Editor's evaluation: GENESPACE tracks regions of interest and gene copy number variation across multiple genomes
Year: 2022
Type: peer-review
Abstract: Article Figures and data Abstract Editor's evaluation eLife digest Introduction Results and discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The development of multiple chromosome-scale reference genome sequences in many taxonomic groups has yielded a high-resolution view of the patterns and processes of molecular evolution. Nonetheless, leveraging information across multiple genomes remains a significant challenge in nearly all eukaryotic systems. These challenges range from studying the evolution of chromosome structure, to finding candidate genes for quantitative trait loci, to testing hypotheses about speciation and adaptation. Here, we present GENESPACE, which addresses these challenges by integrating conserved gene order and orthology to define the expected physical position of all genes across multiple genomes. We demonstrate this utility by dissecting presence–absence, copy-number, and structural variation at three levels of biological organization: spanning 300 million years of vertebrate sex chromosome evolution, across the diversity of the Poaceae (grass) plant family, and among 26 maize cultivars. The methods to build and visualize syntenic orthology in the GENESPACE R package offer a significant addition to existing gene family and synteny programs, especially in polyploid, outbred, and other complex genomes. Editor's evaluation GENESPACE is a new and straightforward computational tool to include synteny information in the calculation of genome-wide sets of orthologs. The development of this tool is very timely as more and more complete chromosome-scale assembled genomes are becoming available. While the assembly problem has been solved, this is not the case for multiple genome comparisons, and GENESPACE is an important step to help remedy this gap in our comparative genomics toolbox. https://doi.org/10.7554/eLife.78526.sa0 Decision letter Reviews on Sciety eLife's review process eLife digest The genome is the complete DNA sequence of an individual. It is a crucial foundation for many studies in medicine, agriculture, and conservation biology. Advances in genetics have made it possible to rapidly sequence, or read out, the genome of many organisms. For closely related species, scientists can then do detailed comparisons, revealing similar genes with a shared past or a common role, but comparing more distantly related organisms remains difficult. One major challenge is that genes are often lost or duplicated over evolutionary time. One way to be more confident is to look at ‘synteny’, or how genes are organized or ordered within the genome. In some groups of species, synteny persists across millions of years of evolution. Combining sequence similarity with gene order could make comparisons between distantly related species more robust. To do this, Lovell et al. developed GENESPACE, a software that links similarities between DNA sequences to the order of genes in a genome. This allows researchers to visualize and explore related DNA sequences and determine whether genes have been lost or duplicated. To demonstrate the value of GENESPACE, Lovell et al. explored evolution in vertebrates and flowering plants. The software was able to highlight the shared sequences between unique sex chromosomes in birds and mammals, and it was able to track the positions of genes important in the evolution of grass crops including maize, wheat, and rice. Exploring the genetic code in this way could lead to a better understanding of the evolution of important sections of the genome. It might also allow scientists to find target genes for applications like crop improvement. Lovell et al. have designed the GENESPACE software to be easy for other scientists to use, allowing them to make graphics and perform analyses with few programming skills. Introduction De novo genome assemblies and gene model annotations represent increasingly common resources that describe the sequence and positions of protein coding and intergenic regions within a single genotype. Evolutionary relationships among these DNA sequences form the foundation of many molecular tools in modern medical, breeding, and evolutionary biology research. Perhaps the most crucial inference to make when comparing genomes revolves around homologous genes, which share an evolutionary common ancestor and ensuing sequence or protein structure similarity. Analyses of homologs, including comparative gene expression, epigenetics, and sequence evolution, require the distinction between orthologs, which arise from speciation events, and paralogs, which arise from sequence duplications (see Box 1 for definitions). In some systems, this is a simple task where most genes are single copy, and orthologs are synonymous with reciprocal best-scoring protein BLAST hits. Other sequence similarity approaches such as OrthoFinder (Emms and Kelly, 2019; Emms and Kelly, 2015) leverage graphs and gene trees to test for orthology, permitting more robust analyses in systems with gene copy number (CNV) or presence–absence variation (PAV). However, whole-genome duplications (WGDs), chromosomal deletions, and variable rates of sequence evolution, such as subgenome dominance in polyploids, can confound the evidence of orthology from sequence similarity alone. Box 1 Definitions Orthogroup — a set of genes across multiple genomes derived from a single ancestral gene. Ortholog — a pair of orthogroup members in two species derived from a single gene in their most recent common ancestor. Paralog — orthogroup members derived from a duplication event since speciation. Homeolog — paralogs derived from a whole-genome duplication. Tandem array — paralogs in proximity on a chromosome within a genome. Gene collinearity — retained order of genes across species due to common ancestry. Synteny — like collinearity but at larger scales, like chromosome arms. Pan-genome annotation — A set of orthogroups across multiple genomes, placed along the coordinate system of a specified reference genome. The physical position of homologs offers a second line of evidence that can help overcome challenges posed by WGDs, tandem arrays, heterozygous-duplicated regions, and other genomic complexities (Drillon et al., 2020; Haug-Baltzell et al., 2017; Wang et al., 2012). Synteny, or the conserved order of DNA sequences among chromosomes that share a common ancestor, is a typical feature of eukaryotic genomes. In some taxa, synteny is preserved across hundreds of millions of years of evolution and is retained over multiple WGDs (Jiao et al., 2014; Simakov et al., 2020; Zhao and Schranz, 2019). Like chromosome-scale synteny, conserved gene order collinearity along local regions of chromosomes can provide evidence of homology, and in some cases enable determination of whether two regions diverged as a result of speciation or a large-scale duplication event (Drillon et al., 2020). Combined, evidence of gene collinearity and sequence similarity should improve the ability to classify paralogous and orthologous relationships beyond either approach in isolation. Integrating synteny and collinearity into comparative genomics pipelines also physically anchors the positions of related gene sequences onto the assemblies of each genome. For example, by exploring only syntenic orthologs it is possible to examine all putatively functional variants within a genomic region of interest, even those in genes that are absent in the focal reference genome (Lovell et al., 2018). Such a ‘pan-genome annotation’ framework (Lovell et al., 2021a) permits access to multi-genome networks of high-confidence orthologs and paralogs, regardless of ploidy or other complicating aspects of genome biology. Here, we present GENESPACE, an analytical pipeline that explicitly links synteny and sequence similarity to provide high-confidence inference about networks of genes that share a common ancestor and represents these networks as a ‘pan-genome annotation’. We then leverage this framework to explore gene family evolution in flowering plants, mammals, and reptiles. Results and discussion GENESPACE syntenic orthology methods to compare multiple complex genomes Comparative genomics across the complex evolutionary histories of eukaryotes typically requires equally varied input and analytical pipelines depending on researchers’ goals and study systems. For example, synteny between closely related haploid assemblies is often inferred by exploring only 1:1 reciprocal best-scoring hits with MCScanX (Wang et al., 2012). Alternatively, polyploid genomes are typically split into subgenomes so that homeologs are viewed by clustering algorithms like OrthoFinder (Emms and Kelly, 2019; Emms and Kelly, 2015) as orthologous and not paralogous. While expert knowledge that informs these analytical decisions can dramatically improve precision, this knowledge is not available in many systems. These issues boil down to a simple circular problem: a priori knowledge of gene copy number is needed to effectively infer orthology and synteny, yet measures of synteny and orthology are needed to infer copy number between a pair of sequences. GENESPACE resolves this circular problem by operating on a foundational assumption: homologs should be exactly single copy within any syntenic region between a pair of genomes. There are two major violations of this assumption that cause copy number variation within a syntenic block: (1) tandem arrays and (2) gene PAV. GENESPACE addresses these complexities directly in two ways. First, physically proximate multigene families (hereon, ‘tandem arrays’, Box 1) are condensed to the physically most central gene of the array. Gene rank order along the genome is recalculated on these ‘array representative’ genes, effectively masking copy number variation due to tandem arrays. Second, synteny is inferred in a pairwise manner only using ‘potential anchor’ protein BLAST hits (hereon ‘hits’) where both the query and target genes are in the same orthogroup. The rank-order positions of these potential anchor genes are also recalculated prior to synteny inference, which effectively masks orthogroups missing a gene in one genome (i.e., PAV). Thus, GENESPACE operates on OrthoFinder-derived orthogroups with one-to-one relationships for any accurately defined syntenic region, regardless of ploidy or level of