Organelle genome evolution in the early‐diverging brown algal lineage, Ishigeales Authors: Hayeon Hwang, Hocheol Kim, Seok‐Wan Choi, Hiroshi Kawai, Hwan Su Yoon Published: 2026 ABSTRACT AbstractThe order Ishigeales is an early‐diverging lineage of brown algae that exhibits remarkable morphological diversity ranging from simple filamentous forms to robust terete and foliose thalli, which may reflect complex evolutionary histories in their organelle genomes. However, limited organellar genomic data for the Ishigeales, presently confined to Ishige okamurae, have hindered a comprehensive understanding of organelle genome dynamics in that order. To address this, we de novo assembled the complete plastomes and mitogenomes of the four Ishigeales species: Pilinia rimosa, Petroderma maculiforme, I. okamurae, and I. foliacea. Phylogenetic analyses based on plastome and mitogenome datasets using both concatenation and coalescent‐based approaches supported the divergence order from the less complex filamentous morphology of Pi. rimosa to the more morphologically complex terete and foliose Ishige species, yet it also revealed a phylogenetic conflict at the branching point of Pe. maculiforme, suggesting an intricate evolutionary history in filamentous species. Both plastomes and mitogenomes exhibited extensive genomic rearrangement. Notably, mitochondrial genome expansion was observed in Pi. rimosa and Pe. maculiforme, and massive cox2 gene duplication was found in Pi. rimosa. These expansions were likely driven by the proliferation of tandem repeats, and most copies are likely under relaxed purifying selection. Analysis of synonymous substitution rate to the nonsynonymous substitution rate revealed the rapid divergence in several plastid genes between I. okamurae and I. foliacea. In particular, the DPOR gene family exhibited a signal of relaxed purifying selection, potentially linked to their distinct morphological changes in Ishige species. Overall, our findings reveal that Ishigeales underwent lineage‐specific differentiation and dynamic evolutionary patterns in the organelle genomes along with morphological variation. Abbreviations brown algal crown radiation coding sequence cetyltrimethylammonium bromide dark‐operative protochlorophyllide oxidoreductase incomplete lineage sorting nonsynonymous substitution rate synonymous substitution rate light‐dependent protochlorophyllide oxidoreductase large single‐copy maximum likelihood Open reading frame small single‐copy INTRODUCTION Brown algae (Phaeophyceae) are an ecologically and economically important marine lineage that forms marine forest providing shelter and habitat for diverse marine organisms. This lineage diverged from the simple filamentous Schizocladiophyceae approximately 450 million year ago, and then the brown algae diversified from simple filaments (Discosporangiales) to complex multicellular orders (e.g., Fucales, Laminariales; Bringloe et al., 2020; Choi et al., 2024; Silberfeld et al., 2010). The ancestral filamentous brown algae underwent a major expansion in morphological complexity during the onset of the Mesozoic Era (ca. 250 million years ago). This period followed the Permian–Triassic mass extinction, which decimated red and green algae that had previously dominated the ocean biosphere (Payne & Clapham, 2012). The breakup of Pangea created new coastal habitats, while nutrient regimes characterized by high phosphate and low iron favored the diversification of chlorophyll‐c–containing algae (brown algae, diatoms, haptophytes, etc.; Knoll et al., 2007). These conditions enabled the remarkable morphological diversification, leading to the brown algal crown radiation (BACR; Choi et al., 2024; Denoeud et al., 2024). However, whether similar evolutionary pressures also shaped the early‐diverging lineages of brown algae has remained unclear until now. The Ishigeales is an early‐evolving order of brown algae that diverged approximately 300 million years ago (Cho et al., 2004; Choi et al., 2024; Denoeud et al., 2024). Interestingly, the genera within the Ishigeales display remarkable morphological diversity, sequentially ranging from simple filamentous forms to robust, terete and foliose thalli (Figure 1). The Ishigeales consist of four genera: Pilinia, Petroderma, Diplura, and Ishige. Pilinia rimosa forms a uniseriate filament with sparse branching. Petroderma maculiforme and Diplura sp. also form uniseriate filaments, but these are packed, forming a dense crustose thallus. Ishige okamurae produces larger multiseriate cylindrical (terete) thalli, and I. foliacea possesses large, multiseriate foliose thalli (Cho et al., 2004; Kawai et al., 2021; Lee et al., 2009; Oteng'o et al., 2020; Poong et al., 2013). Therefore, the Ishigeales encompasses a wide range of morphological forms within a single clade. This makes the Ishigeales an appropriate model for investigating the evolution of complex multicellularity. Morphological diversity and phylogenetic relationship within the Ishigeales. (a) Morphology and cellular organization of Pilinia rimosa, Petroderma maculiforme, Ishige okamurae, and Ishige foliacea. (b) A maximum‐likelihood phylogenetic tree of brown algae using 141 plastid genes. The order Ishigeales, which represents an early‐diverging lineage within brown algae, is highlighted with a yellow box. Bootstrap support values of 100% are marked with an asterisk (*). Organelle genomes exhibit dynamic genomic variation (e.g., changes in gene content and segmental rearrangements), which can reflect lineage‐specific evolutionary histories and responses to environmental pressures (Mower et al., 2012; Starko et al., 2021). Importantly, morphological diversification in multicellular eukaryotes is primarily driven by the nuclear genome, and organelle genomes are not expected to be the direct genetic determinants of complex morphological traits. Nevertheless, because plastids and mitochondria play central roles in energy metabolism, photosynthesis, and stress responses, organelle genome variation is often observed in association with ecological adaptation and diversification within a lineage, including changes in morphology and life‐history strategies (Marechal & Brisson, 2010; Odahara et al., 2021; Wu et al., 2024). For example, organelle genome expansion was correlated to increased organismal size in volvocine green algae (Smith et al., 2013), and mitochondrial gene rearrangements and amplifications were reported in Antarctic notothenioid fishes inhabiting extreme cold environments (Minhas et al., 2023). These examples suggest that changes in organelle genome architecture can occur during lineage diversification under distinct environmental or physiological conditions, although such patterns remain poorly characterized across many algal groups. In brown algae, plastid (plastome) and mitochondrial (mitogenome) genomes exhibit contrasting structural dynamics (Choi et al., 2020; Starko et al., 2021). The plastomes show recurrent structural rearrangements across orders, particularly in both the large single‐copy (LSC) and the small single‐copy (SSC) regions. In contrast, the mitogenomes show relatively conserved gene contents and synteny (Choi et al., 2020; Starko et al., 2021). Consequently, substantial organelle genome