Introduction

In evolutionary biology, the classification of organisms in terms of biological complexity is a fundamental question (McShea, 1996). Genome size and the amount of protein-coding DNA were initially thought to reflect complexity. However, cumulative evidence of the complex relationships in the genotype-phenotype mapping would indicate that the one gene-one protein paradigm is no longer supported and, therefore, genome composition alone is insufficient to quantify organismal complexity (Claverie, 2001; Choi et al., 2020). On one hand, genome sizes display high variability in eukaryotes, even for closely related species (Bennett and Leitch, 2005; Gregory, 2005). On the other hand, a lower-than-expected number of protein-coding genes has been observed in some eukaryotic lineages, a problem known as the G-value paradox (Hahn and Wray, 2002). In humans, it is estimated that 24,000 genes encode a total of 100,000 different proteins (Lander et al., 2001).

Transcript diversification is driven by two main mechanisms: gene duplication (Lynch and Conery, 2000; Holland et al., 2017) and alternative splicing (Bush et al., 2017). Although both mechanisms contribute to phenotype innovation (Su et al., 2006), alternative splicing has been cited as the primary candidate to resolve the G-value paradox (Chen et al., 2014, 2012). It consists of a post-transcriptional mechanism by which different protein isoforms are generated from a single gene by the selective combination of exons and introns (Modrek and Lee, 2002; Keren et al., 2010). Thus, it contributes to increasing biological complexity by the differentiation of entities at different levels of the biological organization, including transcript diversification (Chen et al., 2012; Nilsen and Graveley, 2010), cell differentiation (Fiszbein and Kornblihtt, 2017; Fu et al., 2009) and speciation events, such as morphs (Steward et al., 2022; Grantham and Brisson, 2018), castes (Lyko et al., 2011) or even subspecies (Harr and Turner, 2010).

Alternative splicing profiles are conserved in a species-specific way (Barbosa-Morais et al., 2012). The percentage of alternatively spliced genes is higher in vertebrates than in invertebrates, and only a few examples have been found in bacteria and archaea (Anantharaman et al., 2002; Kim et al., 2006; Schad et al., 2011). By contrast, at least 70% of multi-exon genes undergo alternative splicing in humans, mainly regulated in a tissue-specific way (Pan et al., 2008). However, there are still some unanswered questions regarding the roles played by alternative splicing during genome evolution, its prevalence in the different lineages, and its relation to genome composition. In the present work, we propose a novel measure, namely, the alternative splicing ratio, which quantifies the number of different isoforms that can be transcribed using the same amount of information encoded in the genome and compute it for species spanning the whole tree of life. Accordingly, we perform a comparative analysis of the alternative splicing ratio and identify relations to the genome composition.

Results

Genome composition

A comparison of the genome composition for species spanning the whole tree of life has enabled us to distinguish certain genomic traits characterizing the major evolutionary transitions. This differentiates prokaryotes, unicellular eukaryotes, and the high taxonomic groups of multicellular life forms. Genome, genes, and coding sizes increase proportionally to each other at a lineage-specific level (see Figure 1—figure Supplement 1,Figure 1—figure Supplement 2, and Figure 1—figure Supplement 3). Specifically, we observe a significant correlation between each pair of genome variables in all cases, except for birds, where the genome size is weakly correlated to gene content and coding DNA. Furthermore, the slopes given by the linear regressions differ for each taxonomic group, which suggests key differences in their genome evolution. Figure 1A-C shows the relationship between each pair of genome variables, differentiated by each taxonomic group. Because genome variables are related to each other in an embedding structure, we show in Figure 1D the relation between the percentage of coding composing genes and the respective percentage of gene content in genomes. Complementary to this information, we quantify some statistics characterizing the genome composition found in our data. These results are illustrated in Table 1 and Supplementary file 1.

Figure 1.

Table 1.

In the case of prokaryotes we observe a one-to-one relation among all three genome factors, where at least 97% of the gene content is composed of protein-coding sequences. Thus, genomes of prokaryotes are formed by consecutive single-exon genes, i.e., genes containing only one exon, which encode a unique protein. Despite this result, Table 1 highlights the starting accumulation of non-coding DNA in prokaryotes as intergenic DNA. In particular, we have estimated that genes compose between 61% and 97% of genomes. However, this amount of non-coding DNA is still negligible if compared to eukaryotes, which start to accumulate large amounts of non-coding DNA, both as intergenic DNA and as intronic sequences. In particular, the drastic accumulation of non-coding DNA in eukaryotes is reflected in the deviations from equality observed in Figure 1A-C, which concurs with a genome structure diversification.

