Understanding proteome quantification in an interactive learning module on Google Cloud Platform.

This manuscript describes the development of a resource module that is part of a learning platform named 'NIGMS Sandbox for Cloud-based Learning' https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox at the beginning of this Supplement. This module delivers learning materials on protein quantification in an interactive format that uses appropriate cloud resources for data access and analyses. Quantitative proteomics is a rapidly growing discipline due to the cutting-edge technologies of high resolution mass spectrometry. There are many data types to consider for proteome quantification including data dependent acquisition, data independent acquisition, multiplexing with Tandem Mass Tag reporter ions, spectral counts, and more. As part of the NIH NIGMS Sandbox effort, we developed a learning module to introduce students to mass spectrometry terminology, normalization methods, statistical designs, and basics of R programming. By utilizing the Google Cloud environment, the learning module is easily accessible without the need for complex installation procedures. The proteome quantification module demonstrates the analysis using a provided TMT10plex data set using MS3 reporter ion intensity quantitative values in a Jupyter notebook with an R kernel. The learning module begins with the raw intensities, performs normalization, and differential abundance analysis using limma models, and is designed for researchers with a basic understanding of mass spectrometry and R programming language. Learners walk away with a better understanding of how to navigate Google Cloud Platform for proteomic research, and with the basics of mass spectrometry data analysis at the command line. This manuscript describes the development of a resource module that is part of a learning platform named ``NIGMS Sandbox for Cloud-based Learning'' https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox [1] at the beginning of this Supplement. This module delivers learning materials on the analysis of bulk and single-cell ATAC-seq data in an interactive format that uses appropriate cloud resources for data access and analyses.

Reusable tutorials for using cloud-based computing environments for the analysis of bacterial gene expression data from bulk RNA sequencing.

This manuscript describes the development of a resource module that is part of a learning platform named "NIGMS Sandbox for Cloud-based Learning" https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox at the beginning of this Supplement. This module delivers learning materials on RNA sequencing (RNAseq) data analysis in an interactive format that uses appropriate cloud resources for data access and analyses. Biomedical research is increasingly data-driven, and dependent upon data management and analysis methods that facilitate rigorous, robust, and reproducible research. Cloud-based computing resources provide opportunities to broaden the application of bioinformatics and data science in research. Two obstacles for researchers, particularly those at small institutions, are: (i) access to bioinformatics analysis environments tailored to their research; and (ii) training in how to use Cloud-based computing resources. We developed five reusable tutorials for bulk RNAseq data analysis to address these obstacles. Using Jupyter notebooks run on the Google Cloud Platform, the tutorials guide the user through a workflow featuring an RNAseq dataset from a study of prophage altered drug resistance in Mycobacterium chelonae. The first tutorial uses a subset of the data so users can learn analysis steps rapidly, and the second uses the entire dataset. Next, a tutorial demonstrates how to analyze the read count data to generate lists of differentially expressed genes using R/DESeq2. Additional tutorials generate read counts using the Snakemake workflow manager and Nextflow with Google Batch. All tutorials are open-source and can be used as templates for other analysis.

Transcriptomics and epigenetic data integration learning module on Google Cloud.