variation within a single order is uncommon in brown algae. The Ishigeales represent an early‐diverging order of brown algae that exhibits distinct morphological diversity among closely related species. In this context, organelle genome evolution of Ishigeales may show variation that coincides with morphological differentiation or ecological adaptation, although current data are limited. To date, only one plastome and one mitogenome from Ishige okamurae have been reported (Choi et al., 2024; Liu et al., 2019; Wang et al., 2021), which leaves comparative analyses within the order limited. Here, we newly assembled and analyzed the plastomes and mitogenomes of four Ishigeales species (Pilinia rimosa, Petroderma maculiforme, Ishige okamurae, and I. foliacea). By comparing their phylogenomic relationships, genome architectures, and evolutionary histories of gene families, we aimed both to assess whether lineages that have diversified morphologically within a single brown algal order also exhibit unexpected levels of organelle genome variation and to place these patterns in the broader context of organelle genome evolution in brown algae. MATERIALS AND METHODS Sample selection and collection Culture strains of Pilinia rimosa KU‐0554 and Petroderma maculiforme KU‐0227 were obtained from Kobe University Macroalgal Culture Collection (KU‐MACC). Pilinia rimosa and Pe. maculiforme were cultured in L1 medium at 18°C under a 12:12 h light:dark cycle (Guillard & Hargraves, 1993). Ishige okamurae was collected from Sangjogam, Goseong‐gun, Republic of Korea (34.90506° N, 128.15253° E), on 15 September 2023. Ishige foliacea was sampled from Chaeseokgang, Buan‐gun, Republic of Korea (35.63751° N, 126.46163° E), on 11 February 2024. The collected samples were placed in plastic bags filled with seawater and transported to the laboratory in an ice box. The thalli were wiped using a brush and rinsed with sterilized seawater to remove epiphytes and barnacles. The clean samples were frozen using liquid nitrogen and stored in a −80°C freezer. DNA extraction and genome sequencing Total genomic DNAs (gDNAs) for the four species were extracted with a modified cetyltrimethylammonium bromide (CTAB) protocol (Doyle & Doyle, 1987). For whole‐genome sequencing, the extracted gDNAs were sent to JSLink Inc. (Seoul, Korea). Short‐read sequencing libraries were prepared using the Illumina TruSeq DNA nano kit and short‐read data were generated using Illumina NovaSeq 6000 (Illumina, San Diego, United States) platforms. Due to the discovery of an unusual genome structure in Pilinia rimosa, we generated long‐read sequences of Pi. rimosa using the PacBio HiFi Revio platform to confirm the genome structure. Organelle genome assembly and gene annotation Before analyzing the Illumina data, adapter and low‐quality reads were removed using Trimmomatic v0.39 (Bolger et al., 2014). Plastomes and mitogenomes of four species were assembled using NOVOPlasty 4.2 (Dierckxsens et al., 2017). To verify the genomic variations of Pilinia rimosa, the assembled contigs were confirmed with long‐read base assembly using Hifiasm assembler (Cheng et al., 2021). The assembled genomes were polished using Pilon (Walker et al., 2014) with the information obtained from short‐mapping to the genomes using BWA v0.7.17 (Li, 2013) and SAMtools v1.7 (Li et al., 2009). The gene annotation was initially performed using GeSeq server (Tillich et al., 2017) and subsequently verified with BLASTx (Camacho et al., 2009). Phylogenomic analyses To construct robust phylogenetic relationships and examine evolutionary trends related to multicellularity in brown algae, a concatenated data set was constructed using 141 plastid genes from the four Ishigeales species, along with 36 published brown algal plastomes; Schizocladia ishiensis (Schizocladiophyceae) represented the outgroup (Bustamante et al., 2019; Choi et al., 2020; Choi et al., 2024; Graf et al., 2017; Le Corguille et al., 2009; Lee et al., 2013; Liang et al., 2022; Liu et al., 2017; Phillips et al., 2020; Starko et al., 2021; Wang et al., 2013; Yang et al., 2016; Zhang et al., 2015). We generated a concatenated dataset using plastid genes that were present in at least 70% of the sampled taxa. Genes represented in fewer than 70% of taxa were excluded, and missing sequences in the retained genes were coded as gaps during matrix construction. To evaluate phylogenetic conflict and incongruence of the four Ishigeales species and assess the extent of incomplete lineage sorting (ILS), we compared concatenated and coalescent‐based phylogenies of plastomes and mitogenomes, following the approach of Smith et al. (2015). A total of 143 plastid and 34 mitochondrial genes were analyzed, along with seven species representing the outgroup (Choi et al., 2020; Cock et al., 2010; Le Corguille et al., 2009; Liu et al., 2017; Oudot‐Le Secq et al., 2002, 2006; Phillips et al., 2020). Only genes present in at least 70% of taxa were retained for analysis, and missing genes in the remaining taxa were treated as gaps during matrix construction (Figure S1). All protein sequences of the genes were aligned using MAFFT v7.490 (Katoh & Standley, 2013). The alignments were concatenated using Geneious Prime v2024.0.2 (https://www.geneious.com). Maximum likelihood (ML) trees were constructed using IQ‐TREE v1.6.8, with 1000 ultrafast bootstrap replications (‐bb 1000) and the test model (‐m TEST, Kalyaanamoorthy et al., 2017; Nguyen et al., 2015). From the IQ‐TREE log files generated during single‐gene analyses, we extracted the number of alignment columns, parsimony‐informative sites, singleton sites, constant sites for each gene. Coalescent trees were constructed based on gene trees using ASTRAL‐III (Zhang et al., 2018). Gene tree conflicts were estimated using PhyParts (Smith et al., 2015). To estimate overall substitution rates for each plastid and mitochondrial gene in the Ishigeales and the variation of evolutionary rates across lineages, we calculated two metrics: (1) in root‐to‐tip distances, reflecting overall divergence from the ancestor, measured as the deviation from the midpoint maximum root‐to‐tip distance (i.e., the longest path from the root to any terminal taxon) and (2) total tree length, representing gene‐wide substitution rates, calculated by summing the lengths of all branches from the root (Vankan et al., 2022). Based on these metrics, genes were ranked from lowest to highest variation and grouped cumulatively by sequentially adding five genes. Specific gene sets are described in Table S1. For each group, gene sequences were concatenated to generate combined datasets. Absent genes in particular taxa were treated as missing data during matrix construction (Figure S1). The resulting matrix were then used to reconstruct ML trees using IQ‐TREE v1.6.8, with 1000 ultrafast bootstrap replications (‐bb 1000) and the test model (‐m TEST, Kalyaanamoorthy et al., 2017; Nguyen et al., 2015). For each concatenated dataset, bootstrap support was extracted for the node separating Petroderma maculiforme and Ishige and plotted to evaluate how support for this focal divergence changed as increasing numbers of genes were incorporated. Organelle genome analyses Organelle genome maps were generated using the OGDRAW v1.3.1 server (Greiner et al., 2019). Assembled plastomes and mitogenomes of the Ishigeales species were aligned using the progressive Mauve algorithm in Mauve v2.3.1 (Darling et al., 2010). Given that expansions of mitogenomes were observed, tandem repeats were analyzed to investigate their impact on mitogenome expansion in the expanded genomes. Repeat sequences longer than 20 bp were identified using the REPuter server (Kurtz et al., 2001). Repeat sequence visualization within mitogenomes was performed using the circlize R package (Gu et al., 2014). To measure synonymous (Ks) and nonsynonymous (Ka) substitutions and evolutionary histories of gene families of plastomes and mitogenomes, all genes were reverse‐translated from amino acid alignments to codon‐based alignments using Pal2Nal v14 (Suyama et al., 2006). The KaKs Calculator v2.0 was used for Ka/Ks calculation (Zhang et al., 2006). As the dark‐operative protochlorophyllide oxidoreductase (DPOR) gene family was likely important for the ecological adaptation of the Ishigeales, tests for relaxation or intensification of selection were conducted using the RELAX method implemented in the Datamonkey web server (www.datamonkey.org). For each gene, the focal lineage was designated as the test branch, whereas all remaining branches were assigned as the reference branches (Weaver et al., 2018; Wertheim et al., 2014). Verification of cox2 duplication in Pilinia rimosa Since the duplication events of cox2 genes were observed in Pilinia rimosa, we further verified the origin of the cox2 copy and the relationship with duplicated and fragmented genes using BLAST and phylogenetic analysis. Open reading frames (ORFs) were predicted using the ORF finder implemented in Geneious Prime (Kearse et al., 2012), and gene identity was assigned based on basic local alignment search tool (BLAST) analysis. Multiple ORFs were annotated as cox2, and their sequence similarity to the original cox2 gene was evaluated. Any ORFs showing high sequence similarity (approximately ≥87% identity) to the original cox2 were considered duplicated cox2 copies. An ML tree of cox2 genes was constructed using IQ‐TREE v1.6.8 (Nguyen et al., 2015) with 1000 ultrafast bootstrap replications (‐bb 1000) and the test model (‐m TEST, Kalyaanamoorthy et al., 2017; Nguyen et al., 2015). To assess whether the duplicated cox2 copies were likely to be biologically active, relaxation of selective constraints on cox2 copies was evaluated using RELAX, as described above (Weaver et al., 2018; Wertheim et al., 2014). RESULTS General features of plastomes and mitogenomes of Ishigeales Genome sizes of the four assembled plastomes ranged from 130 to 133 kb (Figure S2). Average GC content of plastomes ranged from 30.7% to 33.5%. The plastomes of Pilinia rimosa and Petroderma maculiforme contained 143 and 142 coding sequences (CDSs), respectively, while those of Ishige okamurae and I. foliacea contained 144 CDSs. In Pi. rimosa, the syfB gene was absent, whereas the rpl32 gene was missing in Pe. maculiforme. All plastomes of the four Ishigeales species included 29 transfer RNA (tRNA) genes and six ribosomal RNA (rRNA) genes (Table 1). TABLE 1Comparison of general features of organelle genomes from the Ishigeales, basal brown algae, and the outgroup Schizocladia ischiensis.OrderSpeciesOrganelleGeneral characteristicsRNAsGene loss/gain relative to other speciesNCBI accessionTotal bpGC %IntronsCDSORFtRNArRNAIshigeales Pilinia rimosa Plastid130,48830.70%01430296 syfB (loss)This study rpl21 (two copies) Petroderma maculiforme Plastid132,21632.90%11420296 rpl32 (loss)This study Ishige okamurae Plastid130,00033.50%01440296 rpl21 (two copies)This study Ishige foliaceae Plastid130,50332.80%01440296 rpl21 (two copies)This studySchizocaldiales Schizocladia ischiensis Plastid138,10130.80%01310296 cp30, cp74, psaE, syfB, ycf17, ycf19, ycf34, ycf35, ycf41, ycf42, ycf47, ycf54, ycf65 (loss)MT226925 rpl21 (two copies)Discosporangiales Choristocarpus tenellus Plastid120,42826.50%01360286 cp74, rpl9, rpl32, tsf, ycf34, ycf35, ycf45, ycf46 (loss)NC_085297 rpl21 (two copies)Dictyotales Dictyopteris divaricate Plastid126,09931.20%01392276 cp30, cp74, psbZ, rbcR, rpl32, ycf45, ycf54 (loss)NC_036804 rpl21 (two copies) Dictyoptosis propagulifera Plastid122,89231.00%01400286 psbZ, rbcR, rpl32, ycf45 (loss)NC_085302 rpl21 (two copies)Ishigeales Pilinia rimosa Mitochondria59,75134.60%1510273 atp8, nad6, nad6b, rpl31 (loss)This study cox2 (17 copies), nad1, nad11 (two copies) Petroderma maculiforme Mitochondria47,11938.70%1330193 atp8, nad6b, rpl31, rps11 (loss)This study nad1 (two copies) Ishige okamurae Mitochondria35,51535.30%1404223This study Ishige foliaceae Mitochondria35,43933.30%1371213 nad6b (loss)This studySchizocaldiales Schizocladia ischiensis Mitochondria41,62841%0330263 nad6b, rpl31, rps11 (loss)MT259947Discosporangiales Choristocarpus tenellus Mitochondria32,35130.60%0350223 nad6b (loss)NC_085326Dictyotales Dictyopteris divaricata Mitochondria32,02138.30%0383243 nad6b (loss)NC_043845 Dictyotopsis propagulifera Mitochondria30,99526.80%0350243 nad6b (loss)NC_085329 Dictyota dichotoma Mitochondria31,61736.50%0383253 nad6b (loss)NC_007685 Comparison of general features of organelle genomes from the Ishigeales, basal brown algae, and the outgroup Schizocladia ischiensis. The assembled mitogenomes exhibited high variability in size. The mitogenome sizes of Ishige okamurae and I. foliacea were 35,515 bp and 35,439 bp, respectively. In contrast, Pilinia rimosa (59,748 bp) and Petroderma maculiforme (47,119 bp) were larger than the two Ishige species (Figure S2) and exceeded the mitogenome sizes observed in closely related brown algae. The average GC content of mitogenomes ranged from 33.3% to 38.7%. The mitogenome of Pi. rimosa comprised 51 CDSs, 27 tRNAs, three rRNAs, and one intron. Among the 51 CDSs, 17 duplicated copies of cox2 were identified, whereas atp8, nad6, nad6b, and rpl31 were absent from the mitogenome of Pi. rimosa. Petroderma maculiforme comprised 33 CDSs, 19 tRNAs, and three rRNAs. This species had a duplicated nad1 gene and one intron in nad6 but lacked the intron in cox2 and had lost atp8, nad6b, and rps11. In contrast, I. okamurae had 40 CDSs, and I. foliacea had 37 CDSs, one intron in cox2, 22 tRNAs, and three rRNAs in mitogenomes (Table 1). Phylogenetic relationship in Ishigeales species and gene tree conflicts To provide an overview of evolutionary trends in multicellularity in brown algae, a phylogeny was reconstructed using a concatenated dataset of 141 plastid protein‐coding genes. In the phylogeny, multiple transitions from simple filamentous forms to complex multicellular organisms with differentiated tissues were observed, including within the order Ishigeales, the order Sphacelariales, and the overall transition from the Discosporangiales to the Larminariales. Within Ishigeales, a clear transition from filamentous Pilinia rimosa to crustose Petroderma maculiforme and two complex multicellular Ishige species was observed (Figure 1b). Given that similar transitions were inferred in two relatively early‐diverging lineages (Ishigeales and Sphacelariales), we reconstructed a phylogeny with richer taxon sampling using the rbcL gene to evaluate further phylogenetic relationships within