Figure 1D shows that the percentage of coding composing genes is the factor that best differentiates single-celled organisms from multicellular ones. It represents between 40% and 99.9% of gene content in unicellular eukaryotes, with typical percentages of about 86.4%. This percentage is much higher than in most multicellular life forms, where coding comprises less than 50% of genes. Specifically, flowering plants have a typical percentage of about 29%, and it decreases gradually for arthropods (14.1%), fish (8.6%), birds (4.8%), and finally mammals, where it comprises only 3%. This result indicates that gene composition of unicellular eukaryotes is closer to that of prokaryotes than multicellular life forms. On the other hand, the percentage of intergenic DNA differs considerably among plants and animals. While it accounts for between 43% and 97% of flowering plants, it drops dramatically in animals, with a range comparable to that of unicellular eukaryotes. Specifically, intergenic regions compose between 17% and 82% of arthropods, 23% and 60% of fish, 35% and 59% of birds, and 38% and 69% of mammals. This result highlights the high accumulation of intergenic DNA in plants, while it maintains at moderate values in animals.

Correlations to Alternative Splicing

With some exceptions, alternative splicing is unrelated to genome factors (see Figure 2—figure Supplement 1, Figure 2—figure Supplement 2, and Figure 2—figure Supplement 3). In the case of genome sizes, only arthropods show a significant correlation, but it is still too weak to draw meaningful conclusions. Conversely, mammals and birds have alternative splicing values positively correlated with the amount of coding and gene content. In particular, we find that alternative splicing and the amount of coding have a correlation coefficient of r = 0.531 in mammals and r = 0.541 in birds. In contrast, the correlation with gene content is r = 0.711 in mammals and r = 0.786 in birds. As we can observe, this correlation becomes very strong for the latter case; therefore, as there is more gene content in the genome of land animals, higher ratios of alternative splicing are detected. According with the definition of the alternative splicing ratio, an increase in the amount of gene composition is not associated to a greater production of protein isoforms, but rather with an increase of transcript diversification. Furthermore, the percentage of coding composing genes appears to be highly correlated to alternative splicing values in mammals (r = −0.631) and birds (r = −0.791), with a negative correlation in both cases (see Figure 2—figure Supplement 5), so these results are in concordance with a key role of intronic sequences in the generation of novel protein isoforms. It is important to recall that these associations are only significant in mammals and birds, which suggests that only sophisticated organisms optimize alternative splicing by increasing intronic sequences within the intron-exon structure. Also, mammals are a unique taxonomic group with alternative splicing highly correlated to both, the number of genes and coding. This fact may indicate a highly conserved gene structure coupled to alternative splicing in highly evolved groups.

The percentage of coding relative to genome sizes does not correlate to alternative splicing for any taxonomical group (See Figure 2—figure Supplement 4). Regarding the proportion of genes composing genomes, we find a lack of correlation with alternative splicing ratios in plants but a high correlation in all groups of animals. As we observe in Figure 2, the corresponding correlation coefficients are r = 0.716 for mammals, r = 0.823 for birds, r = 0.604 for fish, and r = 0.712 for arthropods. This result suggests that alternative splicing may be coupled to a genome-level structure in animals, with increasing activity for gene-rich genomes.

Figure 2.

Evolutionary trends

The average values of the alternative splicing ratio for each taxonomic group are shown in Table 2, together with the standard deviation, the coefficient of variation and the range. As we can observe, the alternative splicing ratio is close to one in all single-celled organisms and increases gradually in flowering plants, arthropods, fish, birds, and mammals. In concordance with its role in the differentiation process, this result suggests that it may only be a relevant mechanism in the genome evolution of multicellularity, especially in highly evolved groups such as mammals. In contrast, the variation coefficients in animals increase with increasing alternative splicing ratios, which suggests that alternative splicing may be coupled to the expected variability appearing throughout the tree-like pattern of the evolution of life. This is also in concordance with the appearance of different modes of alternative splicing with evolutionary time. In the case of plants, we observe a low coefficient of variation, which suggests there is a more constrained mechanism of alternative splicing in plants.

Table 2.