Multi-omics (genomics, transcriptomics, epigenomics, proteomics, metabolomics, etc.) research approaches are vital for understanding the hierarchical complexity of human biology and have proven to be extremely valuable in cancer research and precision medicine. Emerging scientific advances in recent years have made high-throughput genome-wide sequencing a central focus in molecular research by allowing for the collective analysis of various kinds of molecular biological data from different types of specimens in a single tissue or even at the level of a single cell. Additionally, with the help of improved computational resources and data mining, researchers are able to integrate data from different multi-omics regimes to identify new prognostic, diagnostic, or predictive biomarkers, uncover novel therapeutic targets, and develop more personalized treatment protocols for patients. For the research community to parse the scientifically and clinically meaningful information out of all the biological data being generated each day more efficiently with less wasted resources, being familiar with and comfortable using advanced analytical tools, such as Google Cloud Platform becomes imperative. This project is an interdisciplinary, cross-organizational effort to provide a guided learning module for integrating transcriptomics and epigenetics data analysis protocols into a comprehensive analysis pipeline for users to implement in their own work, utilizing the cloud computing infrastructure on Google Cloud. The learning module consists of three submodules that guide the user through tutorial examples that illustrate the analysis of RNA-sequence and Reduced-Representation Bisulfite Sequencing data. The examples are in the form of breast cancer case studies, and the data sets were procured from the public repository Gene Expression Omnibus. The first submodule is devoted to transcriptomics analysis with the RNA sequencing data, the second submodule focuses on epigenetics analysis using the DNA methylation data, and the third submodule integrates the two methods for a deeper biological understanding. The modules begin with data collection and preprocessing, with further downstream analysis performed in a Vertex AI Jupyter notebook instance with an R kernel. Analysis results are returned to Google Cloud buckets for storage and visualization, removing the computational strain from local resources. The final product is a start-to-finish tutorial for the researchers with limited experience in multi-omics to integrate transcriptomics and epigenetics data analysis into a comprehensive pipeline to perform their own biological research.This manuscript describes the development of a resource module that is part of a learning platform named ``NIGMS Sandbox for Cloud-based Learning'' https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox [16] at the beginning of this Supplement. This module delivers learning materials on the analysis of bulk and single-cell ATAC-seq data in an interactive format that uses appropriate cloud resources for data access and analyses.

Discerning structure versus speciation in phylogeographic analysis of Seepage Salamanders (Desmognathus aeneus) using demography, environment, geography, and phenotype.

Numerous mechanisms can drive speciation, including isolation by adaptation, distance, and environment. These forces can promote genetic and phenotypic differentiation of local populations, the formation of phylogeographic lineages, and ultimately, completed speciation. However, conceptually similar mechanisms may also result in stabilizing rather than diversifying selection, leading to lineage integration and the long-term persistence of population structure within genetically cohesive species. Processes that drive the formation and maintenance of geographic genetic diversity while facilitating high rates of migration and limiting phenotypic differentiation may thereby result in population genetic structure that is not accompanied by reproductive isolation. We suggest that this framework can be applied more broadly to address the classic dilemma of "structure" versus "species" when evaluating phylogeographic diversity, unifying population genetics, species delimitation, and the underlying study of speciation. We demonstrate one such instance in the Seepage Salamander (Desmognathus aeneus) from the southeastern United States. Recent studies estimated up to 6.3% mitochondrial divergence and four phylogenomic lineages with broad admixture across geographic hybrid zones, which could potentially represent distinct species supported by our species-delimitation analyses. However, while limited dispersal promotes substantial isolation by distance, microhabitat specificity appears to yield stabilizing selection on a single, uniform, ecologically mediated phenotype. As a result, climatic cycles promote recurrent contact between lineages and repeated instances of high migration through time. Subsequent hybridization is apparently not counteracted by adaptive differentiation limiting introgression, leaving a single unified species with deeply divergent phylogeographic lineages that nonetheless do not appear to represent incipient species.

Speciation Hypotheses from Phylogeographic Delimitation Yield an Integrative Taxonomy for Seal Salamanders (Desmognathus monticola).

