BMC Evolutionary Biology

Early high disparity, where clades reach maximum morphological diversity early in their history, is often observed in the fossil record. Since there are currently no suitable methods to model these Early High Disparity (EHD) processes accurately, we introduce a new phylogenetic comparative method (PCM) to address this gap by combining two evolutionary modes--Ornstein-Uhlenbeck (OU) and Early Burst (EB)--leading to an initial surge in disparity until an early maximum, driven by the EB component, after which disparity declines due to OU constraints. We tested the EHD model on ichthyosaur skull length data and found it to be the best fit in nearly all cases compared to other modes. The median time of peak disparity (TEHD) is about 231.1 million years ago, during the Triassic, roughly 12.4% into the ichthyosaur lineage. To explore how environmental and geological factors might influence peak disparity, we performed changepoint analyses on sea level, sea surface temperature, rock area, and sedimentary sequence ages. While changepoints in temperature at 237.2 Ma and sequence boundaries at 238 Ma were closest to TEHD, no significant changepoints appeared to match TEHD.

Recent studies have shown that the Triassic stem-frog Triadobatrachus lacked the ability to jump off, but nonetheless had the forelimb strength to withstand the impact of landing from a jump. We propose a hypothesis to resolve this pseudoparadox: the strengthened forelimbs are former adaptations to forelimb-based digging that later made jumping possible by exaptation. Micro-CT data from a skeleton of Batropetes palatinus reveal thin cortical bone, confirming Batropetes as terrestrial. Combining adaptations to walking and digging, confirmed by statistical analyses, Batropetes is thought to have searched for food in leaf litter or topsoil. We interpret Batropetes as having used one forelimb at a time to shove leaf litter aside. Batropetes may thus represent an analog or possibly a homolog of the digging stage that preceded the origin of Salientia. We discuss the possibility of homology with the digging lifestyles of other "microsaurs" and other amphibians.

The evolution of genome size and eukaryotic genome organisation has puzzled evolutionary biologists for decades. Are differences in DNA content and protein-coding capacity of eukaryotic genomes a result of natural selection, or neutral processes? If they were shaped by natural selection, which features might be connected with these differences: cell volume, ecological strategy, availability of nutrients, etc.? Kinetoplastids are protists with compact nuclear genomes that almost entirely lack introns, but whose mitochondrial DNA content (kDNA) can rival that of their nuclei. We hypothesised that the reduction in non-protein coding nuclear DNA content during the evolution of kinetoplastids is connected with an increase in kDNA content: species with larger fractions of total cellular DNA that is allocated to kDNA will have concomitantly smaller fractions of non-coding DNA in the nuclear genome. We developed a framework for estimation of DNA content in both the mitochondria and nuclei from quantitative fluorescence microscopy data, and inferred the fraction of non-coding DNA in nuclear genomes based on the combination of transcriptomic and fluorescence imaging data. Across 10 genetically distant isolates from diverse free-living kinetoplastid lineages, there is indeed a trend towards a lower proportion of non-protein-coding nuclear DNA when kDNA proportion is higher relative to the total DNA in the cell. This observation is consistent with the hypothesis that in free-living kinetoplastids, mitochondrial and nuclear genomes might co-evolve in terms of their DNA content. This finding raises the possibility that natural selection acts on genome size evolution in kinetoplastids, and contributes evidence and methodology to address this question more broadly in eukaryotes.

The common marmoset (Callithrix jacchus) is a biomedically important species that is characterized by two unusual biological traits -- a high frequency of twin births and hematopoietic chimerism -- that preclude the application of many commonly used population genomic approaches for quantifying evolutionary processes. In this study, we directly account for both factors in order to estimate fine-scale mutation and recombination rate maps, as well as to infer the demographic and selective processes shaping variation, on the common marmoset X chromosome. Comparing our findings to estimates recently inferred on the autosomes of this species, we find reduced rates of mutation and recombination on the X, as expected. Furthermore, population sex ratios are inferred to be nearly equal, and the appropriately rescaled autosomal population history fits the X chromosome well. Finally, we report evidence of recent selective sweeps targeting a number of X-linked genes, including several of significant biomedical relevance. Overall, these analyses provide novel insights into the evolutionary processes shaping X chromosome evolution in this biomedically-relevant primate model.

