Umeå University's logo

Aspen (Populus tremula L.) is a keystone species and a model system for forest tree genomics. We present an updated resource comprising a chromosome-scale assem- bly, population genetics and genomics data. Using the resource, we explore the genetic basis of natural variation in leaf size and shape, traits with complex genetic architecture.

We generated the genome assembly using long-read sequencing, optical and high-density genetic maps. We conducted whole-genome resequencing of the Umeå Aspen (UmAsp) collection. Using the assembly and re-sequencing data from the UmAsp, Swedish Aspen (SwAsp) and Scottish Aspen (ScotAsp) collections we performed genome-wide association analyses (GWAS) using Single Nucleotide Polymorphisms (SNPs) for 26 leaf physiognomy phenotypes. We conducted Assay of Transposase Accessible Chromatin sequencing (ATAC-Seq), identified genomic regions of accessible chromatin, and subset SNPs to these regions, improving the GWAS detection rate. We identified candidate long non-coding RNAs in leaf samples, quantified their expression in an updated co-expression network, and used this to explore the functions of candidate genes identified from the GWAS.

A GWAS found SNP associations for seven traits. The associated SNPs were in or near genes annotated with developmental functions, which represent candidates for further study. Of particular interest was a !177-kbp region harbouring associations with several leaf phenotypes in ScotAsp.

We have incorporated the assembly, population genetics, genomics, and GWAS data into the PlantGenIE.org web resource, including updating existing genomics data to the new genome version, to enable easy exploration and visualisation. We provide all raw and processed data to facilitate reuse in future studies.

Peptidoglycan (PG), a mesh-like structure which is the primary component of the bacterial cell wall, is crucial to maintain cell integrity and shape. While most bacteria rely on penicillin binding proteins (PBPs) for crosslinking, some species also employ LD-transpeptidases (LDTs). Unlike PBPs, the essentiality and biological functions of LDTs remain largely unclear. The Hyphomicrobiales order of the Alphaproteobacteria, known for their polar growth, have PG which is unusually rich in LD-crosslinks, suggesting that LDTs may play a more significant role in PG synthesis in these bacteria. Here, we investigated LDTs in the plant pathogen Agrobacterium tumefaciens and found that LD-transpeptidation, resulting from at least one of 14 putative LDTs present in this bacterium, is essential for its survival. Notably, a mutant lacking a distinctive group of 7 LDTs which are broadly conserved among the Hyphomicrobiales exhibited reduced LD-crosslinking and tethering of PG to outer membrane β-barrel proteins. Consequently, this mutant suffered severe fitness loss and cell shape rounding, underscoring the critical role played by these Hyphomicrobiales-specific LDTs in maintaining cell wall integrity and promoting elongation. Tn-sequencing screens further revealed non-redundant functions for A. tumefaciens LDTs. Specifically, Hyphomicrobiales-specific LDTs exhibited synthetic genetic interactions with division and cell cycle proteins, and a single LDT from another group. Additionally, our findings demonstrate that strains lacking all LDTs except one displayed distinctive phenotypic profiles and genetic interactions. Collectively, our work emphasizes the critical role of LD-crosslinking in A. tumefaciens cell wall integrity and growth and provides insights into the functional specialization of these crosslinking activities.

Gene regulatory and gene co-expression networks are powerful research tools for identifying biological signal within high-dimensional gene expression data. In recent years, research has focused on addressing shortcomings of these techniques with regard to the low signal-to-noise ratio, non-linear interactions and dataset dependent biases of published methods. Furthermore, it has been shown that aggregating networks from multiple methods provides improved results. Despite this, few useable and scalable software tools have been implemented to perform such best-practice analyses. Here, we present Seidr (stylized Seiðr), a software toolkit designed to assist scientists in gene regulatory and gene co-expression network inference. Seidr creates community networks to reduce algorithmic bias and utilizes noise corrected network backboning to prune noisy edges in the networks.

Using benchmarks in real-world conditions across three eukaryotic model organisms, Saccharomyces cerevisiae, Drosophila melanogaster, and Arabidopsis thaliana, we show that individual algorithms are biased toward functional evidence for certain gene-gene interactions. We further demonstrate that the community network is less biased, providing robust performance across different standards and comparisons for the model organisms.

Finally, we apply Seidr to a network of drought stress in Norway spruce (Picea abies (L.) H. Krast) as an example application in a non-model species. We demonstrate the use of a network inferred using Seidr for identifying key components, communities and suggesting gene function for non-annotated genes.

Parrots (Psittaciformes) are among the most endangered groups of birds today and remain threatened by habitat loss and exploitation for the live bird trade. Under such conditions, reliable and non-invasive monitoring techniques are crucial for successful conservation measures. In this study, we developed a panel of 86 high quality SNPs for genotyping endangered sun parakeets (Aratinga solstitialis) in Guyana, which form one of the last known breeding populations of this South American species in the wild. Genotyping was tested on different types of samples (blood, feathers, feces, beak and cloacal swabs). While blood performed best, feathers and feces also yielded reliable results and could thus be used as non-invasive sources of DNA for future population monitoring. Discriminant Analysis of Principal Components (DAPC) on genotypes revealed that Guyanese sun parakeets clustered separately from other psittacine species as well as conspecifics from a captive population. A priori known first-order kinships were also adequately detected by the SNP panel. Using a series of experimental contaminations, we found that contamination from other psittacine species and slight contamination ( 10%) from conspecifics did not prevent successful genotyping and recognition of individuals. We show that instances of higher conspecific contamination ( 50%) can be detected through an increased level of heterozygosity that falls outside the distribution of uncontaminated samples.