Significant advances have been made in species delimitation and numerous methods can test precisely defined models of speciation, though the synthesis of phylogeography and taxonomy is still sometimes incomplete. Emerging consensus treats distinct genealogical clusters in genome-scale data as strong initial evidence of speciation in most cases, a hypothesis that must therefore be falsified under an explicit evolutionary model. We can now test speciation hypotheses linking trait differentiation to specific mechanisms of divergence with increasingly large data sets. Integrative taxonomy can, therefore, reflect an understanding of how each axis of variation relates to underlying speciation processes, with nomenclature for distinct evolutionary lineages. We illustrate this approach here with Seal Salamanders (Desmognathus monticola) and introduce a new unsupervised machine-learning approach for species delimitation. Plethodontid salamanders are renowned for their morphological conservatism despite extensive phylogeographic divergence. We discover 2 geographic genetic clusters, for which demographic and spatial models of ecology and gene flow provide robust support for ecogeographic speciation despite limited phenotypic divergence. These data are integrated under evolutionary mechanisms (e.g., spatially localized gene flow with reduced migration) and reflected in emergent properties expected under models of reinforcement (e.g., ethological isolation and selection against hybrids). Their genetic divergence is prima facie evidence for species-level distinctiveness, supported by speciation models and divergence along axes such as behavior, geography, and climate that suggest an ecological basis with subsequent reinforcement through prezygotic isolation. As data sets grow more comprehensive, species-delimitation models can be tested, rejected, or corroborated as explicit speciation hypotheses, providing for reciprocal illumination of evolutionary processes and integrative taxonomies. [Desmognathus; integrative taxonomy; machine learning; species delimitation.].

Accelerating genomic workflows using NVIDIA Parabricks.

BACKGROUND: As genome sequencing becomes better integrated into scientific research, government policy, and personalized medicine, the primary challenge for researchers is shifting from generating raw data to analyzing these vast datasets. Although much work has been done to reduce compute times using various configurations of traditional CPU computing infrastructures, Graphics Processing Units (GPUs) offer opportunities to accelerate genomic workflows by orders of magnitude. Here we benchmark one GPU-accelerated software suite called NVIDIA Parabricks on Amazon Web Services (AWS), Google Cloud Platform (GCP), and an NVIDIA DGX cluster. We benchmarked six variant calling pipelines, including two germline callers (HaplotypeCaller and DeepVariant) and four somatic callers (Mutect2, Muse, LoFreq, SomaticSniper). RESULTS: We achieved up to 65 × acceleration with germline variant callers, bringing HaplotypeCaller runtimes down from 36 h to 33 min on AWS, 35 min on GCP, and 24 min on the NVIDIA DGX. Somatic callers exhibited more variation between the number of GPUs and computing platforms. On cloud platforms, GPU-accelerated germline callers resulted in cost savings compared with CPU runs, whereas some somatic callers were more expensive than CPU runs because their GPU acceleration was not sufficient to overcome the increased GPU cost. CONCLUSIONS: Germline variant callers scaled well with the number of GPUs across platforms, whereas somatic variant callers exhibited more variation in the number of GPUs with the fastest runtimes, suggesting that, at least with the version of Parabricks used here, these workflows are less GPU optimized and require benchmarking on the platform of choice before being deployed at production scales. Our study demonstrates that GPUs can be used to greatly accelerate genomic workflows, thus bringing closer to grasp urgent societal advances in the areas of biosurveillance and personalized medicine.

Tecovirimat Resistance in Mpox Patients, United States, 2022-2023.

During the 2022 multinational outbreak of monkeypox virus (MPXV) infection, the antiviral drug tecovirimat (TPOXX; SIGA Technologies, Inc., https://www.siga.com) was deployed in the United States on a large scale for the first time. The MPXV F13L gene homologue encodes the target of tecovirimat, and single amino acid changes in F13 are known to cause resistance to tecovirimat. Genomic sequencing identified 11 mutations previously reported to cause resistance, along with 13 novel mutations. Resistant phenotype was determined using a viral cytopathic effect assay. We tested 124 isolates from 68 patients; 96 isolates from 46 patients were found to have a resistant phenotype. Most resistant isolates were associated with severely immunocompromised mpox patients on multiple courses of tecovirimat treatment, whereas most isolates identified by routine surveillance of patients not treated with tecovirimat remained sensitive. The frequency of resistant viruses remains relatively low (<1%) compared with the total number of patients treated with tecovirimat.

Antigen Receptor T Cells (CAR-T) Effectively Control Tumor Growth in a Colorectal Liver Metastasis Model.