The timing of the origins of the crown-group placental mammals has provided one of the classic battlefields in the long-running debate over when clades arise. Undoubted fossil crown-group placentals appear only in the Paleogene, but even so most molecular analyses, and many palaeontologists, have suggested their true origin is somewhere between 70-100 Ma. However, apart from the fact of the fossil record itself, there are several reasons to believe that the true origin is indeed post-Cretaceous, including consideration of the dynamics of stem and crown groups, which strongly favour crown-group origins to lie just after, and not just before, mass extinctions. Here we consider this "hard explosive" model in the light of the newly-developed "Covariant Evolutionary Tempo (CET)" model which allows diversification and molecular evolution rates to covary. It predicts "early bursts" in both lineage creation and molecular evolution at the base of major radiations which lead to highly unequally-sized clades; and an inheritance of rapid rates from this initial event by extant rapidly-evolving clades. We show that when the placentals are constrained to emerge after the K-Pg boundary, they indeed show elevated rates of both diversification and molecular evolution, which rapidly decline. Nevertheless, although elevated, these rates are comparable to the fastest rates seen in extant clades such as the rodents. In addition, the contiguous lineages leading from the origin to the rodents and other fast evolving clades also show elevated rates. These patterns suggest that not only is a Paleogene origin for the placental crown-group plausible, as fossil evidence suggests, but they also provide support for the CET model, which should be considered in other cases of pronounced fossil record/molecular clock mismatch.

AbstractFossils and ancient crude pottery vessels function as physical "DNA containers," preserving oriDNA (original, in situ DNA) and accumulating eDNA (environmental DNA) over time. Ancient DNA (aDNA) serves as molecular fossils, chronicling evolutionaryhistory. Using "nano-affinitybead technology", we extracted and sequenced DNAfrom Lycoptera fossils in the Jehol region and a "round-bellied jar (Jar)" from the Erlitou period in Guangwu Town, yielding 236,545 primate sequences from the fossils and 86,908 from the pottery. We observed that the AFF value of DNA sequences negatively corre lates with species divergence time1, offering a quantitative measure of DNA preservation and host divergence. Some fragments distinct from modern genomic sequences, termed "intermediate DNAsequences" (IDS), have been identified. Many IDS exhibit an upper age limit, preserving characteristics of the last common ancestor (LCA) of the Hominidae and offering molecular insights into "Darwins puzzle". Among IDS, the SRRA subtype (Sequence Reversal and Rearrangement), identified in mRNA-coding exonic sequences, arises from the incorporation of a complementary antisense strand upstream of the sense strand, either adjacent to it or separated bya sequence interval. This introduces a novel post-transcriptional regulatory mechanism at the mRNA level, driven by SRRA, which presets hairpin structures and Indels in the UTR or CDS of exonic sequences, modulating gene variation. We propose: "SRRAs played a critical trial-and-error role in early Hominidae evolution, facilitating adaptive genomic changes, with some SRRA sequences later excised from exons", positioning SRRAas an evolutionary "genetic switch". Additionally, seven species -specific fragments (SSFs) of non-human primates (NHPs) linked to Asian Homo erectus were identified in the fossils, and pottery DNA reveale d sequences from tropical species (e.g., zebrass, oil palms), providing evidence of climate-driven local extinction and supporting paleo-ecological and paleo-environmental reconstruction. This method of analyzing aDNA from non-skeletal materials opens new avenues in paleontology, archaeology, and geology, guiding the tracing of ancient migration patterns and fossil searches. DNAfragments preserved within "DNA containers " exhibit an "old-few, new-many" turnover pattern, with many aDNA fragments displaying non-deamination. This evidence challenges prevailing perspectives, the authenticity criteria for aDNA, and the capabilities and scope of the traditional research method, necessitating a thorough reevaluation of the relevant knowledge framework. Furthermore, this study opens new opportunities for frontier research.

The allorecognition system of the hydroid Hydractinia symbiolongicarpus plays an essential role in mediating self vs. non-self recognition among cnidarian colonies, analogous to the role of the major histocompatibility complex (MHC) in bilaterian immune systems. Both complexes have contiguous stretches of similar, highly duplicated genes, suggesting a homologous relationship. Protein structural prediction methods such as AlphaFold2 also indicate structural similarity between the proteins encoded within the Hydractinia ARC and the MHC. We have determined the structure of the ARC in the related hydrozoan Podocoryna americana to determine the degree of conservation of the cnidarian allorecognition complex. Protein structural prediction methods have revealed a number of candidate allorecognition proteins that form a long contiguous complex. Phylogenies suggest homology between cnidarian allorecognition genes, as well as syntenic conservation of this gene complex across hydrozoans. Ongoing comparative genomic analyses across bilaterian species aim to clarify the relationship between the ARC and the MHC, thereby informing our understanding of the evolutionary origins of bilaterian adaptive immunity.