The relation of alternative splicing to genome composition is also analyzed. First, the percentage of intergenic DNA and its relation to alternative splicing for the different taxonomic groups is shown in Figure 3A. While prokaryotes contain less than 30% of intergenic DNA, unicellular eukaryotes and fungi show a percentage that falls within a similar range to that of animals, from 17% to 57%. The absence of alternative splicing characterizes both single-celled groups. At the other end of the spectrum we find plants with a high prevalence of intergenic DNA, in between 50% and 96% of genomes, and with alternative splicing ratios of about 1.554. This result suggests that, while alternative splicing is maintained at moderated values in plants, genome evolution occurs by increasing the intergenic content. A tendency to increase alternative splicing values is only observed in animals, and it is optimized in organisms where the proportion of intergenic DNA represents about 50% of the genome. This result concurs with the constraints imposed by the eukaryotic cell, where the energetic cost of having large genomes is high.

Figure 3.

A clear tendency of genome evolution is observed in Figure 3B, where alternative splicing increases for the taxonomic groups having high intron-rich genes. Specifically, the trend starts from prokaryotes, where alternative splicing is almost absent, followed by unicellular eukaryotes, fungi, plants, arthropods, fish, birds, and finally mammals, where alternative splicing reaches maximum values. This result is also in concordance with the prevalence of intron retention in plants and the evolution of alternative splicing towards more complex modes of splicing, such as exon skipping, found in mammals, where genes usually have a highly conserved intron-rich structure. Interestingly, we observe an opposite tendency between prokaryotes and the most complex lineages, such as mammals, while the other groups distribute as intermediate stages of evolution according to their divergence from mammals. Thus, this may indicate an evolutionary tendency whereby alternative splicing ratios increase in the evolution of complex organisms, shaped by an intron-rich gene structure.

Figure 4 shows the contribution of the alternative splicing ratio to genome size, gene content, coding, and the percentage of coding composing the genome. A qualitative comparison shows that the different taxonomic groups are segregated into different spectrum regions and shape certain evolutionary trends. In all figures, the slopes correspond to variability relations, computed as m_s = s_x/s_y, where s_x denotes the standard deviation of alternative splicing and s_y to the corresponding genome variable. Furthermore, to quantify potential trends we compare the variability of each pair of variables regardless of scale. Thus, we have first computed the variation coefficient for each variable, defined as , and obtained the relation of variability r = d_x/d_y for each taxonomic group, as shown in Table 3.

Figure 4.

Table 3.

As we can observe in Figure 4A, plants assume values in a restricted range of alternative splicing ratios at low values while showing a high variability of genome sizes. Indeed, while a small group of plants have huge genomes, the distribution is skewed to small genomes. Conversely, mammals typically have higher genome sizes than plants but span different alternative splicing values. The same pattern observed in mammals appears in birds, with smaller genome sizes and a shorter range of alternative splicing values. There is no clear tendency in fish and arthropods, in which values span both spectrum ranges. The variability relation is r > 1 for mammals and birds, whereas for the other groups it is r < 1, with plants reaching a minimum value close to zero. A similar pattern shaped by the different taxonomic groups is observed for gene content (Figure 4B). We also compare alternative splicing values to the amount of coding in Figure 4C. On the one hand, mammals, birds and arthropods conserve the relative positions found in the genome size and gene content, whereas fish accumulate higher amounts of coding. Considering the results in Figure 3, we discern that marine animals differ from land animals; they have shorter genomes, higher amounts of coding and, in some cases, a lower percentage of intergenic DNA. Conversely, while plants typically have smaller genome sizes than most animals, their genomes accumulate large amounts of coding DNA, which forms exon-rich genes, and large quantities of intergenic DNA.