BACKGROUND: Effective treatment of solid tumors requires multi-modality approaches. In many patients with stage IV liver disease, current treatments are not curative. Chimeric antigen receptor T cells (CAR-T) are an intriguing option following success in hematological malignancies, but this has not been translated to solid tumors. Limitations include sub-optimal delivery and elevated interstitial fluid pressures. We developed a murine model to test the impact of high-pressure regional delivery (HPRD) on trafficking to liver metastases (LM) and tumor response. MATERIALS AND METHODS: CAR-T were generated from CD45.1 mice and adoptively transferred into LM-bearing CD45.2 mice via regional or systemic delivery (RD, SD). Trafficking, tumor growth, and toxicity were evaluated with flow cytometry, tumor bioluminescence (TB, photons/sec log2-foldover baseline), and liver function tests (LFTs). RESULTS: RD of CAR-T was more effective at controlling tumor growth versus SD from post-treatment days (PTD) 2-7 (P = 0.002). HPRD resulted in increased CAR-T penetration versus low-pressure RD (LPRD, P = 0.004), suppression of tumor proliferation (P = 0.03), and trended toward improved long-term control at PTD17 (TB=3.7 versus 6.1, P = 0.47). No LFT increase was noted utilizing HPRD versus LPRD (AST/ALT P = 0.65/0.84) while improved LFTs in RD versus SD groups suggested better tumor control (HPRD AST/ALT P = 0.04/0.04, LPRD AST/ALT P = 0.02/0.02). CONCLUSIONS: Cellular immunotherapy is an emerging option for solid tumors. Our model suggests RD and HPRD improved CAR-T penetration into solid tumors with improved short-term tumor control. Barriers associated with SD can be overcome using RD techniques to maximize therapeutic delivery and HPRD may further augment efficacy without increased toxicity.

Speciation and secondary contact in a fossorial island endemic, the São Tomé caecilian.

A period of isolation in allopatry typically precedes local adaptation and subsequent divergence among lineages. Alternatively, locally adapted phenotypes may arise and persist in the face of gene flow, resulting in strong correlations between ecologically-relevant phenotypic variation and corresponding environmental gradients. Quantifying genetic, ecological, and phenotypic divergence in such lineages can provide insights into the abiotic and biotic mechanisms that structure populations and drive the accumulation of phenotypic and taxonomic diversity. Low-vagility organisms whose distributions span ephemeral geographic barriers present the ideal evolutionary context within which to address these questions. Here, we combine genetic (mtDNA and genome-wide SNPs) and phenotypic data to investigate the divergence history of caecilians (Amphibia: Gymnophiona) endemic to the oceanic island of São Tomé in the Gulf of Guinea archipelago. Consistent with a previous mtDNA study, we find two phenotypically and genetically distinct lineages that occur along a north-to-south axis with extensive admixture in the centre of the island. Demographic modelling supports divergence in allopatry (~300 kya) followed by secondary contact (~95 kya). Consequently, in contrast to a morphological study that interpreted latitudinal phenotypic variation in these caecilians as a cline within a single widespread species, our analyses suggest a history of allopatric lineage divergence and subsequent hybridization that may have blurred species boundaries. We propose that late Pleistocene volcanic activity favoured allopatric divergence between these lineages with local adaptation to climate maintaining a stable hybrid zone in the centre of São Tomé Island. Our study joins a growing number of systems demonstrating lineage divergence on volcanic islands with stark environmental transitions across small geographic distances.