Proper mitochondrial function is reliant on favourable mitonuclear epistatic interactions between two genomes, the mitochondrial (mtDNA) and the nuclear with contrasting rates of mutation, copy-number and modes of replication and transmission. The compactness and limited coding capacity of animal mtDNA along with their high mutation rates, uniparental inheritance and lack of recombination has contributed to the view that mtDNA is prone to accumulating deleterious mutations, with mitonuclear adaptation primarily attributed to the nuclear genome. Employing a prospective evolution framework in Caenorhabditis elegans, we investigated the adaptive capacity of the mtDNA in ETC-deficient lines bearing four focal ETC mutants across three different breeding systems following 60 generations of adaptation. The mtDNA genome can adapt rapidly to mitochondrial dysfunction caused by deleterious mutations in nuclear genes. Nonsynonymous changes in mtDNA protein-coding genes had significantly higher frequencies and were positively correlated with fitness, indicative of positive selection. Mitochondrial gene nd-1 emerged as a key driver of compensatory mitonuclear adaptation. Notably, compensatory mutations reach higher frequency when they affected ETC subunits in the same complex as the original deleterious mutation. Our results argue for a disproportionately outsized role for mitochondria in mitonuclear adaptation despite their diminutive genome size and protein-coding capacity.

Signals of positive selection around the SLC30A9 gene have been reported in human populations outside Africa. Selection likely acted on a highly differentiated single-nucleotide polymorphism, rs1047626, leading to a non-synonymous substitution in the encoded zinc transporter. Because of the striking similarity between the putatively selected SLC30A9 haplotype observed in several current human populations and the Denisovan individual, previous work has proposed adaptive introgression. Yet alternative explanations, including ancient human variation, and the precise archaic source -Neanderthal or Denisovan- remained unresolved. Considering the potentially complex evolution of SLC30A9, we applied Approximate Bayesian Computation (ABC) algorithms coupled to machine learning to investigate the most plausible evolutionary origin of this substitution. After modelling different evolutionary scenarios with forward-in-time simulations, our results highlight that the most probable scenario is a Denisovan origin of the rs1047626 polymorphism. However, the allele likely introgressed into Neanderthals first and was then passed into non-African modern humans. Moreover, the derived allele frequency for rs1047626 across several African populations is consistent with back-to-Africa migrations. Finally, our ABC analyses indicate strong positive selection in East Asian populations and other out-of-Africa populations, whereas in Neanderthal populations, the selection coefficient was probably neutral or slightly deleterious.

Lineages that have undergone little apparent evolutionary change over long geological intervals are often described as "living fossils", with iconic examples including the coelacanth, tuatara, and Ginkgo. However, studies of morphological and molecular data have produced mixed evidence regarding whether living fossils have evolved more slowly than their relatives. We tested for reduced evolutionary rates using a sister-pairs approach, analysing a comprehensive set of morphological and molecular data for 45 living-fossil lineages and their close relatives across the Tree of Life. Living-fossil lineages showed substantially reduced rates of morphological evolution, whereas evidence for reduced molecular rates was weaker. To test whether this pattern could be explained by species poverty, we also analysed other depauperate lineages (species-poor clades contrasted with species-rich sister clades) and found a weak tendency towards reduced molecular rates but without a corresponding slowdown in morphological evolution. Our study reveals that the living-fossil concept captures a quantifiable pattern of prolonged morphological conservatism, but also suggests that living fossils lack other shared biological or life-history characteristics that might unite them as a group.

The relative roles of genetic drift, inbreeding, and natural selection in island evolution have been fundamental themes in evolutionary biology for nearly a century. However, empirical studies have yet to reach a consensus, and research integrating both genotypes and phenotypes in primates remains limited. The present study aims to estimate the demographic history of an island primate population, the Yakushima macaque (Macaca fuscata yakui), and evaluate the genomic and morphological consequences of its isolation. To this end, we conducted population genetic analyses based on whole-genome sequencing and geometric morphometric analyses of cranial shape. Our results indicate that M. f. yakui has been isolated with an effective population size of several thousand individuals for over 100,000 years. Although we detected signatures of purifying selection, deleterious alleles have not been sufficiently purged. The degree of fluctuating asymmetry in males was slightly, yet significantly, higher in M. f. yakui than in mainland populations, whereas this trend was not observed in females. Marginal evidence of directional selection was found for cranial size, but not for shape. Overall, our findings suggest that due to long-term isolation, genetic drift and inbreeding have exerted a substantial influence on shaping genotypic and phenotypic variation in this island population.