Figure 4D shows the percentage of coding DNA composing genomes and its relation to the alternative splicing ratio. In mammals, the coding percentage is between 0.92% and 1.73%. Only about 1% of mammalian genomes comprise protein-coding DNA, while their alternative splicing ratios range from ρ_min = 1.396 to ρ_max = 6.919. Birds show a similar pattern, where the coding percentages are also low, between 2.4% and 3.12%, while splicing ratios span from ρ_min = 1.845 to ρ_max = 3.759. By contrast, fish show a decrease in the maximum value of alternative splicing, which spans the range of ρ_min = 1.185 to ρ_max = 3.247, while showing a higher coding variability. In particular, it composes between the 1.63% and 11.26% of their genomes. Similarly, arthropods have a coding percentage of between 1% and 17.83%, while their splicing ratios are low, ranging from ρ_min = 1.116 to ρ_max = 3.279. Finally, flowering plants are the taxonomic group exhibiting the maximum coding variability, which forms in between 1.28% and 28.11% of genomes but with a maximal reduction in the range of splicing ratios, going from ρ_min = 1.188 to ρ_max = 1.918. As we can observe, the corresponding variability coefficients are r = 0.174 for plants, r = 0.382 for arthropods, r = 0.503 for fish, r = 3.191 for birds, and r = 2.550 for mammals, forming a progressive evolutionary pattern. In plants, alternative splicing does not seem to play a relevant role as compared to mammals, which show an opposite pattern. Moreover, we have found that the other taxonomic groups are between these two spectrums, highlighting the relative roles of genome composition and alternative splicing in the evolution of each taxonomic group. As observed in Table 3, the variability relations have values of r > 1 for mammals and birds in all cases, which indicates a higher variability of alternative splicing compared to the other genome variables. On the other hand, we obtain values of less than one for the other taxonomic groups, with a minimum corresponding to plants. These results suggest that genome evolution follows a progressive trend that goes from prokaryotes, unicellular eukaryotes, fungi, flowering plants, arthropods, fish, birds and finally mammals, where the relative roles of introns, exons and intergenic regions become more constrained in highly evolved organisms.

Discussion

Genomes of prokaryotes are mainly composed of single-exon genes; therefore, increasing genome sizes reflects an increase in phenotype innovation. However, this scenario is very different in eukaryotes. Despite the increased genome size, the amount of protein-coding DNA seems to be limited to a size of about 10 MB, which may be related to biophysical constraints (Choi et al., 2020). Thus, the limitation of forming new proteins may be overcome by the selective combination of exons through alternative splicing. This has contributed to the transition from a one-to-one relationship between the genotypic and phenotypic spaces in prokaryotes to a complex relationship in eukaryotes. Genome architecture has allowed this transition in eukaryotes, which differ from bacteria and archaea in different aspects. The organization of DNA into chromosomes and the presence of the nuclear membrane have constrained genome evolution of eukaryotes (Archibald, 2015; Thomas, 1971; Volff and Altenbuchner, 2000). In particular, the compartmentalization of the eukaryotic cell has provided physical support for the accumulation of non-coding DNA and the development of new modes of alternative splicing (Ros-Rocher et al., 2021; Dey et al., 2014; Mattick, 2003). Various authors highlight the role of alternative splicing in the differentiation process, even at varying levels of the biological organization (Chen et al., 2012; Nilsen and Graveley, 2010). In concordance with ancestral reconstructions (Anantharaman et al., 2002) and comparative studies (Chaudhary et al., 2019), our results suggest that alternative splicing is almost exclusive to multicellular life forms.

Genome composition in eukaryotes, however, differs among the different taxonomic groups. As summarized in Table 4, we observe the starting accumulation of non-coding DNA in the eukaryotic cell, reaching very high levels in multicellular organisms. In particular, the coding DNA of unicellular eukaryotes makes up about 50% of genomes, but it decreases drastically for all multicellular organisms, comprising less than 2% in mammals. In general, we observe that coding, gene, and genome size increase proportionally to each other at the lineage-specific level. Although this relationship is linear in almost all cases, the slopes differ for each taxonomic group. On the one hand, plants accumulate large amounts of intergenic DNA, composing about 74% of genomes, and have high exon-rich genes compared to animals. On the other hand, intergenic DNA in animals is much lower, about 50%, and coding reduces drastically, making up less than 14% of genes for all animal groups.

Table 4.

