Colin Kern

Three crocodilian genomes reveal ancestral patterns of evolution among archosaurs

INTRODUCTION Crocodilians and birds are the two extant clades of archosaurs, a group that includes the extinct dinosaurs and pterosaurs. Fossils suggest that living crocodilians (alligators, crocodiles, and gharials) have a most recent common ancestor 80 to 100 million years ago. Extant crocodilians are notable for their distinct morphology, limited intraspecific variation, and slow karyotype evolution. Despite their unique biology and phylogenetic position, little is known about genome evolution within crocodilians. Evolutionary rates of tetrapods inferred from DNA sequences anchored by ultraconserved elements. Evolutionary rates among reptiles vary, with especially low rates among extant crocodilians but high rates among squamates. We have reconstructed the genomes of the common ancestor of birds and of all archosaurs (shown in gray silhouette, although the morphology of these species is uncertain). RATIONALE Genome sequences for the American alligator, saltwater crocodile, and Indian gharial—representatives of all three extant crocodilian families—were obtained to facilitate better understanding of the unique biology of this group and provide a context for studying avian genome evolution. Sequence data from these three crocodilians and birds also allow reconstruction of the ancestral archosaurian genome. RESULTS We sequenced shotgun genomic libraries from each species and used a variety of assembly strategies to obtain draft genomes for these three crocodilians. The assembled scaffold N50 was highest for the alligator (508 kilobases). Using a panel of reptile genome sequences, we generated phylogenies that confirm the sister relationship between crocodiles and gharials, the relationship with birds as members of extant Archosauria, and the outgroup status of turtles relative to birds and crocodilians. We also estimated evolutionary rates along branches of the tetrapod phylogeny using two approaches: ultraconserved element–anchored sequences and fourfold degenerate sites within stringently filtered orthologous gene alignments. Both analyses indicate that the rates of base substitution along the crocodilian and turtle lineages are extremely low. Supporting observations were made for transposable element content and for gene family evolution. Analysis of whole-genome alignments across a panel of reptiles and mammals showed that the rate of accumulation of micro-insertions and microdeletions is proportionally lower in crocodilians, consistent with a single underlying cause of a reduced rate of evolutionary change rather than intrinsic differences in base repair machinery. We hypothesize that this single cause may be a consistently longer generation time over the evolutionary history of Crocodylia. Low heterozygosity was observed in each genome, consistent with previous analyses, including the Chinese alligator. Pairwise sequential Markov chain analysis of regional heterozygosity indicates that during glacial cycles of the Pleistocene, each species suffered reductions in effective population size. The reduction was especially strong for the American alligator, whose current range extends farthest into regions of temperate climates. CONCLUSION We used crocodilian, avian, and outgroup genomes to reconstruct 584 megabases of the archosaurian common ancestor genome and the genomes of key ancestral nodes. The estimated accuracy of the archosaurian genome reconstruction is 91% and is higher for conserved regions such as genes. The reconstructed genome can be improved by adding more crocodilian and avian genome assemblies and may provide a unique window to the genomes of extinct organisms such as dinosaurs and pterosaurs. To provide context for the diversification of archosaurs—the group that includes crocodilians, dinosaurs, and birds—we generated draft genomes of three crocodilians: Alligator mississippiensis (the American alligator), Crocodylus porosus (the saltwater crocodile), and Gavialis gangeticus (the Indian gharial). We observed an exceptionally slow rate of genome evolution within crocodilians at all levels, including nucleotide substitutions, indels, transposable element content and movement, gene family evolution, and chromosomal synteny. When placed within the context of related taxa including birds and turtles, this suggests that the common ancestor of all of these taxa also exhibited slow genome evolution and that the comparatively rapid evolution is derived in birds. The data also provided the opportunity to analyze heterozygosity in crocodilians, which indicates a likely reduction in population size for all three taxa through the Pleistocene. Finally, these data combined with newly published bird genomes allowed us to reconstruct the partial genome of the common ancestor of archosaurs, thereby providing a tool to investigate the genetic starting material of crocodilians, birds, and dinosaurs.