sequence conservation. Given GENESPACE’s reliance on syntenic regions between genomes, errors in syntenic block coordinates can have major effects on downstream estimates. Therefore, we crafted a sensitive pipeline to infer syntenic regions (see pipeline overview in Figure 1). To demonstrate its functionality, we ran GENESPACE on seven pairs of genomes spanning closely related maize cultivars to ancient mammal–bird divergence (~300 M ya; Table 1, see Materials and methods). The three comparisons between vertebrate genomes have no recent history of WGDs, while WGDs predated evolutionary divergence of the four plant contrasts: maize is a 12 M ya paleo-tetraploid, cotton is a 1.6 M ya meso-tetraploid, and the ~70 M ya Rho WGD predated grass diversification. Paralogs derived from these WGDs notoriously obscure contrasts between orthologs in these systems. As such, we treated all genomes as haploid and did not include an outgroup; therefore, orthogroups should not include homeologs since the genomes share a common ancestor that arose after the WGD (see ‘additional considerations’ methods section: Outgroups and the phylogenetic context of orthology inference). In each run, we contrasted GENESPACE-derived orthogroups and syntenic blocks to the defaults from OrthoFinder and MCScanX, respectively (Table 1). Since it is a common practice to refine hits prior to running MCScanX, we also included syntenic block coverage from MCScanX run on orthogroup-constrained hits where both the target and query genes must be in the same orthogroup. To contrast each approach, we calculated the percent of genes in (1) orthogroups and (2) syntenic blocks that were placed on exactly one chromosome per genome (Table 1). Figure 1 Download asset Open asset GENESPACE synteny and pan-genome annotation methods. (A, grey panel) GENESPACE runs and parses OrthoFinder results into a synteny-constrained pan-genome annotation. (B, purple panel) Chromosome, gene rank order, and orthogroup membership are added to BLAST hits, which allows direct integration between estimates of orthology and synteny. The three dotplots present the efficacy of GENESPACE syntenic blocks by exploring a particularly challenging region on human (x-axis) and chimpanzee (y-axis) chr. 6. Each point is a BLAST hit rank-order position, colored by syntenic block; colors are recycled if there are more than eight blocks. (C, green panel) Synteny-constrained orthogroups and optionally non-syntenic orthologs are decomposed into a pan-genome annotation where each orthogroup is placed at its inferred syntenic position. Table 1 Comparison of synteny and orthogroup methods. To test the precision of GENESPACE syntenic orthogroups estimates, we contrasted seven pairs of haploid genome assemblies. We present the percent of genes that were found in an orthogroup that hit a single chromosome per genome from the default OrthoFinder and GENESPACE runs. The precision of syntenic block breakpoint estimates was calculated similarly, where the percentage of genes that are placed in a single syntenic block per genome are presented for MCScanX run on all hits, those where both the query and target genes are in the same orthogroup (‘OG’) or via the GENESPACE pipeline. (a) % genes in single-copy OGs(b) % genes in single-copy syntenic blocksAge (~M ya)OrthoFinderGENESPACEMCScanXMCScanX OGGENESPACEB73 vs. B97 maize*<0.0151.573.650.879.093.4Human Hg38 vs. T2T0–0.187.795.981.195.097.7Cotton*,+0.535.685.72.714.196.2HAL2 vs. FIL2 panicgrass*1.174.883.262.389.392.0Human-chimpanzee781.190.278.691.293.3Sorghum-Brachypodium*5046.750.249.367.476.3Human-chicken31066.768.566.471.273.0 * The plant genomes all have one or more WGDs that predate divergence of the genomes,.+Cotton species Gossypium barbadense and G. darwinii have the most recent WGD of ~1.6 M ya, which causes a large number of blocks to be included as two copies; to avoid confusion between subgenomes, blkSize, and nGaps parameters were increased from 5 (default) to 10 genes. For every pair of genomes, GENESPACE produced a greater percentage of single-copy syntenic blocks and genes in single chromosome orthogroups than either OrthoFinder or MCScanX in isolation. GENESPACE also outperformed simple integration between the two methods through MCScanX on orthogroup-constrained hits. The improved performance of GENESPACE synteny-constrained orthogroups was most subtle between highly diverged haploid animal assemblies. For example, in the comparison between human and chicken, GENESPACE resolved 2% and 1.8% more single-copy orthogroups and syntenic blocks than default OrthoFinder and orthogroup-constrained MCScanX, respectively. In contrast, the benefits of GENESPACE were most pronounced in recently diverged genomes with a history of WGDs. Single-copy orthogroups between two meso-tetraploid cotton species that share a WGD that predated speciation by ~1 M ya, were uncommon in the default OrthoFinder run (35.6%) but far more prevalent in GENESPACE syntenic orthogroups (85.7%). Similarly, homeologs derived from the cotton WGD impacted estimates of syntenic blocks: only 14% of the genomes were single-copy syntenic in orthogroup-constrained MCScanX but 96.2% were single copy in GENESPACE syntenic blocks. Combined, these