Gene duplication is an important evolutionary mechanism contributing to genomic innovation and adaptation. While the long term fate of duplicated genes has been extensively studied, immediate effects such as those resulting from increase in gene dosage remain understudied. Of particular interest is the question of how expression level relates to the optimum after a duplication, as it could fall short, exceed or directly reach the optimal expression. In this study, we investigated the relationship between expression and fitness for 11 genes previously reported to increase fitness when duplicated in Saccharomyces cerevisiae to compare the expression level achieved by duplication to the one maximizing fitness. While most of the studied genes show little fitness variation across the expression levels tested, we found that ESP1, which encodes a separase essential to anaphase initiation, exhibits a clear fitness optimum at an expression level above the native single-copy state. However, the expression level associated with gene duplication exceeds the optimum. We find that ESP1 overexpression is associated with a reduced cell size and a lower proportion of the cell cycle spent in G2 phase, consistent with its role in cell cycle progression. We propose that the duplication of ESP1 accelerates the cell cycle, resulting in smaller, faster-dividing cells. When measured in different environments, the fitness functions also exhibit similar shapes, with slight variation in the optimal expression. Our findings highlight how gene duplication can confer an immediate selective advantage through dosage increase while being non-optimal, reversing the directionality of natural selection on expression level after duplication.

Random genetic drift, rather than selection, is expected to determine the dynamics of alleles with fitness effects that fall below the reciprocal of effective population size (Ne). This is the drift barrier. The drift barrier hypothesis suggests that adaptation may be hopelessly constrained in organisms with low Ne, with much of the complexity of life reflecting a byproduct of ineffective natural selection. Yet decades of empirical research in evolutionary ecology provides substantial evidence of adaptation in finite metazoan populations. Here I explore the extent to which the drift barrier may, or may not, constrain contemporary adaptive divergence under directional selection in the wild. Compensatory directional selection on polygenic traits occurring after the fixation of a deleterious allele will be twice as efficient compared to the initial action of stabilizing selection. Consistent with this result, meta analysis reveals a negative relationship between contemporary estimates of Ne and the strength of directional selection in natural populations. Analysis of empirical phenotypic selection gradients from wild metazoans further reveals that directional selection is often strong enough to overcome the constraints imposed by finite population size, across a range of genetic architectures. Moreover, on a multi-peak adaptive landscape, the probability of a peak shift between adaptive zones increases exponentially as Ne declines, with observed values of contemporary Ne typically being well within the range needed for peak shift models to be a viable explanation of diversity. Together, these results suggest that widespread adaptation in metazoans is not inconsistent with the fundamental constraints on perfection imposed by drift.

Aquaporin-10 (AQP10) is an aquaglyceroporin that transports small, uncharged molecules, such as water and glycerol. In some species, such as mice and cattle, the AQP10 gene (Aqp10) is pseudogenized. In rodents, pseudogenization of Aqp10 is limited to the Myomorpha suborder, while Aqp10 remains intact in other rodent lineages. The order Cetartiodactyla encompasses diverse species, including Suina, Tylopoda, Ruminantia, and Whippomorpha. The distribution of intact Aqp10 across the entire Cetartiodactyla and analyses estimating the timing of Aqp10 pseudogenization have not been sufficiently investigated. In this study, we identified species within Cetartiodactyla that retain intact Aqp10 and analyzed the detailed patterns of pseudogenization to elucidate the specifics of pseudogenization within this order. We collected Aqp10 loci by BLAST and synteny analyses of publicly available genome databases of 51 cetartiodactyl species from 17 families. Dot plot and alignment analyses revealed that seudogenization of Aqp10 occurred in 42 Cetartiodactyla species, including the Camelidae, warthog (Suidae), Antilocapridae, okapi (Giraffidae), Cervidae, Bovidae, Hippopotamidae, and Cetacea. In contrast, nine species retained an intact Aqp10: pig and babirusa (Suidae), peccary (Tayassuidae), mouse deeres (Tragulidae), giraffes (Giraffidae), and musk deers (Moschidae). The timing of pseudogenization was estimated by comparing patterns of exon deletions, frameshifts, and nonsense mutations in Aqp10 pseudogenes. RT-PCR analysis of cattle and sheep tissues revealed that these species express Aqp3 and Aqp7 in their intestines, but not Aqp10. Collectively, these findings indicate that Aqp10 pseudogenization occurred independently in multiple cetartiodactyl lineages and provide new insights into the evolutionary history of Aqp10 loss within this order.