Large numbers of novel miRNAs originate from DNA transposons and are coincident with a large species radiation in bats

Vesper bats (family Vespertilionidae) experienced a rapid adaptive radiation beginning around 36 Ma that resulted in the second most species-rich mammalian family (>400 species). Coincident with that radiation was an initial burst of DNA transposon activity that has continued into the present in some species. Such extensive and recent DNA transposon activity has not been seen in any other extant mammal. Indeed, retrotransposon activity is much more common in all other sequenced mammal genomes. Deep sequencing of the small RNA fraction from a vespertilionid bat, Eptesicus fuscus, as well as a dog and horse revealed large numbers of 17-24 bp putative miRNAs (p/miRNAs). Although the origination rate of p/miRNAs is similar in all three taxa, 61.1% of postdivergence p/miRNAs in Eptesicus are derived from transposable elements (TEs) compared with only 23.9% and 16.5% in the dog and horse, respectively. Not surprisingly, given the retrotransposon bias of dog and horse, the majority of TE-derived p/miRNAs are associated with retrotransposons. In Eptesicus, however, 58.7% of the TE-derived and 35.9% of the total p/miRNAs arose not from retrotransposons but from bat-specific DNA transposons. Notably, we observe that the timing of the DNA transposon expansion and the resulting introduction of novel p/miRNAs coincide with the rapid diversification of the family Vespertilionidae. Furthermore, potential targets of the DNA transposon-derived p/miRNAs are identifiable and enriched for genes that are important for regulation of transcription. We propose that lineage-specific DNA transposon activity lead to the rapid and repeated introduction of novel p/miRNAs. Some of these p/miRNAs are likely functional miRNAs and potentially influenced the diversification of Vespertilionidae. Our observations suggest a mechanism for introducing functional genomic variation rapidly through the expansion of DNA transposons that fits within the TE-thrust hypothesis.

Predicting Interacting Residues Using Long-Distance Information and Novel Decoding in Hidden Markov Models

Identification of interacting residues involved in protein-protein and protein-ligand interaction is critical for the prediction and understanding of the interaction and has practical impact on mutagenesis and drug design. In this work, we introduce a new decoding algorithm, ETB-Viterbi, with an early traceback mechanism, and apply it to interaction profile hidden Markov models (ipHMMs) to enable optimized incorporation of long-distance correlations between interacting residues, leading to improved prediction accuracy. The method was applied and tested to a set of domain-domain interaction families from the 3DID database, and showed statistically significant improvement in accuracy measured by F-score. To gauge and assess the methods effectiveness and robustness in capturing the correlation signals, sets of simulated data based on the 3DID dataset with controllable correlation between interacting residues were also used, as well as reversed sequence orientation. It was demonstrated that the prediction consistently improves as the correlations increase and is not significantly affected by sequence orientation.

Improved genome assembly of American alligator genome reveals conserved architecture of estrogen signaling

The American alligator, Alligator mississippiensis, like all crocodilians, has temperature-dependent sex determination, in which the sex of an embryo is determined by the incubation temperature of the egg during a critical period of development. The lack of genetic differences between male and female alligators leaves open the question of how the genes responsible for sex determination and differentiation are regulated. Insight into this question comes from the fact that exposing an embryo incubated at male-producing temperature to estrogen causes it to develop ovaries. Because estrogen response elements are known to regulate genes over long distances, a contiguous genome assembly is crucial for predicting and understanding their impact. We present an improved assembly of the American alligator genome, scaffolded with in vitro proximity ligation (Chicago) data. We use this assembly to scaffold two other crocodilian genomes based on synteny. We perform RNA sequencing of tissues from American alligator embryos to find genes that are differentially expressed between embryos incubated at male- versus female-producing temperature. Finally, we use the improved contiguity of our assembly along with the current model of CTCF-mediated chromatin looping to predict regions of the genome likely to contain estrogen-responsive genes. We find that these regions are significantly enriched for genes with female-biased expression in developing gonads after the critical period during which sex is determined by incubation temperature. We thus conclude that estrogen signaling is a major driver of female-biased gene expression in the post-temperature sensitive period gonads.

Evolution of avian encephalomyelitis virus during embryo-adaptation

Wild-type avian encephalomyelitis virus (AEV) causes neurological signs in young chicks but no disease in pullets after oral or intracutaneous infection. However, if the virus gets embryo-adapted by serial passaging in chicken embryos, it will cause AE after intracutaneous infection in chickens of all ages. Recently, several cases of AE in layer pullets occurring shortly after intracutaneous vaccination were described. The present investigation was initiated to determine if vaccines that had inadvertently been embryo-adapted were responsible for these outbreaks. Virus isolation was done from two vaccines and one field sample. One of the vaccines had been used in one of the flocks before the outbreak. After the first passage, regardless of the inoculum, no embryo was paralyzed, indicating that the vaccines and the field isolate were not embryo-adapted. After seven passages all three strains were fully embryo-adapted causing typical lesions in the embryos. Viral load as determined by RT-qPCR remained constant during the passages. Partial sequences of the VP2 gene of vaccines, the field sample and four other field isolates were nearly identical and highly similar to published sequences from all over the world; only sequences originating from non-vaccinated birds were clearly set apart. Analysis of whole genomes identified two single nucleotide polymorphisms (SNPs) that distinguished wild-type and embryo-adapted strains. Sanger sequencing brains and nerves of the five field isolates and of the first, third and fifth passages of the isolates showed that the mutations indicating embryo-adaptation were first observed in the fifth passage.