results demonstrate significant flexibility and utility of GENESPACE across a range of evolutionary histories and divergence. It is important to note that some evolutionary processes, including small-scale translocations, can cause true orthologs to exist outside of syntenic regions. Closely related genomes without a history of WGDs tend to have few non-syntenic orthologs. For example, there are 1096 non-syntenic orthologs (6.3% of all orthologs) between human and chimpanzee. In contrast, the 50 M ya diverged Sorghum and Brachypodium genomes have 9002 non-syntenic orthologs, many of which are the result of over-retained Rho WGD-derived paralogs (see below). Since the non-syntenic orthologs can be important in some situations, GENESPACE embraces this complexity by including and flagging non-syntenic orthologs within the pan-genome annotation (Figure 1). Synteny-anchored exploration of vertebrate sex chromosomes GENESPACE facilitates the exploration and analysis of sequence evolution across multiple genomes within regions of interest (e.g., quantitative trait loci [QTL] intervals, see the next section). One particularly instructive example comes from the origin and evolution of the mammalian XY and avian ZW sex chromosome systems. To explore these chromosomes, we ran GENESPACE on 15 haploid avian and mammalian genome assemblies, spanning most major clades of birds, placental mammals, monotremes, and marsupials with available chromosome-scale annotated reference genomes (See Materials and methods). We also included two reptile genomes as outgroups to the avian genomes. The heteromorphic chromosomes (Y and W) are often unassembled, or, where assemblies exist, lack sufficient synteny to provide a useful metric for comparative genomics. As such, we chose to focus on the homomorphic X and Z chromosomes, which have remained surprisingly intact over the >100 M years of independent mammalian (Murphy et al., 1999) and avian evolution (Zhou et al., 2014; Figure 2, Figure 2—figure supplement 1). Figure 2 with 1 supplement see all Download asset Open asset Sex chromosome syntenic network across 17 representative vertebrate genomes. The plot was generated by the plot_riparian GENESPACE function. Genomes are ordered vertically to minimize the number of translocations between each pairwise combination. Chromosomes are ordered horizontally to maximize synteny with the human chromosomes [X, 1–22]. Regions containing syntenic orthogroup members to the mammalian X (gold) or avian Z (blue) chromosomes are highlighted. All sex chromosomes are represented by red segments while autosomes are white. Chromosome segment sizes are scaled by the total number of genes in syntenic networks and positions of the braids are the gene order along the chromosome sequence. See Figure 2—figure supplement 1 for the full synteny graph including autosomes and chromosome labels. While the same or similar genomic regions often recurrently evolve into sex chromosomes, perhaps due to ancestral gene functions involved in gonadogenesis, evidence about the nonrandomness of sex chromosome evolution is still contentious (Kratochvíl et al., 2021). Given our analysis, the avian Z chromosome clearly did not evolve from either of the two reptile Z chromosomes sampled here, but instead likely arose from autosomal regions or unsampled ancestral sex chromosomes. The situation in mammals is less clear, in part because both reptile genomes are more closely related to avian than mammalian genomes, which makes ancestral state reconstructions between the two groups less accurate. Nonetheless, the mammalian X and sand lizard Z chromosomes partially share syntenic orthology, an outcome that would be consistent with common descent from a shared ancestral sex chromosome or autosome containing sex-related genes. The shared 91.7 M bp region between the human X and sand lizard Z represents 59.0% of the human X chromosome genic sequence. The remaining 64.0 M bp of human X linked sequences are syntenic with autosomes in the sand lizard and garter snake genomes (Figure 2). The eutherian mammalian X chromosome is largely composed of two regions, an X-conserved ancestral sex chromosome region that arose in the common ancestor of therian mammals, and an X-added region that arose in the common ancestor of eutherians (Ross et al., 2005). Consistent with this evolutionary history, the X chromosome is syntenic across all five eutherian mammals studied here. Further, a 107.2 M bp (68.8%) segment of the human X, which corresponds with the X-conserved region, is syntenic with 77.8 M bp (93.9%) of the Tasmanian devil X chromosome and represents the entire syntenic region between the human and all three marsupial X chromosomes (Figure 2). Similarly, the chicken Z chromosome is retained in its entirety across all five avian genomes. The only notable exception being the budgie Z chromosome, which features a partial fusion between the Z and an otherwise autosomal 19.5 M bp segment of chicken chromosome 11 (Figure 2), potentially a chromosome fusion that has not yet been In contrast to conserved eutherian and avian sex chromosomes, the complex sex chromosomes are only partially syntenic between the two sampled genomes. the X chromosomes