Um período de isolamento em alopatria geralmente precede adaptação local e divergência subsequente entre linhagens evolutivas. Alternativamente, fenótipos adaptados localmente podem surgir e persistir apesar de fluxo gênico, resultando em fortes correlações entre variação fenotípica ecologicamente relevante e os gradientes ambientais correspondentes. Quantificar divergência genética, ecológica e fenotípica em tais linhagens pode ajudar a clarificar os mecanismos abióticos e bióticos que estruturam as populações e levam ao acúmulo de diversidade fenotípica e taxonômica. Organismos de baixa vagilidade, cujas áreas de distribuição incluem barreiras geográficas efêmeras, representam um contexto evolutivo ideal para abordar essas questões. Neste estudo, combinamos dados genéticos (mtDNA e SNPs genômicos) e fenotípicos para investigar a história de divergência de cecílias endêmicas da ilha oceânica de São Tomé, no arquipélago do Golfo da Guiné. Consistentemente com um estudo anterior de mtDNA, encontramos duas linhagens fenotipicamente e geneticamente distintas que ocorrem ao longo de um eixo norte-sul, com extensa mistura genética no centro da ilha. Modelagem demográfica suportou um cenário de divergência em alopatria (~ 300 mil anos atrás) seguida de contato secundário (~ 95 mil anos atrás). Ao contrário de um estudo morfológico que interpretou a variação fenotípica latitudinal nessas cecílias como uma clina dentro de uma única espécie amplamente difundida, nossas análises sugerem uma história de divergência de linhagens em alopatria e subsequente hibridização que pode ter confundido os limites das espécies. Propomos que atividade vulcânica durante o Pleistoceno tardio favoreceu divergência alopátrica entre essas linhagens, com adaptação local ao clima mantendo uma zona híbrida estável no centro da Ilha de São Tomé. Nosso estudo se une a um número crescente de sistemas que demonstram divergência entre linhagens em ilhas vulcânicas com transições ambientais marcantes ao longo de distâncias geográficas curtas.

Impacts of the Toba eruption and montane forest expansion on diversification in Sumatran parachuting frogs (Rhacophorus).

Catastrophic events, such as volcanic eruptions, can have profound impacts on the demographic histories of resident taxa. Due to its presumed effect on biodiversity, the Pleistocene eruption of super-volcano Toba has received abundant attention. We test the effects of the Toba eruption on the diversification, genetic diversity, and demography of three co-distributed species of parachuting frogs (Genus Rhacophorus) on Sumatra. We generate target-capture data (~950 loci and ~440,000 bp) for three species of parachuting frogs and use these data paired with previously generated double digest restriction-site associated DNA (ddRADseq) data to estimate population structure and genetic diversity, to test for population size changes using demographic modelling, and to estimate the temporal clustering of size change events using a full-likelihood Bayesian method. We find that populations around Toba exhibit reduced genetic diversity compared with southern populations, and that northern populations exhibit a shift in effective population size around the time of the eruption (~80 kya). However, we infer a stronger signal of expansion in southern populations around ~400 kya, and at least two of the northern populations may have also expanded at this time. Taken together, these findings suggest that the Toba eruption precipitated population declines in northern populations, but that the demographic history of these three species was also strongly impacted by mid-Pleistocene forest expansion during glacial periods. We propose local rather than regional effects of the Toba eruption, and emphasize the dynamic nature of diversification on the Sunda Shelf.

Phylogenomic data reveal reticulation and incongruence among mitochondrial candidate species in Dusky Salamanders (Desmognathus).

Gene flow between evolutionarily distinct lineages is increasingly recognized as a common occurrence. Such processes distort our ability to diagnose and delimit species, as well as confound attempts to estimate phylogenetic relationships. A conspicuous example is Dusky Salamanders (Desmognathus), a common model-system for ecology, evolution, and behavior. Only 22 species are described, 7 in the last 40 years. However, mitochondrial datasets indicate the presence of up to 45 "candidate species" and multiple paraphyletic taxa presenting a complex history of reticulation. Some authors have even suggested that the search for species boundaries in the group may be in vain. Here, we analyze nuclear and mitochondrial data containing 161 individuals from at least 49 distinct evolutionary lineages that we treat as candidate species. Concatenated and species-tree methods do not estimate fully resolved relationships among these taxa. Comparing topologies and applying methods for estimating phylogenetic networks, we find strong support for numerous instances of hybridization throughout the history of the group. We suggest that these processes may be more common than previously thought across the phylogeography-phylogenetics continuum, and that while the search for species boundaries in Desmognathus may not be in vain, it will be complicated by factors such as crypsis, parallelism, and gene-flow.