these groups. Within Ishigeales, the topology placed Pilinia rimosa as the earliest‐diverging lineage, followed by Petroderma maculiforme, Diplura, and Ishige, which was consistent with the concatenated plastid genes tree, exhibiting a trend toward increasing morphological complexity. The bootstrap value at the Pe. maculiforme and Diplura divergence node was low (<50%, see Figure S3). A similar increase in morphological complexity along evolutionary patterns was also observed in the closely related clade Sphacelariales (Figure S4). To evaluate the gene tree discordance and assess conflicting topologies, we compared concatenated phylogenies with coalescent‐based species trees. For this analysis, 143 plastid and 34 mitochondrial genes from four Ishigeales species and seven outgroup species were included. Both the concatenated and coalescent trees strongly supported the monophyly of the four Ishigeales species from Pilinia rimosa to Ishige species; however, topological discordance was observed at the node representing the divergence of Petroderma maculiforme (Figure 2a,b). Phylogenetic discordance between concatenated and coalescent species tree of the Ishigeales. (a) Phylogenetic trees based on the concatenation (left) and coalescence methods (right) using amino acid sequences of plastomes and (b) mitogenomes. Bootstrap support values of 100% are marked with an asterisk (*). (c) Schematic of tree topologies of concordance tree consistent with concatenated tree and two types of discordance tree at the divergence of Petroderma maculiforme. The proportions of concordant and discordant gene trees supporting each topology are indicated. Estimation of evolutionary rate variation across lineages in the Ishigeales based on (d) plastomes and (e) mitogenomes. Bootstrap support values were calculated using gene datasets (from low to high variation according to two criteria: The root‐to‐tip distances (divergence from the ancestor) and total tree length (overall substitution rate). In the PhyParts analysis, gene trees were scored as concordant if they supported the same bipartition as the corresponding node in the species tree, regardless of alternative overall relationships within each partition. To quantify the source of this conflict, we counted the number of gene trees supporting the concordant Ishigeales topology (i.e., the species‐tree topology) and the two main alternative topologies (Figure 2c). Discordance topology‐1 suggested an alternative scenario in which Petroderma maculiforme diverged first, whereas discordance topology‐2 supported the monophyly of Pilinia rimosa and Pe. maculiforme. Others represented gene trees that do not correspond to any of these three predefined topologies. As a result of comparing their ratios, among plastid genes, 24% supported the concordant topology, whereas the discordant topologies 1 and 2 accounted for 22% and 17%, respectively. Mitochondrial genes exhibited a slightly more concordant pattern, with 47% of loci recovering the concordant topology and 24% and 21% supporting discordant topologies 1 and 2, respectively (Figure 2c). To evaluate whether these topological conflicts were due to insufficient phylogenetic signal, we examined how phylogenetic resolution changed with increasing gene number and sequence variation. Gene variation was quantified using root‐to‐tip distance and total tree length, reflecting rate heterogeneity and overall sequence divergence, respectively. Bootstrap support values at the conflicting node in the ML tree were calculated by sequentially adding genes in sets of five. For plastid genes, bootstrap support for the Petroderma maculiforme–Ishige divergence fluctuated substantially when approximately 10–30 (root‐to‐tip distance) and 15–40 (total branch length) loci were included. During this range, discordant topology 1 were repeatedly recovered; however, as additional loci were incorporated, the concordant topology became stable (Figure 2d). Mitochondrial genes also exhibited discordant topology 1 when only the first five loci were concatenated. As additional loci were included, the concordant topology was consistently recovered and maintained high bootstrap support (Figure 2e). To evaluate whether the datasets contained sufficient phylogenetic signal for reliable tree reconstruction, we examined the number of parsimony‐informative sites in each alignment. Gene trees with lower root‐to‐tip distances and total tree lengths tended to exhibit lower proportions of parsimony‐informative sites (Figure S5 and Table S1). Structural rearrangements in Ishigeales plastomes and mitogenomes To investigate structural rearrangements in plastomes and mitogenomes, we conducted whole genome alignments. The plastomes in brown algae showed distinct patterns of rearrangement depending on their mode of inheritance. In isogamous species, genomic rearrangements occurred even among species within the same order. In contrast, oogamous species largely maintained a conserved structure within their orders but showed gene rearrangements between different orders (Figure S6). In plastomes of Ishigeales, two major inversion events were identified: one occurred during the divergence of Pilinia rimosa and Petroderma maculiforme, and the other was found when the Ishige species diverged from Pe. maculiforme (Figures 3a and S6). Genome structure and contents variation in organelle genomes of the Ishigeales. (a) Comparison of plastome and (b) mitogenome structure among the Ishigeales species using Mauve alignment. Locally collinear blocks represent conserved genomic regions across species. The analysis reveals structural conservation and lineage‐specific rearrangements within the Ishigeales plastomes and mitogenomes. (c) Genome content of plastomes and (d) mitogenomes in the Ishigeales species. Bar graphs show the proportion of different genomic components, including coding sequences (CDSs), intergenic regions, tRNAs, rRNAs and duplicated cox2. Values are presented as percentages of total genome size in each organelle. (e) Distribution of tandem repeat sequences (black line) in the mitogenomes of the Ishigeales species. The abundance and location of repeat sequences in each species were shown. A notable increase in repeat content was observed in intergenic regions. In Pilinia rimosa, repeat sequences were accumulated around the duplicated cox2 genes (pink arrow). The mitogenomes of brown algae generally maintain conserved structures regardless of the mode of reproduction. However, gene rearrangements have been observed as exceptions in the Ishigeales and Sphacelariales (Figure S7). The mitogenomes of the four Ishigeales species exhibited substantial structural variations. In particular, Pilinia rimosa and Petroderma maculiforme showed extensive gene rearrangements (Figures 3b and S7). Mitogenome size expansion and cox2 gene duplication in Pilinia rimosa We confirmed that plastome sizes are broadly similar across brown algae (Figure S6). Consistent with this pattern, the proportions of genomic components were highly conserved across species, with no significant differences in gene number, intergenic region length, or the abundance of tandem repeats (Figures 3c, S8 and S9). In contrast, mitogenome expansion was observed in four brown algal species (Figure S7), including two Ishigeales species (Pilinia rimosa and Petroderma maculiforme), one Sphacelariales species (Phaeostrophion irregularis), and one Ectocarpales species (Pylaiella littoralis). In Pi. rimosa and Pe. maculiforme, the significant mitogenome increase was largely attributed to the accumulation of tandem repeats (black line) in intergenic regions (Figure 3d,e). Pilinia rimosa and Pe. maculiforme harbored 53 and 58 tandem repeats, respectively, which represent markedly higher repeat numbers than those observed in other brown algal species (Figure S9). Notably, tandem repeat sequences between the two species did not overlap (Figure S10). Such extensive tandem repeat expansions were not detected in their sister species, Ishige okamurae or I. foliacea, nor in Ph. irregularis and Py. littoralis that had increased mitogenome sizes (Figures 3e and S11). In addition to repeat accumulation, we identified extensive duplication of the cox2 gene in Pilinia rimosa. A total of 17 cox2 copies were detected, including one original copy corresponding to the conserved genomic position (purple arrow) and 16 duplicated copies (pink arrows; Figure 3d,e). To find the original copy, we examined mitochondrial gene synteny across brown algae. In most species, a single cox2 gene is located between cob and nad4, with rpl31 variably present in this region (Figure S12). In the Pi. rimosa mitogenome, several copies of the cox2 gene were identified. To determine the original copy, BLASTn analysis was performed using cox2 sequences from Ishige okamurae and I. foliacea as queries. A single cox2 copy was located immediately upstream of the nad4 gene, corresponding to the conserved genomic context observed in brown algae (Figure 4a). Furthermore, this copy showed the highest sequence similarity among the cox2 copies. The total length of these duplicated cox2 copies was 10,470 bp, accounting for a substantial proportion of its mitogenome expansion (Figure 3d). Tandem repeats were also concentrated around the duplicated cox2 regions and adjacent non‐coding regions (Figure 3e). Lineage‐specific amplification of cox2 genes in mitogenomes of Pilinia rimosa. (a) Distribution of the duplicated cox2 gene in mitogenomes of Pi. rimosa, showing 16 tandemly arranged copies. (b) The gene tree and sequence alignment of cox2 genes from Pi. rimosa and other brown algae. All amplified cox2 copies form Pi. rimosa‐specific clades, indicating that the duplications occurred uniquely within this species. Bootstrap support values of 100% are marked with an asterisk (*). (c) Pairwise Ka/Ks comparisons between the Ishige okamurae cox2 locus and 17 copies in Pi. rimosa. (d) Relaxation of selective constraints in duplicated cox2 copies of Pi. rimosa. RELAX K estimates are shown for the original cox2 locus and 16 duplicated cox2 copies. Statistical significance (p < 0.05) is indicated by an asterisk (*). To examine the evolutionary relationships among the duplicated cox2 copies, we reconstructed a cox2 gene tree (Figure 4). Phylogenetic analysis confirmed the monophyly of cox2 genes within the Ishigeales. Among these copies, the copy showing the highest sequence similarity to cox2 genes from other brown algae above was positioned basally relative to the duplicated copies. This copy was annotated as the original cox2 copy based on its sequence similarity and phylogenetic position. The duplicated copies in Pilinia rimosa showed substantial sequence divergence from the original copy and formed two distinct groups based on sequence length: one comprising truncated sequences shorter than 200 amino acids (~600 bp), and the other containing longer sequences exceeding 200 amino acids (Figure 4b). To evaluate the potential functionality of the duplicated cox2 copies in Pilinia rimosa, we performed pairwise Ka/Ks analyses between the Ishige okamurae cox2 locus and each of the 17 copies in Pi. rimosa. All copies exhibited Ka/Ks ratios below 1, consistent with purifying selection. Notably, the original cox2 locus showed a Ka/Ks value of approximately 0.1, whereas the 16 duplicated copies displayed higher ratios, ranging from approximately 0.6 to 0.9 (Figures 4c, S13 and Table S2). In addition, RELAX analyses indicated significant relaxation of selective constraints in all duplicated cox2 copies except the original cox2, despite their overall Ka/Ks ratios remaining below 1 (Figure 4d and Table S2). General plastid and mitochondrial gene evolution Nonsynonymous substitution rate (Ka), Ks, and Ka/Ks were analyzed to assess sequence divergence and gene evolution in plastomes and mitogenomes of the Ishigeales through pairwise species comparison. Across all genes, Ka/Ks ratios between species pairs were below 1, indicating that these genes are generally under purifying selection. Overall, mitochondrial genes exhibited higher Ka/Ks ratios than plastid genes (Figure 5a,b). Notably, the pairwise comparison of the two Ishige species showed lower Ka and Ks values but similar Ka/Ks ratios compared to other species pairs (Figures 5a,b, and S14). Pairwise synonymous/nonsynonymous substitution ratios (Ka/Ks) in plastomes and mitogenomes among the Ishigeales. Overall, Ka/Ks ratios in (a) plastomes and (b) mitogenomes. Ka/Ks values are shown for each gene family across all groups of (c) plastomes and (d) mitogenomes. Each box represents the distribution of Ka/Ks values calculated from all pairwise comparisons within each six pairwise group (left). Ka/Ks values are shown for each gene family across all groups, with outliers removed (right). Boxes indicate the interquartile range (IQR); the median is shown as a horizontal line. Asterisks (*) indicate statistically significant differences in Ka/Ks ratios between gene families (p < 0.05). In both plastomes and mitogenomes, the ribosomal protein gene family exhibited significantly higher Ka/Ks values compared to other gene families (Wilcoxon rank‐sum test; plastome: W = 64,369, p < 0.05; mitogenome: W = 5631, p < 0.05). To assess whether this pattern could reflect analytical artifacts, we examined synonymous substitution rates and gene length distributions. Ribosomal genes did not exhibit unusually low Ks values or exceptionally short coding sequences relative to other functional categories, suggesting that the elevated Ka/Ks ratios are unlikely to be explained solely by gene length or rate‐estimation bias (Table S3). In the plastomes, genes involved in ATP synthase, cytochrome complexes, and photosystems did not show significant differences in Ka/Ks across species pairs. However, in the Ishige okamurae / I. foliacea pair, the DPOR (W = 45, p < 0.05) and ribosomal protein genes (W = 7007, p < 0.05) showed significantly higher Ka/Ks ratios, along with lower Ka and Ks values than in other groups (Figures 5c and S14). The RELAX analyses of the DPOR gene family revealed significantly reduced selection intensity in the Ishige lineage, with chlL (K = 0.59, p < 0.05) and chlB (K = 0.71, p < 0.05) showing clear signatures of relaxed purifying selection, whereas chlN did not exhibit a statistically significant shift (K = 0.87, p = 0.078). In the mitogenomes, the I. okamurae / I. foliacea pair showed lower Ka/Ks, Ka, and Ks values than other pairs in all gene families (Figures 5d and S14). Notably, the