are ancestral to both and et al., and all are to the mammalian X chromosomes (Figure 2—figure supplement consistent with their independent evolution et al., the entirety of the and M bp of the genic region of the chromosomes are syntenic with the avian Z chromosome (Figure 2). The phylogenetic of the genomes presented evolutionary inference about the origin of these shared sex chromosome the of evolution of sex chromosomes between such diverged an line of synteny to track candidate genes in The Poaceae grass plant family is one of the studied of all eukaryotes and model species and many of the most wheat, and the genetic resources of these and other genomic comparisons among are in part because of an ancient polyploid origin (see the next and because WGDs are a feature of most clades of For example, maize is an M ya and M ya (Lovell et al., and arose about ya et al., 2019). In some gene duplications from have generated genetic diversity that can be for crop in other cases the genetic of trait variation be to sequences that arose in a single Thus, it is crucial to genomics and explicitly explore only the orthologous or regions of interest when for or candidate genes trait variation — a significant challenge in the complex and polyploid grass genomes. To help overcome this challenge and provide tools for grass comparative we a GENESPACE run and an on et al., 2012). to its of orthology and synteny (Figure GENESPACE is to comparisons across species with parameters produced a largely of synteny even across notoriously comparisons like the between the maize subgenomes (Figure the sensitive synteny pipeline by GENESPACE effectively masks paralogous regions like those from the Rho duplication that to all Figure with 1 supplement see all Download asset Open asset genomics in the The GENESPACE syntenic of orthologous regions among eight grass genomes. Chromosomes are ordered horizontally to maximize synteny with and are by synteny to chromosomes. Genomes are ordered vertically by phylogenetic The the of maize gene without syntenic orthologs in each split by the full and genes The of absent genes is in the genomes even when for more gene orthologs, homeologs among the 26 maize genomes, with two quantitative trait loci on chromosome and chromosome regions that in where only the genome that the the and genomes are presented and the region is to the physical and a M bp on either that the into two to a larger number of potential candidate genes, the larger region, with is explored in Figure supplement and copy number variation are presented for two of the three as where each is a genome each is a pan-genome (see Figure and the of each single copy and of the focal genome is For each the to of each genome is as to the of the and molecular can two approaches to the genetic of complex studying variation by a genes of interest or candidate genes from genomic regions of As an example of the exploration of of a candidate genes, we of genes to be involved in the between and et al., the permitting dominance in and increased of the next To this analysis, we pan-genome annotations across the seven to maize, which was the genome in which these genes were This in pan-genome nearly two for each gene in the maize genome. Given that many of these genes were in part because of sequence similarity to genes in and other diverged plant species, it is not that among genes was than the vs. 1 Figure However, these were highly variable among genomes, particularly among the species which far more than the species vs. Figure This is due in part to the increased evolutionary between maize and the species to the other However, when for the level of absent genes in species, the was still very the of species more of these genes at syntenic pan-genome positions than the was than the species (Figure these only a single phylogenetic split in this it is not possible to test evolutionary hypotheses the causes of such PAV. Nonetheless, this result a possible of gene or as an evolutionary for and Like the exploration of a sets of genes, finding candidate genes within a single reference genome and genes with annotations or putatively functional In the case of a a single this is a process where genes within physical of a are the However, many genetic have reference genome sequences for all this offers an to explore variation among functional and which would be with a single reference genome. GENESPACE is for this of and was designed to this problem between the two reference genomes and their (Lovell et al., using synteny to the positions of genes across multiple genomes onto the physical positions of a To this approach, we generated from the maize et al., for these were defined by the proximate gene only in the reference genome et al., with GENESPACE and the recently genomes et al., it is possible to candidate genes present in the genomes of other but either absent or in the reference genome. To explore this we a single-copy synteny across all 26 to the genome (Figure We to focus on where the of a single genome was an to all other Such in
Source: eLife
Author: Detlef Weigel
Institution: Max Planck Institute for Biology
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $2,500
OpenAlex ID: W4297676842