Diversification of bent-toed geckos (Cyrtodactylus) on Sumatra and west Java.

Complex geological processes often drive biotic diversification on islands. The islands of Sumatra and Java have experienced dramatic historical changes, including isolation by marine incursions followed by periodic connectivity with the rest of Sundaland across highland connections. To determine how this geological history influenced island invasions, we investigated the colonization history and diversification of bent-toed geckos (genus Cyrtodactylus) on Sumatra and west Java. We used mitochondrial and nuclear sequence data to explore species boundaries, estimate phylogenetic relationships and divergence times, and to reconstruct ancestral range evolution. We found that Sumatran and Javan Cyrtodactylus were closely related to species from the Thai-Malay Peninsula, rather than from Borneo, and that Cyrtodactylus most likely dispersed to Sumatra three times during the late Oligocene and early Miocene. Similarly, Cyrtodactylus invaded west Java from Sumatra once in the early Miocene. Our results suggest that despite isolation by marine incursions during much of the Miocene, Cyrtodactylus dispersed to and from Sumatra and west Java likely via land bridges, and that in situ diversification occurred several times on Sumatra.

The effect of missing data on coalescent species delimitation and a taxonomic revision of whipsnakes (Colubridae: Masticophis).

A stable alpha taxonomy is essential to understanding evolutionary processes and achieving effective conservation aims. Taxonomy depends on the identification of independently evolving lineages, and the delimitation of these lineages based on multiple lines of evidence. Coalescent species delimitation within an integrative framework has increased the rigor of the delimitation process. Here we use genome-wide SNP data and coalescent species delimitation to explore lineage relationships within several North American whipsnake species, test the species status of several lineages, and test the effect of missing data on species delimitation. We find support for the elevation of several previously recognized subspecies to full species status, and formally elevate two species. This study demonstrates the power of molecular data and model-based delimitation methods to identify evolutionary relationships, and finds that missing data have little impact on the outcome of delimitation analyses.

Synchronous diversification of parachuting frogs (Genus Rhacophorus) on Sumatra and Java.

Geological and climatological processes can drive the synchronous diversification of co-distributed species. The islands of Sumatra and Java have experienced complex geological and climatological histories, including extensive sea-level changes and the formation of valleys between northern, central, and southern components of the Barisan Mountain Range, which may have promoted diversification of their resident species. We investigate diversification on these islands using 13 species of the parachuting frog genus Rhacophorus. We use both mitochondrial and nuclear sequence data, along with genome-wide SNPs to estimate phylogenetic structure and divergence times, and to test for synchronous diversification. We find support for synchronous divergence among sister-species pairs from Sumatra and Java ∼9 Ma, as well as of populations of four co-distributed taxa on Sumatra ∼5.6 Ma. We found that sister species diverged in allopatry on Sumatra and conclude that divergence on Sumatra and Java was affected by sea-level fluctuations that promoted isolation in allopatry.

Geographical features are the predominant driver of molecular diversification in widely distributed North American whipsnakes.

Allopatric divergence following the formation of geographical features has been implicated as a major driver of evolutionary diversification. Widespread species complexes provide opportunities to examine allopatric divergence across varying degrees of isolation in both time and space. In North America, several geographical features may play such a role in diversification, including the Mississippi River, Pecos River, Rocky Mountains, Cochise Filter Barrier, Gulf of California and Isthmus of Tehuantepec. We used thousands of nuclear single nucleotide polymorphisms (SNPs) and mitochondrial DNA from several species of whipsnakes (genera Masticophis and Coluber) distributed across North and Central America to investigate the role that these geographical features have played on lineage divergence. We hypothesize that these features restrict gene flow and separate whipsnakes into diagnosable genomic clusters. We performed genomic clustering and phylogenetic reconstructions at the species and population levels using Bayesian and likelihood analyses and quantified migration levels across geographical features to assess the degree of genetic isolation due to allopatry. Our analyses suggest that (i) major genetic divisions are often consistent with isolation by geographical features, (ii) migration rates between clusters are asymmetrical across major geographical features, and (iii) areas that receive proportionally more migrants possess higher levels of genetic diversity. Collectively, our findings suggest that multiple features of the North American landscape contributed to allopatric divergence in this widely distributed snake group.

