Publications

Dental caries, the most common chronic infectious disease worldwide, has a complex etiology involving the interplay of microbial and host factors that are not completely understood. In this study, the oral microbiome and 38 host cytokines and chemokines were analyzed across 23 children with caries and 24 children with healthy dentition. De novo assembly of metagenomic sequencing obtained 527 metagenome-assembled genomes (MAGs), representing 150 bacterial species. Forty-two of these species had no genomes in public repositories, thereby representing novel taxa. These new genomes greatly expanded the known pangenomes of many oral clades, including the enigmatic Saccharibacteria clades G3 and G6, which had distinct functional repertoires compared to other oral Saccharibacteria. Saccharibacteria are understood to be obligate epibionts, which are dependent on host bacteria. These data suggest that the various Saccharibacteria clades may rely on their hosts for highly distinct metabolic requirements, which would have significant evolutionary and ecological implications. Across the study group, Rothia, Neisseria, and Haemophilus spp. were associated with good dental health, whereas Prevotella spp., Streptococcus mutans, and Human herpesvirus 4 (Epstein-Barr virus [EBV]) were more prevalent in children with caries. Finally, 10 of the host immunological markers were significantly elevated in the caries group, and co-occurrence analysis provided an atlas of potential relationships between microbes and host immunological molecules. Overall, this study illustrated the oral microbiome at an unprecedented resolution and contributed several leads for further study that will increase the understanding of caries pathogenesis and guide therapeutic development.

The recent detection of GW190521 stimulated ideas on how to populate the predicted black hole (BH) pair-instability (PI) mass gap. One proposal is the dynamical merger of two stars below the PI regime forming a star with a small core and an oversized envelope. We outline the main challenges this scenario faces to form one BH in the gap. In particular, the core needs to avoid growing during the merger, and the merger product needs to retain enough mass, including in the subsequent evolution, and at core collapse (CC). We explore this scenario with detailed stellar evolution calculations, starting with ad hoc initial conditions enforcing no core growth during the merger. We find that these massive merger products are likely to be helium-rich and spend most of their remaining lifetime within regions of instabilities in the Herzsprung–Russell diagram, such as luminous blue variable eruptions. An energetic estimate of the amount of mass loss neglecting the back reaction of the star suggests that the total amount of mass that can be removed at low metallicity is ≲1 M⊙. This is small enough that at CC our models are retaining sufficient mass to form BHs in the PI gap similar to the recent ones detected by LIGO/Virgo. However, mass loss at the time of merger, the resulting core structure, and the mass loss at CC still need to be quantified for these models to confirm the viability of this scenario.

An identification is found between meta-learning and the problem of determining the ground state of a randomly generated Hamiltonian drawn from a known ensemble. A model-agnostic meta-learning approach is proposed to solve the associated learning problem and a preliminary experimental study of random Max-Cut problems indicates that the resulting Meta Variational Monte Carlo accelerates training and improves convergence.

In cells, cytoskeletal filament networks are responsible for cell movement, growth, and division. Filaments in the cytoskeleton are driven and organized by crosslinking molecular motors. In reconstituted cytoskeletal systems, motor activity is responsible for far-fromequilibrium phenomena such as active stress, self-organized flow, and spontaneous nematic defect generation. How microscopic interactions between motors and filaments lead to larger-scale dynamics remains incompletely understood. To build from motor-filament interactions to predict bulk behavior of cytoskeletal systems, more computationally efficient techniques for modeling motor-filament interactions are needed. Here we derive a coarsegraining hierarchy of explicit and continuum models for crosslinking motors that bind to and walk on filament pairs. We compare the steady-state motor distribution and motorinduced filament motion for the different models and analyze their computational cost. All three models agree well in the limit of fast motor binding kinetics. Evolving a truncated moment expansion of motor density speeds the computation by 103–106
compared to the explicit or continuous-density simulations, suggesting an approach for more efficient simulation of large networks. These tools facilitate further study of motor-filament networks on micrometer to millimeter length scales.

BACKGROUND
The efficacy of public health measures to control the transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has not been well studied in young adults.

METHODS
We investigated SARS-CoV-2 infections among U.S. Marine Corps recruits who underwent a 2-week quarantine at home followed by a second supervised 2-week quarantine at a closed college campus that involved mask wearing, social distancing, and daily temperature and symptom monitoring. Study volunteers were tested for SARS-CoV-2 by means of quantitative polymerase-chain-reaction (qPCR) assay of nares swab specimens obtained between the time of arrival and the second day of supervised quarantine and on days 7 and 14. Recruits who did not volunteer for the study underwent qPCR testing only on day 14, at the end of the quarantine period. We performed phylogenetic analysis of viral genomes obtained from infected study volunteers to identify clusters and to assess the epidemiologic features of infections.

RESULTS
A total of 1848 recruits volunteered to participate in the study; within 2 days after arrival on campus, 16 (0.9%) tested positive for SARS-CoV-2, 15 of whom were asymptomatic. An additional 35 participants (1.9%) tested positive on day 7 or on day 14. Five of the 51 participants (9.8%) who tested positive at any time had symptoms in the week before a positive qPCR test. Of the recruits who declined to participate in the study, 26 (1.7%) of the 1554 recruits with available qPCR results tested positive on day 14. No SARS-CoV-2 infections were identified through clinical qPCR testing performed as a result of daily symptom monitoring. Analysis of 36 SARS-CoV-2 genomes obtained from 32 participants revealed six transmission clusters among 18 participants. Epidemiologic analysis supported multiple local transmission events, including transmission between roommates and among recruits within the same platoon.