Complex traits represent valuable research targets as many highly desirable properties of plants and animals (such as growth rate and height) fall into this group. However, associating biological markers with these traits is incredibly challenging, in part due to their small effect sizes. For the two species at the core of our research, European aspen (Populus tremula) and Norway spruce (Picea abies), association studies are even more challenging, primarily due to the fragmented state of their genome assemblies. These assemblies represent the gene space well, but poorly represented inter-genic regions hinder variant discovery and large scale association studies.

In this thesis, I present my work to improve association discovery of complex traits in forest trees. Firstly, to overcome the issues with assembly fragmentation, I have created an updated version of the P. tremula genome, which is highly contiguous and anchored in full chromosomes. To calculate the dense linkage map required to order and orient the aspen assembly, I developed "BatchMap", a parallel implementation of linkage mapping software. BatchMap has been successfully applied to several dense linkage maps, including aspen and Norway spruce, and was essential to the progress in improving the aspen genome assembly. Further, I developed seidr, which represents a starting point in multi-layer, network-based systems biology, an analysis technique with promising prospects for complex trait association analysis. As a case study, I applied some of the methods developed to the analysis of leaf shape in natural populations of European aspen, a complex, omnigenic trait.

The multi-layer model of systems biology and related analysis techniques offer promise in the analysis of complex traits, and this thesis represents a starting point toward an intricate, holistic model of systems biology that may help to unravel the overwhelmingly complicated nature of complex traits.

The rate of meiotic recombination is one of the central factors determining genome-wide levels of linkage disequilibrium which has important consequences for the efficiency of natural selection and for the dissection of quantitative traits. Here we present a new, high-resolution linkage map for Populus tremula that we use to anchor approximately two thirds of the P. tremula draft genome assembly on to the expected 19 chromosomes, providing us with the first chromosome-scale assembly for P. tremula (Table 2). We then use this resource to estimate variation in recombination rates across the P. tremula genome and compare these results to recombination rates based on linkage disequilibrium in a large number of unrelated individuals. We also assess how variation in recombination rates is associated with a number of genomic features, such as gene density, repeat density and methylation levels. We find that recombination rates obtained from the two methods largely agree, although the LD-based method identifies a number of genomic regions with very high recombination rates that the map-based method fails to detect. Linkage map and LD-based estimates of recombination rates are positively correlated and show similar correlations with other genomic features, showing that both methods can accurately infer recombination rate variation across the genome. Recombination rates are positively correlated with gene density and negatively correlated with repeat density and methylation levels, suggesting that recombination is largely directed toward gene regions in P. tremula.

Leaf shape is a defining feature of how we recognize and classify plant species. Although there is extensive variation in leaf shape within many species, few studies have disentangled the underlying genetic architecture. We characterized the genetic architecture of leaf shape variation in Eurasian aspen (Populus tremula L.) by performing genome‐wide association study (GWAS) for physiognomy traits. To ascertain the roles of identified GWAS candidate genes within the leaf development transcriptional program, we generated RNA‐Seq data that we used to perform gene co‐expression network analyses from a developmental series, which is publicly available within the PlantGenIE resource. We additionally used existing gene expression measurements across the population to analyze GWAS candidate genes in the context of a population‐wide co‐expression network and to identify genes that were differentially expressed between groups of individuals with contrasting leaf shapes. These data were integrated with expression GWAS (eQTL) results to define a set of candidate genes associated with leaf shape variation. Our results identified no clear adaptive link to leaf shape variation and indicate that leaf shape traits are genetically complex, likely determined by numerous small‐effect variations in gene expression. Genes associated with shape variation were peripheral within the population‐wide co‐expression network, were not highly connected within the leaf development co‐expression network, and exhibited signatures of relaxed selection. As such, our results are consistent with the omnigenic model.

Plant mitogenomes can be difficult to assemble because they are structurally dynamic and prone to intergenomic DNA transfers, leading to the unusual situation where an organelle genome is far outnumbered by its nuclear counterparts. As a result, comparative mitogenome studies are in their infancy and some key aspects of genome evolution are still known mainly from pregenomic, qualitative methods. To help address these limitations, we combined machine learning and in silico enrichment of mitochondrial-like long reads to assemble the bacterial-sized mitogenome of Norway spruce (Pinaceae: Picea abies). We conducted comparative analyses of repeat abundance, intergenomic transfers, substitution and rearrangement rates, and estimated repeat-by-repeat homologous recombination rates. Prompted by our discovery of highly recombinogenic small repeats in P. abies, we assessed the genomic support for the prevailing hypothesis that intramolecular recombination is predominantly driven by repeat length, with larger repeats facilitating DNA exchange more readily. Overall, we found mixed support for this view: Recombination dynamics were heterogeneous across vascular plants and highly active small repeats (ca. 200 bp) were present in about one-third of studied mitogenomes. As in previous studies, we did not observe any robust relationships among commonly studied genome attributes, but we identify variation in recombination rates as a underinvestigated source of plant mitogenome diversity.