Recent findings in humans and other species have revealed the presence of "mutator" alleles that increase germline mutation rate across the genome. Such mutators are expected to be selected against because of the additional deleterious alleles that they generate, to a degree that will depend on how much they increase the mutation rate in heterozygotes and homozygotes. To describe their dynamics, we develop a population genetic model of mutation rate modifiers with arbitrary dominance coefficients, in which fitness effects stem from additional germline mutations. We then use it to interpret findings for the seven human mutators identified to date. For six of the seven known mutators, the observed frequencies are well fit by the model and thus consistent with purifying selection arising solely due to their effects on germline mutation rates, although also consistent with a wide range of parameters; in two, the observed frequencies are readily explained by purely recessive fitness effects. The exception is a variant in MUTYH, which is more common than predicted under plausible parameters, for reasons that remain unclear. We also use the model to explore what types of mutators are most likely to be discovered in parents, based on identifying offspring with unexpectedly high numbers of de novo mutations. Although the two first mutators identified by this approach seem to be recessive, our modeling suggests that, for the same effect size, semi-dominant mutators are much more likely to be detected. These findings therefore imply that there are many more modifier sites with recessive effects than semi-dominant ones. More generally, our model provides a framework for interpreting properties of mutators in humans and other species and for learning about the genetic architecture of germline mutation rate variation.

Many pathogenic and microbial species undergo complex life cycles wherein they experience drastic bottlenecks during invasion into a host and rapid growth within hosts. The effects of such a complex population history on patterns of genetic variation is not currently understood. We model a pathogen population as a metapopulation where each deme represents a host and incorporate transmission dynamics as well as the effects of transmission bottlenecks. We employ a coalescent framework, obtaining analytical expressions for pairwise times to coalescence and genetic differentiation within and between hosts. We find that recurrent bottlenecks rescale the coalescent process within hosts to the size of the bottleneck, reducing within-host variation. In addition, the pairwise distribution of time to coalescence under our model deviates substantially from the Wright-Fisher process when the number of hosts (d) is of similar order to the size of the bottleneck (k), i.e., k [~] d and when m <<1/(2k) +1/(2d), where m is the migration rate. Otherwise, the pairwise process under our model converges in distribution to the Wright-Fisher process. Coalescent simulations show that under the same conditions the site frequency spectrum (SFS) under our model deviates from that under the Kingman coalescent and is drastically skewed towards singletons. Our results suggest that the inference of selection or demographic history from pathogen population genetics data may be systematically biased by the mechanics of pathogen reproduction, indicating a need for further theoretical work that can inform inference.

The classical model of chromosomal speciation is based on the observation that divergent species often differ in chromosome numbers and that heterokaryotypic hybrids show reduced fitness. Furthermore, positive correlations between the rates of karyotype evolution and speciation have been established in several groups of organisms. Yet, the mechanisms that make some lineages more prone to chromosomal fissions and fusions than others, and the causal role of such karyotype changes in speciation, are poorly understood. Here, we focus on the taxonomically and karyotypically diverse group of African monkeys, the guenons (tribe Cercopithecini), consisting of ~35 species with diploid chromosome numbers ranging from 48 to 72. We generated de novo genome assemblies for eight species with different karyotypes, and incorporated publicly available genomes from additionally three guenon species. Using the rhesus macaque as an outgroup, we identified 28 independent chromosomal fissions and 4 fusions in guenons. In line with previous cytogenetic studies, repeat composition around chromosomal breakpoints suggested that most fissions were non-centromeric, requiring the evolution of a new centromere in at least one of the fission products. In contrast to the expectation under strong underdominance, we found a paraphyletic distribution of several chromosomal fissions. Some of these phylogenetically discordant events could have resulted from interspecific hybridization, whereas others are more likely caused by incomplete lineage sorting, suggesting that fissioned and non-fissioned homologs segregated in ancestral guenon lineages for millions of years. Nevertheless, we find evidence for reduced introgression on rearranged chromosomes in an ancient cross-karyotypic gene flow event between Cercopithecus cephus and C. pogonias. Although this suggests that chromosomal fissions may have constituted barriers to gene flow in the guenon evolution, the signal was inconclusive or absent in two other investigated gene flow events. This study provides important first insights into the remarkable karyotypic diversification of guenons, and its role in generating reproductive barriers.