ribosomal protein gene family showed a significant difference between the I. okamurae / I. foliacea pair and the other pairs (W = 266, p < 0.05). DISCUSSION Morphological transitions in brown algae and organelle gene content variation in filamentous Ishigeales species The Ishigeales have a remarkable transition from filamentous forms to complex thalli. However, this pattern is not confined to Ishigeales; similar transitions have occurred at least three times independently within brown algae, including in Ishigeales and Sphacelariales and along the broader transition from the Discosporangiales to the Laminariales. This repeated emergence of complex multicellularity indicates that morphological complexity in brown algae did not arise from a single evolutionary event but instead evolved convergently across multiple lineages (Figure 1b). Interestingly, two cases of simple to complex multicellularity transition were observed in early‐diverging groups (Ishigeales and Sphacelariales; Figures S3 and S4). Simple filamentous Sphacelaria spp. were basal to the clades of crustose Lithodermatacean species and foliose thallus Phaeostrophilacean species (Kawai et al., 2005; Kawai & Prud'homme van Reine, 1998; Oteng'o et al., 2024). During initial diversification in brown algae, the evolution of complex forms from filamentous ancestors may have occurred more frequently, potentially facilitated by early genomic features and cellular developmental mechanisms (Rana et al., 2019). Within Ishigeales, we observed that gain and loss in organelle gene content—particularly, gene loss—were more pronounced in the filamentous species Pilinia rimosa and Petroderma maculiforme than in the morphologically more complex species Ishige okamurae and I. foliacea. This pattern suggests that dynamic variations in organelle genomes occurred preferentially in early‐diverging, filamentous lineages during the initial stages of diversification, rather than accumulating progressively with increasing morphological complexity (Table 1). Robust phylogenetic relationships within Ishigeales despite gene tree conflicts in filamentous species Since the previous study on Ishigeales phylogeny was conducted based solely on five plastid genes (Kawai et al., 2021), we reconstructed the phylogenetic tree using complete plastomes and mitogenomes to enhance the accuracy and robustness. Our result confirmed the congruent species order—Pilinia rimosa, Petroderma maculiforme, and two Ishige species—as reported previously (Figure 2a,b). Diplura is currently treated as incertae sedis, and its phylogenetic position within the Ishigeales remains uncertain. Sequence data for Diplura in public databases are limited, and phylogenies reconstructed from the available data show low support at the relevant node (Figure S3). In this study, we were unable to obtain a culture or field material of this genus and, therefore, excluded it from our analyses. Thus, further study should re‐evaluate the phylogenetic relationship of all the Ishigeales species using these comprehensive genomic datasets. The recovery of consistent monophyly among the four species in both concatenated trees and coalescent‐based trees indicates that the fundamental species relationships within Ishigeales are generally stable (Figure 2a,b). Nevertheless, some discordance in gene trees was observed at the node separating Petroderma maculiforme in the phyparts analysis. The relatively short branch length between Pilinia rimosa and Pe. maculiforme in the concatenated trees suggests that ancestral polymorphisms have not been fully sorted between these two species, which may have led to ILS (Huang et al., 2025). The distribution of gene‐tree topology frequencies is also consistent with this interpretation. In both plastomes and mitogenomes, the concordant topology accounted for the highest proportion of gene trees; however, two alternative branching scenarios were also supported at substantial frequencies (Figure 2c). Notably, the discordant topologies occurred at comparable frequencies rather than being skewed toward a single alternative, a pattern more consistent with genome‐wide ILS than with directionally biased evolutionary events affecting specific loci (Dias et al., 2020). Such a distribution suggests that different loci do not uniformly reflect a single evolutionary history. To more rigorously evaluate the processes underlying this phylogenetic conflict, additional analyses incorporating nuclear genomic data will be necessary. To assess whether phylogenetic inference changes consistently with increasing gene sampling, we performed ranked‐loci concatenation analyses. Ranked‐loci concatenation analyses showed that phylogenetic discordance was most pronounced when only a small number of genes were included in both plastid and mitochondrial datasets (Figure S5). In plastid datasets, discordant topologies were primarily observed when approximately 10–40 genes were concatenated, whereas in mitochondrial datasets, discordance was detected only in the tree constructed from the first five concatenated loci. This early instability may reflect limited phylogenetic signal, as genes with shorter root‐to‐tip distances and lower total tree lengths tended to contain fewer parsimony‐informative sites. As additional loci were incorporated, the inferred topology progressively stabilized, suggesting that the accumulation of loci strengthened the species‐level signal. Overall, concatenation‐based and coalescent‐based analyses recovered broadly consistent interspecific relationships. Although the early discordance likely reflects limited data and reduced information content, biological processes cannot be entirely excluded. Further nuclear genomic analyses will be necessary to clarify the underlying evolutionary processes. Organelle genome rearrangements associated with inheritance Brown algae exhibit three distinct types of fertilization strategies—isogamy, anisogamy, and oogamy—reflecting a continuum of gamete differentiation (Choi et al., 2024; Luthringer et al., 2014; Motomura et al., 2010). Plastome inheritance in brown algae varies according to gamete type. Biparental transmission of plastomes occurs in isogamous species, whereas uniparental inheritance is observed in oogamous species (Choi et al., 2020; Mignerot et al., 2019). Consistent with this pattern, the plastome structure exhibited inversions among Pilinia rimosa, Petroderma maculiforme, and Ishige. Given that the Ishigeales possess isogamous gametes (Cho et al., 2004), biparental inheritance of plastomes may have also occurred, potentially contributing to genomic rearrangements that we observed (Figures 3a and S6, Choi et al., 2020). Mitogenomes are typically inherited uniparentally (maternally in most cases) regardless of fertilization strategy and, therefore, exhibit highly conserved gene content and synteny across most brown algal lineages (Choi et al., 2020; Greiner et al., 2015; Mignerot et al., 2019; Figure S7). Notably, the filamentous Ishigeales species Pilinia rimosa and Petroderma maculiforme exhibited dynamic changes in mitogenome architecture, contrasting with the conserved patterns observed in other Ishigeales species (Figure 3b). An exception to strict uniparental inheritance has been reported in the