The relative roles of sequence duplication and alternative splicing show lineage-specific patterns and shape a trend where the role of alternative splicing is relatively low in plants, followed by arthropods, fish, birds, and finally mammals, where alternative splicing plays a key role and coding is maintained at low values. The high variability of genome factors characterizing eukaryotic genomes and its coupling to alternative splicing can be partially explained by the limitations imposed by cell physiology. Genome sizes fluctuate through the expansion and reduction of DNA sequences coupled with cell activity (Cavalier-Smith, 1978). Whole-genome duplication and poly-ploidization are recurrent mechanisms in plants and are the mainly responsible for huge genomes (Clark and Donoghue, 2018; Bennett and Leitch, 2005). About 70% of plants are polyploids (Soltis et al., 2003), and most species in the plant kingdom have undergone at least one round of whole-genome duplication (Clark and Donoghue, 2018). This coupling of genome sizes to cell activity may explain why plants, which are less restricted in terms of storing DNA sequences compared to animals, do not prioritize alternative splicing as a mechanism of transcript diversification. In particular, we observe that the alternative splicing ratio is fixed at a moderate value in all flowering plants, a much lower value than in most animals. By contrast, we observe a high variability in their genome composition, suggesting that plants’ primary mechanism of genome evolution is by expanding their genomes. By contrast, animals are more constrained in terms of energy consumption, so maintaining excessive amounts of repeat DNA requires a high energy cost to the cell (Choi et al., 2020). As a counterbalancing response to the expansion of genomes, repetitive sequences are reduced through a process termed down-sizing (Doyle and Coate, 2019), which prevents the uncontrolled expansion of genomes and controls the proper functioning of the cell (Francis et al., 2008; Bennett and Leitch, 2005; Gregory, 2005). In animals, it has been observed that genome sizes are continuously adjusted to cell activity (Cavalier-Smith, 1978; Knight and Beaulieu, 2008), which may explain why alternative splicing is the most relevant mechanism of transcript diversification. These results show that animals and plants follow very different evolutionary pathways and suggest that alternative splicing reflects some organismal complexity in animals alone.

The different mechanisms of transcript diversification shaping a trend that goes from plants, arthropods, fish, birds, and mammals is also in concordance with the fact that alternative splicing has evolved and diversified by different modes. The molecular processes behind alternative splicing can be described in terms of how introns and exons are defined, i.e., the constraints imposed by the splicing recognition machinery (Xing and Lee, 2006; Keren et al., 2010). In particular, it is observed during the evolution of a shift from intron to exon recognition, associated with increased intron size for highly evolved organisms (Rogozin et al., 2012; Schwartz et al., 2009, 2008). Among the most important modes of alternative splicing we find exon skipping, a mechanism by which some exons are spliced out of the transcript to generate mature mRNA. This is the most frequent mode in higher eukaryotes, and its prevalence gradually decreases further down the eukaryotic tree, being very rare in lower eukaryotes (Kim et al., 2008). Intron retention is other mechanism that occurs when an intron is retained in the mature mRNA transcript. While it only represents about 5% of events in animals, it is the prevalent mechanism in plants, fungi, and protozoa (Jacob and Smith, 2017; Alekseyenko et al., 2007). In accordance with previous results, we observe a gradual increase in alternative splicing ratios coupled with a rise in intron-rich gene composition, ranging from unicellular eukaryotes, where genes are mainly composed by single-exons and alternative splicing does not play a key role, to flowering plants, arthropods, fish, birds, and finally mammals. A comparison of these values with the relative amounts of genome composition, summarized in Table 4, indicates that alternative splicing is closely linked to gene composition structures at each specific lineage level. Thus, this progressive trend differentiates each of the taxonomic groups in terms of their genome compositional structure, where alternative splicing is maximized for mammalian species with high intron-rich genes and an intergenic DNA composition of about 50% of genomes. Because exons flanked by long introns are more prone to exon skipping, this result concurs with a prevalence of exon skipping in mammals and its association with an intron-rich gene structure, which becomes increasingly conserved in highly evolved organisms. Finally, the variability of alternative splicing values increases for increasing average values, suggesting lineage-specific evolutionary trends.

In summary, transcript diversification by the selective combination of intron and exon sequences is reflected in the alternative splicing ratios. So, we propose it as a potential candidate to quantify organismal complexity. We observe a gradual increase of alternative splicing ratios toward highly evolved taxonomical groups. In particular, we find low values for flowering plants (ρ = 1.55), and an increase for arthropods (ρ = 2.16), fish (ρ = 2.26), amphibians (ρ = 2.38), birds (ρ = 2.83) and finally mammals (ρ = 3.05). In the case of single-celled organisms, alternative splicing does not appear to play a crucial role in genome evolution, thus we observe values close to one in most cases. The genome compositional trait that best differentiates multicellular life forms to single-celled organisms is the presence of multi-exon, intron-rich genes, which seems to occur coupled to increasing transcript diversification through alternative splicing. This result suggests that alternative splicing is only a relevant mechanism in multicelular life forms, where differentiation regulated at the genome level becomes crucial.