CONCLUSIONS
Among Marine Corps recruits, approximately 2% who had previously had negative results for SARS-CoV-2 at the beginning of supervised quarantine, and less than 2% of recruits with unknown previous status, tested positive by day 14. Most recruits who tested positive were asymptomatic, and no infections were detected through daily symptom monitoring. Transmission clusters occurred within platoons. (Funded by the Defense Health Agency and others.)

In lung cancer, enrichment of the lower airway microbiota with oral commensals commonly occurs and ex vivo models support that some of these bacteria can trigger host transcriptomic signatures associated with carcinogenesis. Here, we show that this lower airway dysbiotic signature was more prevalent in group IIIB-IV TNM stage lung cancer and is associated with poor prognosis, as shown by decreased survival among subjects with early stage disease (I-IIIA) and worse tumor progression as measured by RECIST scores among subjects with IIIB-IV stage disease. In addition, this lower airway microbiota signature was associated with upregulation of IL-17, PI3K, MAPK and ERK pathways in airway transcriptome, and we identified Veillonella parvula as the most abundant taxon driving this association. In a KP lung cancer model, lower airway dysbiosis with V. parvula led to decreased survival, increased tumor burden, IL-17 inflammatory phenotype and activation of checkpoint inhibitor markers.

Galactic outflows have density, temperature, and velocity variations at least as large as that of the multiphase, turbulent interstellar medium (ISM) from which they originate. We have conducted a suite of parsec-resolution numerical simulations using the TIGRESS framework, in which outflows emerge as a consequence of interaction between supernovae (SNe) and the star-forming ISM. The outflowing gas is characterized by two distinct thermal phases, cool (T10^6 K), with most mass carried by the cool phase and most energy and newly-injected metals carried by the hot phase. Both components have a broad distribution of outflow velocity, and especially for cool gas this implies a varying fraction of escaping material depending on the halo potential. Informed by the TIGRESS results, we develop straightforward analytic formulae for the joint probability density functions (PDFs) of mass, momentum, energy, and metal loading as distributions in outflow velocity and sound speed. The model PDFs have only two parameters, SFR surface density \Sigma_SFR and the metallicity of the ISM, and fully capture the behavior of the original TIGRESS simulation PDFs over \Sigma_SFR~(10^{-4},1)M_sun/kpc^2/yr. Employing PDFs from resolved simulations will enable galaxy formation subgrid model implementations with wind velocity and temperature (as well as total loading factors) that are based on theoretical predictions rather than empirical tuning. This is a critical step to incorporate advances from TIGRESS and other high-resolution simulations in future cosmological hydrodynamics and semi-analytic galaxy formation models. We release a python package to prototype our model and to ease its implementation.

We engineered a machine learning approach, MSHub, to enable auto-deconvolution of gas chromatography–mass spectrometry (GC–MS) data. We then designed workflows to enable the community to store, process, share, annotate, compare and perform molecular networking of GC–MS data within the Global Natural Product Social (GNPS) Molecular Networking analysis platform. MSHub/GNPS performs auto-deconvolution of compound fragmentation patterns via unsupervised non-negative matrix factorization and quantifies the reproducibility of fragmentation patterns across samples.

The effective diffusivity of a Brownian tracer in unidirectional flow is well known to be enhanced due to shear by the classic phenomenon of Taylor dispersion. At long times, the average concentration of the tracer follows a simplified advection–diffusion equation with an effective shear-dependent dispersivity. In this work, we make use of the generalized Taylor dispersion theory for periodic domains to analyze tracer dispersion by peristaltic pumping. In channels with small aspect ratios, asymptotic expansions in the lubrication limit are employed to obtain analytical expressions for the dispersion coefficient at both small and high Péclet numbers. Channels of arbitrary aspect ratios are also considered using a boundary integral formulation for the fluid flow coupled to a conservation equation for the effective dispersivity, which is solved using the finite-volume method. Our theoretical calculations, which compare well with results from Brownian dynamics simulations, elucidate the effects of channel geometry and pumping strength on shear-induced dispersion. We further discuss the connection between the present problem and dispersion due to Taylor’s swimming sheet and interpret our results in the purely diffusive regime in the context of Fick–Jacobs theory. Our results provide the theoretical basis for understanding passive scalar transport in peristaltic flow, for instance, in the ureter or in microfluidic peristaltic pumps.

Computing sequence similarity is a fundamental task in biology, with alignment forming the basis for the annotation of genes and genomes and providing the core data structures for evolutionary analysis. Standard approaches are a mainstay of modern molecular biology and rely on variations of edit distance to obtain explicit alignments between pairs of biological sequences. However, sequence alignment algorithms struggle with remote homology tasks and cannot identify similarities between many pairs of proteins with similar structures and likely homology. Recent work suggests that using machine learning language models can improve remote homology detection. To this end, we introduce DeepBLAST, that obtains explicit alignments from residue embeddings learned from a protein language model integrated into an end-to-end differentiable alignment framework. This approach can be accelerated on the GPU architectures and outperforms conventional sequence alignment techniques in terms of both speed and accuracy when identifying structurally similar proteins.