percentile among ancestry-matched cases, asking whether individual-level genomic and transcriptomic analyses could help identify potentially novel schizophrenia-relevant variants. We used exome analysis with multiple variant-prediction tools to generate a panel of variants and corresponding candidate genes potentially implicated in the pathological process in SZ07. The prioritised variants were then assessed using RNA sequencing of three cortical regions. Of these, only an IZUMO4 splice-acceptor variant showed direct RNA-level support. Beyond variant-specific analyses, we characterised broader transcriptomic abnormalities in the SZ07 brain. Pathway analysis of gene-level expression deviations implicated innate immune processes, calcium signalling, synaptic biology, and axon remodelling. Notably, compared with conventional single-sample biological pathway analysis, the analysis we used showed only a weak association with neuronal specificity, suggesting their reduced sensitivity to the main axis of variation related to cellular composition in bulk brain tissue gene expression data. Although the candidate genes showed no significant pathway-level convergence, the broader transcriptomic profile revealed coherent biological alterations largely compatible with the documented history of cerebrovascular disease in SZ07. These results demonstrate how the integration of exome and transcriptome data at the individual level enables the formulation and constraint of hypotheses regarding patient-specific genetic features and their molecular consequences, while highlighting important limitations related to tissue composition, comorbidity, and the choice of reference cohort.

genetics↗
bioRxiv2026

A genetic screen to identify host factors required for commensal bacterial adherence to the gut epithelia

Bacterial adherence to gut epithelial cells likely involves host genetic components. We isolated a gut commensal bacterium, Lelliottia jeotgali, that can robustly colonize C. elegans through attachment to intestinal epithelial cells. Here, we developed a novel screen in C. elegans to identify host genes required for microbiome bacterial colonization. Using transgenic L. jeotgali strain expressing tdTomato, we conducted forward genetic screens in C. elegans to select for mutants that completely lost colonization. We identified multiple allelic mutations in five different C. elegans genes, termed knockout of Lelliottia adherence (kola) mutants, with a near complete loss of adherence. Each of kola-1-5 is uncharacterized and predicted to encode for a secreted or membrane-bound protein. We also identified gly-5, a GalNAc transferase that initiates O-linked glycosylation. Overall, we describe the first C. elegans screen to identify host factors that control microbiome bacterial colonization in the gut.

genetics↗