We find a strong negative correlation between alternative splicing ratios and the percentage of coding composing genes in mammals and birds, which highlights the key role of intronic sequences in highly evolved organisms. In general, alternative splicing is maximized within a genome structure that maintains about 50% of intergenic DNA and with intron-rich genes. We also observe an interesting pattern, whereby plants maintain alternative splicing values at moderate values while showing high variability in their genome composition, specifically in the percentage of coding. The opposite trend is observed for mammals, which show a highly conserved genome compositional structure, with a percentage of coding around 1.35%, while presenting a high variation in alternative splicing values. Finally, the tendencies for other taxonomic groups lie between these two spectrums, highlighting the relative roles of sequence duplication and alternative splicing. Indeed, the different modes of alternative splicing and their association to genome compositional structure are in concordance with the energy contraints imposed by the cell. This may explain why sophisticated groups, such as mammals, evolve towards an efficient organization of information processing, while other groups play with the relative roles of storage capacity and information-processing efficiency. Indeed, plants, which have less energy constraints if compared to animals, seem to play with large accumulations of intergenic DNA.

Methods

Genome variables

We consider three genome variables characterizing DNA sequence composition: genome size, gene content, and coding. While gene content refers to the amount of DNA composing all genes, coding represents the amount of DNA within genes that codes for a protein. Furthermore, we defined a novel measure, the alternative splicing ratio, which quantifies the ratio of the different protein isoforms (CDSs) that can build up through the alternative combination of protein-coding sequences. Figure 5 shows a representative example of how it is computed.

Figure 5.

Mathematically, it is defined as follows. We consider the genome as the sequential order of nucleotides, where the ith position of a genome is denoted as G(i). We define f(i,j) as the number of times the nucleotides i and j appear together in a CDS. Thus, f (i,i) corresponds to the number of different CDSs where nucleotide G(i) is inserted. The symmetric matrix M_ij =f (i,j), the transcription matrix, captures the connectivity among nucleotides shaping the genotype-phenotype map. The binary version of the transcription matrix is:

Thus, the alternative splicing ratio is defined as:

where tr(⋅) denotes the trace of a matrix, i.e., the sum of the diagonal elements.

Data source

All genome features are computed from high-quality and whole-genome assembly annotated files from the NCBI database (Kitts et al., 2015). All files have been downloaded the December of 2021, except for the human genome. Due to the recent advances in the genome sequencing of Homo sapiens, we have downloaded the associated annotated file the December of 2023. We have only considered files assembled at the genome level in order to reduce noisy data. All files are in GFF3 format, which provides well-structured data in tab-delimited columns. In particular, information is provided about the starting and ending positions of the following genome attributes: regions, genes, mRNAs, and CDSs, which are all organized in a tree-like pattern. Thus, the genome variables have been computed by counting the number of positions associated with each attribute. Note that we have only considered CDSs with a well-defined parent-child relationship. Pseudogenes and duplicated genes have been excluded from this study. It is important to recall that estimation of alternative splicing ratios can be highly biased. On the one hand, various methods exist in genome sequencing and data assembly, which can be a source of noise. On the other hand, there is a significant bias causing higher alternative splicing ratios for the more frequently studied species.

We have collected annotation files for organisms spanning the whole tree of life. However, the sampling process was limited by the availability of whole-genome annotation files, which reduced the number of organisms to 449 in the case of eukaryotes. Regarding bacteria and archaea, we randomly collected 396 and 288 species, respectively. We also excluded the organisms with extreme values from the study, such as Triticum dicoccoides, which has a genome size larger than 900 Mb. Also, in order to get statistical significance in the different analysis, we need a high number of species representing each taxonomical group. We had only eight organisms representing Amphibians, so we have excluded them from the study. For the same reason we also restricted the plant kingdom to flowering plants. This procedure has resulted in a total of 423 eukaryotes, which are classified as follows, according to the NCBI Taxonomy Database (Federhen, 2011, 2014): 70 mammals, 26 birds, 88 fish, 61 arthropods, 75 flowering plants, 69 Fungi, and 34 unicellular eukaryotes (see Figure 6).

Figure 6.

Additional files

Supplementary files

• Supplementary file 1. Supplementary tables. Range, average, variance, and coeficient of variation of genome sizes (Table S1), gene content (Table S2), coding DNA (Table S3), and alternative splicing (Table S4), computed for each taxonomical group.

Data availability

Availability of data and materials: Data has been downloaded from the NCBI database (Kitts et al., 2015). The list of species under study, code, and the datasets generated in this study are available in Zenodo, at https://zenodo.org/doi/10.5281/zenodo.10064095 (de la Fuente, 2023).