Mammalian brains exhibit extensive diversity in size, cellular composition, and organization, yet the molecular evolutionary changes associated with this phenotypic diversification remain incompletely understood. Here, we investigated whether protein-sequence evolution covaries with quantitative differences in brain phenotypes across mammals. We integrated 29 brain and body traits with orthologous protein sequences from 18 mammalian species and obtained gene-specific molecular evolutionary measures for 10,395 genes. To reduce redundancy among strongly correlated phenotypes, we selected five representative traits that explained 88.12% of the standardized trait variance under linear reconstruction. Using RERconverge, we tested for associations between gene-specific relative evolutionary rates and evolutionary changes in these five representative traits. The primary analysis, based on log-transformed phenotypes, identified 25 significant gene--trait associations involving 23 genes, 19 of which were also recovered using untransformed phenotypes, indicating that a substantial subset of the associations was robust to phenotype transformation. Expression profiling of these shared candidate genes across adult human GTEx brain tissues revealed heterogeneous patterns ranging from broad expression across the examined tissues to low adult brain expression. Functional enrichment analysis further showed that the gene set retained for comparative analysis was enriched for diverse biological processes, including metabolic, cellular, and neural pathways. These results identify gene-specific evolutionary-rate associations with quantitative mammalian phenotypes and provide functional context for the genes represented in the comparative analysis. Our study provides a comparative framework linking relative evolutionary rates of protein-coding genes to quantitative variation in mammalian brain cellular architecture while accounting for shared evolutionary history.

Recombination shapes allelic combinations across the genome and plays a central role in adaptation and speciation. While recombination rates are known to vary and evolve, how recombination landscapes diverge between populations and which processes drive this divergence remains poorly understood. Here, we use the Population Recombination Divergence Index (PRDI) to quantify divergence in effective recombination landscapes between seven pairs of populations and species from four percomorph fish lineages (cichlids, sticklebacks, killifish, and clownfish). While accounting for stochasticity and confounders of linkage disequilibrium (LD) -based recombination rate estimates, we find that recombination landscapes diverged substantially in all comparisons. This divergence generally increases with genetic differentiation and divergence time, but the amount of PRDI accumulated per unit of genetic divergence (FST), net nucleotide divergence (da), or divergence time varies markedly among population and species pairs. Comparisons across systems suggest that PRDI accumulation tends to be faster (per unit of FST, da, or divergence time) in pairs with higher inferred gene flow. We also introduce a local version of PRDI (_l) to better characterize genomic heterogeneity in effective recombination divergence. We find that higher inferred gene flow between populations tends to be associated with stronger intragenomic correlation between local FST and _l. Finally, we propose a conceptual framework for investigating how divergent selection and gene flow may interact to shape effective recombination landscape divergence. Our work highlights the heterogeneous dynamics of effective recombination landscape evolution, and motivates further theoretical and empirical efforts to integrate recombination landscape evolution into studies of adaptation and speciation.

Standard methods for ancestral sequence reconstruction (ASR) rely on substitution models for the residues in a biological sequence and assume independent evolution across these sites, ignoring the epistatic interactions that shape molecular evolution. In contrast, deep learning models like variational autoencoders (VAEs) can learn low-dimensional representations ("embeddings") of sequences in a protein family that may implicitly handle these dependencies, raising the possibility of performing more accurate ASR by interpolating between extant sequence embeddings within the VAE's latent space. In this study, we test this hypothesis by developing and evaluating a VAE-based ASR pipeline. Benchmarking this approach against established likelihood-based and parsimony methods using various simulations of protein evolution, including scenarios with and without epistasis, we find that the VAE-based approach is consistently and significantly outperformed by standard methods, even in epistatic regimes where it was hypothesized to have an advantage. We further show that this failure is not due to a lack of phylogenetic structure in the latent space, which does contain evolutionary signal. Rather, the primary limitation is the information loss inherent to the autoencoding process: the VAE's decoder cannot generate sequences with sufficient fidelity for the precise demands of ASR.