The Draft Genome Sequences of 50 Salamander species (Caudata, Amphibia).

We present partial genome sequences of 50 salamander species (Urodela) from 10 genera and 4 families. These span nearly the entire range of genome sizes in salamanders, from ~14-130GB, the latter of which is among the largest of all animal genomes. Only three salamander genomes were available to this point, from Ambystomatidae (one species) and Salamandridae (two species from two genera), to which we have added Amphiumidae (one species), Plethodontidae (45 species from 6 genera), Proteidae (one species), and Sirenidae (three species from two genera). These span ~140 million years of evolutionary divergence, leaving only Cryptobranchidae, Hynobiidae, and Rhyacotritonidae as salamander families without genome assemblies. These data should facilitate additional future work on speciation and genome evolution, both within Urodela and across Animalia.

Investigation of machine learning algorithms for taxonomic classification of marine metagenomes.

Microbial communities play key roles in ocean ecosystems through regulation of biogeochemical processes such as carbon and nutrient cycling, food web dynamics, and gut microbiomes of invertebrates, fish, reptiles, and mammals. Assessments of marine microbial diversity are therefore critical to understanding spatiotemporal variations in microbial community structure and function in ocean ecosystems. With recent advances in DNA shotgun sequencing for metagenome samples and computational analysis, it is now possible to access the taxonomic and genomic content of ocean microbial communities to study their structural patterns, diversity, and functional potential. However, existing taxonomic classification tools depend upon manually curated phylogenetic trees, which can create inaccuracies in metagenomes from less well-characterized communities, such as from ocean water. Herein, we explore the utility of deep learning tools-DeepMicrobes and a novel Residual Network architecture-that leverage natural language processing and convolutional neural network architectures to map input sequence data (k-mers) to output labels (taxonomic groups) without reliance on a curated taxonomic tree. We trained both models using metagenomic reads simulated from marine microbial genomes in the MarRef database. The performance of both models (accuracy, precision, and percent microbe predicted) was compared with the standard taxonomic classification tool Kraken2 using 10 complex metagenomic data sets simulated from MarRef. Our results demonstrate that time, compute power, and microbial genomic diversity still pose challenges for machine learning (ML). Moreover, our results suggest that high genome coverage and rectification of class imbalance are prerequisites for a well-trained model, and therefore should be a major consideration in future ML work. IMPORTANCE Taxonomic profiling of microbial communities is essential to model microbial interactions and inform habitat conservation. This work develops approaches in constructing training/testing data sets from publicly available marine metagenomes and evaluates the performance of machine learning (ML) approaches in read-based taxonomic classification of marine metagenomes. Predictions from two models are used to test accuracy in metagenomic classification and to guide improvements in ML approaches. Our study provides insights on the methods, results, and challenges of deep learning on marine microbial metagenomic data sets. Future machine learning approaches can be improved by rectifying genome coverage and class imbalance in the training data sets, developing alternative models, and increasing the accessibility of computational resources for model training and refinement.

Candidate-species delimitation in Desmognathus salamanders reveals gene flow across lineage boundaries, confounding phylogenetic estimation and clarifying hybrid zones.