isogamous, filamentous Ectocarpus species (Ectocarpales), in which biparental inheritance and mitochondrial recombination can occur (Mignerot et al., 2019). In this context, similar mitogenome rearrangements in Sphacelariales are a comparative example (Figure S7). These observations suggest that atypical organelle inheritance dynamics may occasionally arise in early‐diverging, isogamous, filamentous brown algal lineages. Importantly, because most other isogamous brown algae retain highly conserved mitogenomes, the structural changes observed in the taxa including Pi. rimosa and Pe. maculiforme may have occurred independently. Repeat‐mediated mitogenome expansion and cox2 duplication Plastome size expansion is not typically observed in brown algae, including within the Ishigeales (Figures 3c, S6 and S8). In contrast, mitogenome size expansion was observed in several brown algal lineages. Mitogenome size expansion occurred through various mechanisms, including gene duplication, accumulation of repeat sequences, and insertion of foreign genetic material (Gandini & Sanchez‐Puerta, 2017; Wang et al., 2024). In brown algae, mitogenome expansion in previously published cases has been attributed either to intergenic region expansion in Phaeostrophion irregularis (Sphacelariales) or to the insertion of a foreign group II intron in Pylaiella littoralis (Ectocarpales; Figure S7, Fontaine et al., 1997). In contrast, the mitogenome expansions observed in Pilinia rimosa and Petroderma maculiforme are characterized by extensive proliferation of tandem repeats within intergenic regions, and this pattern has not been reported previously in brown algae (Figures 3b,d and S7). Furthermore, the low sequence similarity of tandem repeats between these two species suggests that mitogenome expansion occurred independently in each lineage rather than through a shared ancestral event (Figure S10). In Pilinia rimosa, mitogenome expansion was additionally associated with extensive duplication of the cox2 gene (Figure 3d,e). Phylogenetic analysis indicated the monophyly among these genes, suggesting that the duplicated genes originated from internal replication events rather than external insertions (Figure 4). Repeat sequences are known to mediate recombination between homologous regions and can facilitate gene duplication (Cai et al., 2024; Kim et al., 2021). Numerous repeats were identified around the duplicated cox2 genes, supporting their possible involvement in the duplication process. Moreover, the high sequence divergence and frequent deletions among duplicated cox2 copies suggest that the duplication events were likely mediated by random recombination of repeats rather than certain selection pressure on cox2 (Figure 4b). Repeat expansion and gene duplication can be related to persistent or repeated environmental stress (Kim et al., 2025; Zhong et al., 2023). The Ishigeales species inhabit the intertidal zone, where they are frequently exposed to diverse environmental changes, such as fluctuations in salinity, desiccation, and UV radiation. Notably, Pilinia rimosa and Petroderma maculiforme are simple filamentous organisms in which all cells are directly exposed to the external environment, unlike the two complex multicellular Ishige species. Simple filamentous organisms may therefore be more vulnerable to genome instability under environmental stress (Knoll, 2011). Additionally, Pi. rimosa lives in freshwater and brackish habitats, whereas Pe. maculiforme has been reported to form lichen associations with fungi in the intertidal environments (Sanders et al., 2004; Waern, 1952). Together, these observations raise the possibility that mitogenome expansion in filamentous Ishigeales reflects lineage‐specific responses to environmental and physiological stress. An alternative scenario, in which such mitogenome features were present in a common ancestor and subsequently lost in most brown algal lineages, cannot be entirely excluded. However, given the strong structural conservation observed in mitochondrial genomes of early‐diverging brown algae (i.e., Choristocarpus tenellus) and other brown algal lineages, the independent emergence of these features in Pilinia rimosa and Petroderma maculiforme represents a more parsimonious explanation, although direct causal links remain to be tested. Patterns of selection across organelle gene families and duplicated cox2 genes in Pilinia rimosa Comparative analyses of Ka/Ks ratios across functional gene categories showed results largely consistent with previous studies (Liu et al., 2024; Starko et al., 2021). Overall, mutation rates were higher in mitogenomes than in plastomes (Figure 5a,b). Among functional categories, ribosomal protein genes showed consistently higher Ka/Ks ratios than other gene families in both plastomes and mitogenomes, primarily driven by increased Ka (Figures 5c,d and S14). Similar acceleration of organelle‐encoded ribosomal protein genes was reported in multiple photosynthetic lineages, including brown algae (Maier et al., 2013; Starko et al., 2021; Weng et al., 2016). Because organelle‐encoded ribosomal proteins interact closely with nuclear‐encoded ribosomal subunits, lineage‐specific changes in ribosome structure or assembly can impose distinct selective constraints on these genes. In flowering plants, elevated Ka in ribosomal protein genes has been interpreted in the context of cytonuclear coevolution rather than adaptive diversification (Maier et al., 2013; Weng et al., 2016). Similar cytonuclear interactions may also have occurred during Ishigeales diversification; however, testing this hypothesis will require integrative analyses of nuclear and organelle‐encoded ribosomal components in further study. In the comparison between Ishige okamurae and I. foliacea, both Ka and Ks values were lower than in other species pairs, likely due to their relatively recent divergence (Figures 5 and S14). Despite this, Ka/Ks patterns between the two Ishige species differed in plastid and mitochondrial gene families. Although mitochondrial gene families remained under strong purifying selection, many plastid gene families displayed comparable or elevated Ka/Ks ratios relative to other species pairs, indicating a relaxation of purifying selection in plastid genes within the Ishige lineage (Figure 5). Collectively, these results suggest that selective constraints on organelle genes have evolved heterogeneously in Ishigeales, with plastomes showing greater flexibility in selective regimes than mitogenomes. Among plastid gene families, the DPOR gene family in the Ishige okamurae and I. foliacea pair showed particularly higher Ka/Ks ratios compared to other species pairs, further supporting the evidence of rapid diversification in plastomes. The DPOR genes are essential for chlorophyll synthesis in low‐light or dark conditions, and this elevated ratio may reflect functional or morphological adaptations specific to light environments (Uthanumallian et al., 2022). Notably, Pilinia rimosa and Petroderma maculiforme appear to be adapted to low‐light environments such as caves or rock crevices (Wilce et al., 1970), where the conservation of DPOR function may be advantageous. In contrast, DPOR genes are frequently lost in many lineages adapted to high‐light