Genome-wide identification of tissue-specific long non-coding RNA in three farm animal species

BackgroundNumerous long non-coding RNAs (lncRNAs) have been identified and their roles in gene regulation in humans, mice, and other model organisms studied; however, far less research has been focused on lncRNAs in farm animal species. While previous studies in chickens, cattle, and pigs identified lncRNAs in specific developmental stages or differentially expressed under specific conditions in a limited number of tissues, more comprehensive identification of lncRNAs in these species is needed. The goal of the FAANG Consortium (Functional Annotation of Animal Genomes) is to functionally annotate animal genomes, including the annotation of lncRNAs. As one of the FAANG pilot projects, lncRNAs were identified across eight tissues in two adult male biological replicates from chickens, cattle, and pigs.ResultsComprehensive lncRNA annotations for the chicken, cattle, and pig genomes were generated by utilizing RNA-seq from eight tissue types from two biological replicates per species at the adult developmental stage. A total of 9393 lncRNAs in chickens, 7235 lncRNAs in cattle, and 14,429 lncRNAs in pigs were identified. Including novel isoforms and lncRNAs from novel loci, 5288 novel lncRNAs were identified in chickens, 3732 in cattle, and 4870 in pigs. These transcripts match previously known patterns of lncRNAs, such as generally lower expression levels than mRNAs and higher tissue specificity. An analysis of lncRNA conservation across species identified a set of conserved lncRNAs with potential functions associated with chromatin structure and gene regulation. Tissue-specific lncRNAs were identified. Genes proximal to tissue-specific lncRNAs were enriched for GO terms associated with the tissue of origin, such as leukocyte activation in spleen.ConclusionsLncRNAs were identified in three important farm animal species using eight tissues from adult individuals. About half of the identified lncRNAs were not previously reported in the NCBI annotations for these species. While lncRNAs are less conserved than protein-coding genes, a set of positionally conserved lncRNAs were identified among chickens, cattle, and pigs with potential functions related to chromatin structure and gene regulation. Tissue-specific lncRNAs have potential regulatory functions on genes enriched for tissue-specific GO terms. Future work will include epigenetic data from ChIP-seq experiments to further refine these annotations.

A post-decoding re-ranking algorithm for predicting interacting residues in proteins with hidden Markov models incorporating long-distance information

Protein-protein interactions play a central role in the biological processes of cells. Accurate prediction of the interacting residues in protein-protein interactions enhances understanding of the interaction mechanisms and enables in silico mutagenesis, which can help facilitate drug design and deepen our understanding of the inner workings of cells. Correlations have been found among interacting residues as a result of selection pressure to retain the interaction during evolution. In previous work, incorporation of such correlations in the interaction profile hidden Markov models with a special decoding algorithm (ETB-Viterbi) has led to improvement in prediction accuracy. In this work, we first demonstrated the sub-optimality of the ETB-Viterbi algorithm, and then reformulated the optimality of decoding paths to include correlations between interacting residues. To identify optimal decoding paths, we propose a post-decoding re-ranking algorithm based on a genetic algorithm with simulated annealing and show that the new method gains an increase of near 14% in prediction accuracy over the ETB-Viterbi algorithm.

Improving interacting residue prediction using long-distance information in hidden Markov models

Identification of interacting residues involved in protein-protein and protein-ligand interaction is critical for the prediction and understanding of the interaction and has practical impact on mutagenesis and drug design. In this work, we introduce a new decoding algorithm, ETB-Viterbi, with early trace back mechanism built into interaction profile hidden Markov models (ipHMMs) that can incorporate the long-distance correlations between interacting residues to improve prediction accuracy. The method was applied and tested to a set of domain-domain interaction families from the 3DID database, and showed statistically significant improvement in accuracy measured by F-score. To gauge and assess the methods effectiveness in capturing the correlation signals, sets of simulated data based on the 3DID dataset with controllable correlation between interacting residues were also used, and it was demonstrated that the prediction consistently improves as the correlations increase.