Dusky Salamanders (genus Desmognathus) currently comprise only 22 described, extant species. However, recent mitochondrial and nuclear estimates indicate the presence of up to 49 candidate species based on ecogeographic sampling. Previous studies also suggest a complex history of hybridization between these lineages. Studies in other groups suggest that disregarding admixture may affect both phylogenetic inference and clustering-based species delimitation. With a dataset comprising 233 Anchored Hybrid Enrichment (AHE) loci sequenced for 896 Desmognathus specimens from all 49 candidate species, we test three hypotheses regarding (i) species-level diversity, (ii) hybridization and admixture, and (iii) misleading phylogenetic inference. Using phylogenetic and population-clustering analyses considering gene flow, we find support for at least 47 candidate species in the phylogenomic dataset, some of which are newly characterized here while others represent combinations of previously named lineages that are collapsed in the current dataset. Within these, we observe significant phylogeographic structure, with up to 64 total geographic genetic lineages, many of which hybridize either narrowly at contact zones or extensively across ecological gradients. We find strong support for both recent admixture between terminal lineages and ancient hybridization across internal branches. This signal appears to distort concatenated phylogenetic inference, wherein more heavily admixed terminal specimens occupy apparently artifactual early-diverging topological positions, occasionally to the extent of forming false clades of intermediate hybrids. Additional geographic and genetic sampling and more robust computational approaches will be needed to clarify taxonomy, and to reconstruct a network topology to display evolutionary relationships in a manner that is consistent with their complex history of reticulation.

Genomic library preparation and hybridization capture of formalin-fixed tissues and allozyme supernatant for population genomics and considerations for combining capture- and RADseq-based single nucleotide polymorphism data sets.

Until recently many historical museum specimens were largely inaccessible to genomic inquiry, but high-throughput sequencing (HTS) approaches have allowed researchers to successfully sequence genomic DNA from dried and fluid-preserved museum specimens. In addition to preserved specimens, many museums contain large series of allozyme supernatant samples, but the amenability of these samples to HTS has not yet been assessed. Here, we compared the performance of a target-capture approach using alternative sources of genomic DNA from 10 specimens of spring salamanders (Plethodontidae: Gyrinophilus porphyriticus) collected between 1985 and 1990: allozyme supernatants, allozyme homogenate pellets and formalin-fixed tissues. We designed capture probes based on double-digest restriction-site associated sequencing (RADseq) derived loci from frozen blood samples available for seven of the specimens and assessed the success and consistency of capture and RADseq approaches. This study design enabled direct comparisons of data quality and potential biases among the different data sets for phylogenomic and population genomic analyses. We found that in phylogenetic analyses, all enrichment types for a given specimen clustered together. In principal component space all capture-based samples clustered together, but RADseq samples did not cluster with corresponding capture-based samples. Single nucleotide polymorphism calls were on average 18.3% different between enrichment types for a given individual, but these discrepancies were primarily due to differences in heterozygous/homozygous single nucleotide polymorphism calls. We demonstrate that both allozyme supernatant and formalin-fixed samples can be successfully used for population genomic analyses and we discuss ways to identify and reduce biases associated with combining capture and RADseq data.

Assessing the Future of Solid Tumor Immunotherapy.

With the advent of cancer immunotherapy, there has been a major improvement in patient's quality of life and survival. The growth of cancer immunotherapy has dramatically changed our understanding of the basics of cancer biology and has altered the standards of care (surgery, radiotherapy, and chemotherapy) for patients. Cancer immunotherapy has generated significant excitement with the success of chimeric antigen receptor (CAR) T cell therapy in particular. Clinical results using CAR-T for hematological malignancies have led to the approval of four CD19-targeted and one B-cell maturation antigen (BCMA)-targeted cell therapy products by the US Food and Drug Administration (FDA). Also, immune checkpoint inhibitors such as antibodies against Programmed Cell Death-1 (PD-1), Programmed Cell Death Ligand-1 (PD-L1), and Cytotoxic T-Lymphocyte-Associated Antigen 4 (CTLA-4) have shown promising therapeutic outcomes and long-lasting clinical effect in several tumor types and patients who are refractory to other treatments. Despite these promising results, the success of cancer immunotherapy in solid tumors has been limited due to several barriers, which include immunosuppressive tumor microenvironment (TME), inefficient trafficking, and heterogeneity of tumor antigens. This is further compounded by the high intra-tumoral pressure of solid tumors, which presents an additional challenge to successfully delivering treatments to solid tumors. In this review, we will outline and propose specific approaches that may overcome these immunological and physical barriers to improve the outcomes in solid tumor patients receiving immunotherapies.