environments, where light‐dependent protochlorophyllide oxidoreductase (LPOR) can compensate for their function (Kusumi et al., 2006; Uthanumallian et al., 2022). Inhabiting rocky surfaces in the intertidal zone, Ishige species are likely exposed to higher light intensities (Cho et al., 2005), suggesting that habitat‐specific light regimes may have driven increased nonsynonymous substitutions in DPOR genes. The duplicated cox2 genes identified in Pilinia rimosa provide a contrasting example of recent mitochondrial gene evolution driven by structural genome dynamics. Although all duplicated copies remained under overall purifying selection, both Ka/Ks (0.6–0.9) and RELAX analyses indicated a substantial relaxation of selective constraints following duplication. In contrast, the original copy retained a markedly lower Ka/Ks value (0.1) and did not show significant relaxation in RELAX analyses, suggesting that it continues to experience purifying selection. This pattern indicates that, although the duplicated copies remain under overall purifying selection, they are evolving under reduced selective constraints relative to the ancestral copy. Overall, these results suggest that the duplicated cox2 genes of Pi. rimosa represent an intermediate evolutionary state, in which the original copy maintains functional integrity while the duplicated copies experience relaxed selective pressures and ongoing divergence. CONCLUSIONS We reported the plastomes and mitogenomes for four species within the Ishigeales, revealing unique heterogeneity in organelle genome evolution across the order. Dynamic structural variations and repeat expansions were observed in the filamentous species Pilinia rimosa and Petroderma maculiforme, including extensive cox2 gene duplication in Pi. rimosa. Although intricate gene histories and incomplete genetic divergence were found in the filamentous species, the morphological changes observed in Ishige okamurae and I. foliacea may be linked to relaxed purifying selection in certain plastid genes, such as the DPOR gene family. Our findings indicate plastomes and mitogenomes in the Ishigeales have followed complex and largely independent evolutionary trajectories, which may be related to different morphologies and their environmental adaptations. To fully elucidate the genomic basis of morphological transitions in this order, nuclear genome analyses are essential in future studies. In particular, three independent origins of complex multicellularity evolution in brown algae should be comparatively investigated to clarify whether they were driven by shared genomic features or by distinct lineage‐specific traits. AUTHOR CONTRIBUTIONS Hayeon Hwang: Conceptualization (supporting); data curation (lead); formal analysis (lead); visualization (lead); writing – original draft (lead); writing – review and editing (lead). Hocheol Kim: Conceptualization (lead); data curation (supporting); formal analysis (supporting); supervision (lead); writing – original draft (lead); writing – review and editing (supporting). Seok‐Wan Choi: Methodology (supporting); writing – review and editing (supporting). Hiroshi Kawai: Resources (supporting); writing – review and editing (supporting). Hwan Su Yoon: Conceptualization (lead); funding acquisition (lead); project administration (supporting); supervision (lead); writing – review and editing (lead). FUNDING INFORMATION This study was supported by the National Research Foundation of Korea (RS‐2022‐NR068987, RS‐2022‐NR070837, and RS‐2025‐25434447), the Bio&Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. RS‐2024‐00411768), the Korea Institute of Marine Science and Technology Promotion (KIMST) funded to the Ministry of Oceans and Fisheries (MOF) of Korea (RS‐2021‐KS211476, RS‐2025‐02304428), the Cooperative Research Program for Agriculture Science and Technology Development of Korea (RS‐2023‐00231243), and Rural Development Administration, Republic of Korea. Supporting information Figure S1. Comparative gene composition of organelle genomes in brown algae and Schizocladia ischiensis. Heatmaps show the presence and absence of (a) Gene composition of plastomes and (b) Gene composition of mitogenomes in ten brown algal species and Schizocladia ischiensis. Blue indicates gene presence, whereas white indicates gene absence. Figure S2. Assembled plastomes (left) map and mitogenomes (right) map of four Ishigeales species. Figure S3. rbcL gene phylogenetic tree of the Ishigeales showing the unstable position of Diplura. Numbers on branches indicate bootstrap value (%) from maximum likelihood (ML) phylogenetic analysis. Figure S4. rbcL gene phylogenetic tree of the Sphacelariales. Numbers on branches indicate bootstrap value (%) from maximum likelihood (ML) phylogenetic analysis. Figure S5. The proportion of parsimony‐informative (PI) sites for each gene is plotted against total branch length (top) and root‐to‐tip distance (bottom). Each point represents a single gene. Figure S6. Plastome structural comparison among 41 brown algae species using Mauve alignment. Isogamous species show gene rearrangements even within the same order, whereas in oogamous lineages, rearrangements occur mainly between orders and plastid genomes are largely conserved within an order. Figure S7. Mitogenome structural comparison among 41 brown algae species using Mauve alignment. The mitogenome is generally conserved; however, gene rearrangements were detected in the Ishigeales and the Sphacelariales. Four instances of genome‐size expansion: in Pilinia rimosa through tandem repeats and amplification of the cox2 gene; in Petroderma maculiforme through tandem repeats; in Phaeostrophion irregulare through an increase in intergenic regions; and in Pylaiella littoralis through a group II intron. Figure S8. Distribution of tandem repeats in plastomes of the four Ishigeales species. Figure S9. Numbers and proportions of tandem repeats in plastomes and mitogenomes across brown algae. Blue bars indicate the number of tandem repeats detected in each species, whereas the orange line shows the proportion of tandem‐repeat sequence relative to (a) plastomes and (b) mitogenomes length of each species. Figure S10. Multiple sequence alignment of tandem repeat detected in Pilinia rimosa and Petroderma maculiforme. The tandem repeat sequences of Pi. rimosa and Pe. maculiforme show little to no alignment. Figure S11. Mitogenome compositions and tandem repeat in brown algal species exhibiting genome expansion. (a) Mitogenome compositions of the four Ishigeales species, the two Sphacelariales species and Pylaiella littoralis (Ectocarpales) and (b) distribution of tandem repeat sequences in the mitogenomes. Figure S12. Distribution and structural organization of the mitochondrial cox2 gene across brown algae. Figure S13. Pairwise Ka/Ks comparisons between the original cox2 locus and duplicated copies in Pilinia rimosa. Figure S14. Synonymous (Ks) and nonsynonymous (Ka) values by pairwise comparison between Ishigeales species. Table S1. Summary statistics of gene trees and alignment characteristics. Table S2. Pairwise Ka/Ks values for protein‐coding genes and relaxation of selective constraints of Pilinia rimosa. Table S3. Pairwise Ka/Ks values for protein‐coding genes in Ishigeales.