Search for: All records

Abstract. We introduce physics-informed multimodal autoencoders (PIMA) - a variational inference framework for discovering shared information in multimodal datasets. Individual modalities are embedded into a shared latent space and fused through a product-of-experts formulation, enabling a Gaussian mixture prior to identify shared features. Sampling from clusters allows cross-modal generative modeling, with a mixture-of-experts decoder that imposes inductive biases from prior scientific knowledge and thereby imparts structured disentanglement of the latent space. This approach enables cross-modal inference and the discovery of features in high-dimensional heterogeneous datasets. Consequently, this approach provides a means to discover fingerprints in multimodal scientific datasets and to avoid traditional bottlenecks related to high-fidelity measurement and characterization of scientific 

Control volume analysis models physics via the exchange of generalized fluxes between subdomains. We introduce a scientific machine learning framework adopting a partition of unity architecture to identify physically-relevant control volumes, with generalized fluxes between subdomains encoded via Whitney forms. The approach provides a differentiable parameterization of geometry which may be trained in an end-to-end fashion to extract reduced models from full field data while exactly preserving physics. The architecture admits a data-driven finite element exterior calculus allowing discovery of mixed finite element spaces with closed form quadrature rules. An equivalence between Whitney forms and graph networks reveals that the geometric problem of control volume learning is equivalent to an unsupervised graph discovery problem. The framework is developed for manifolds in arbitrary dimension, with examples provided for 𝐻(div) problems in ℝ2 establishing convergence and structure preservation properties. Finally, we consider a lithium-ion battery problem where we discover a reduced finite element space encoding transport pathways from high-fidelity microstructure resolved simulations. The approach reduces the 5.89M finite element simulation to 136 elements while reproducing pressure to under 0.1% error and preserving conservation. 

Abstract Spin-orbit coupling in noncentrosymmetric crystals leads to spin-momentum locking – a directional relationship between an electron’s spin angular momentum and its linear momentum. Isotropic orthogonal Rashba spin-momentum locking has been studied for decades, while its counterpart, isotropic parallel Weyl spin-momentum locking has remained elusive in experiments. Theory predicts that Weyl spin-momentum locking can only be realized in structurally chiral cubic crystals in the vicinity of Kramers-Weyl or multifold fermions. Here, we use spin- and angle-resolved photoemission spectroscopy to evidence Weyl spin-momentum locking of multifold fermions in the chiral topological semimetal PtGa. We find that the electron spin of the Fermi arc surface states is orthogonal to their Fermi surface contour for momenta close to the projection of the bulk multifold fermion at the Γ point, which is consistent with Weyl spin-momentum locking of the latter. The direct measurement of the bulk spin texture of the multifold fermion at the R point also displays Weyl spin-momentum locking. The discovery of Weyl spin-momentum locking may lead to energy-efficient memory devices and Josephson diodes based on chiral topological semimetals. 

Fewer than half of individuals with a suspected Mendelian or monogenic condition receive a precise molecular diagnosis after comprehensive clinical genetic testing. Improvements in data quality and costs have heightened interest in using long-read sequencing (LRS) to streamline clinical genomic testing, but the absence of control data sets for variant filtering and prioritization has made tertiary analysis of LRS data challenging. To address this, the 1000 Genomes Project (1KGP) Oxford Nanopore Technologies Sequencing Consortium aims to generate LRS data from at least 800 of the 1KGP samples. Our goal is to use LRS to identify a broader spectrum of variation so we may improve our understanding of normal patterns of human variation. Here, we present data from analysis of the first 100 samples, representing all 5 superpopulations and 19 subpopulations. These samples, sequenced to an average depth of coverage of 37× and sequence read N50 of 54 kbp, have high concordance with previous studies for identifying single nucleotide and indel variants outside of homopolymer regions. Using multiple structural variant (SV) callers, we identify an average of 24,543 high-confidence SVs per genome, including shared and private SVs likely to disrupt gene function as well as pathogenic expansions within disease-associated repeats that were not detected using short reads. Evaluation of methylation signatures revealed expected patterns at known imprinted loci, samples with skewed X-inactivation patterns, and novel differentially methylated regions. All raw sequencing data, processed data, and summary statistics are publicly available, providing a valuable resource for the clinical genetics community to discover pathogenic SVs. 

ABSTRACT Gaining a more complete understanding of a species' dietary variability is crucial to properly discern distribution, population growth trends, and conservation actions. Endangered mountain gorillas live in topographically complex forests covering a wide elevational range and diverse habitat matrices. Since 1967, mountain gorillas have been studied at high elevations in the southwest of the Volcanoes National Park (VNP) in Rwanda, where groups use different compositions of habitats and have been growing at higher rates than groups in the northeast VNP region, which is characterized by lower elevations. Building on previous efforts, we describe dietary variability among VNP mountain gorilla groups by integrating data from groups ranging in the northeast VNP. We assessed and compared nutritional components of key foods (making up 80% of the diet) to better understand whether variation in diet quality could be linked to within‐population growth differences. Feeding and ranging data were collected between November 2019 and December 2022, using long‐term monitoring data, group scans, and focal animal sampling. To compare diet quality, we combined nutritional values from newly collected food plants and previously collected and assessed food plant samples using comparable field and laboratory methods. We recorded 57 new foods for the study population. Groups in the southwest (N = 8) and the northeast (N = 4) regions of VNP used different vegetation zones, and there was high dietary variability with low diet overlap among these regions. Although northeast groups rely on more diverse diets, key foods (making up ~80% of the diet) had comparable nutrient concentrations to southwest groups. This suggests that diet quality is unlikely to be a main driver of observed heterogeneous population growth. For follow‐up research, we discuss alternative explanations linked to food distribution, biomass, and energy expenditure to access foods. Our findings add important information for future habitat suitability assessments essential for mountain gorilla conservation management and habitat restoration and expansion efforts. 

Abstract Background Introgressive hybridization can reassort genetic variants into beneficial combinations, permitting adaptation to new ecological niches. To evaluate evolutionary patterns and dynamics that contribute to introgression, we investigate six wild Vitis species that are native to the Southwestern United States and useful for breeding grapevine ( V. vinifera ) rootstocks. Results By creating a reference genome assembly from one wild species, V. arizonica , and by resequencing 130 accessions, we focus on identifying putatively introgressed regions (pIRs) between species. We find six species pairs with signals of introgression between them, comprising up to ~ 8% of the extant genome for some pairs. The pIRs tend to be gene poor, located in regions of high recombination and enriched for genes implicated in disease resistance functions. To assess potential pIR function, we explore SNP associations to bioclimatic variables and to bacterial levels after infection with the causative agent of Pierce’s disease ( Xylella fastidiosa ). pIRs are enriched for SNPs associated with both climate and bacterial levels, suggesting that introgression is driven by adaptation to biotic and abiotic stressors. Conclusions Altogether, this study yields insights into the genomic extent of introgression, potential pressures that shape adaptive introgression, and the evolutionary history of economically important wild relatives of a critical crop. 

Serial crystallography of membrane proteins often employs high-viscosity injectors (HVIs) to deliver micrometre-sized crystals to the X-ray beam. Typically, the carrier medium is a lipidic cubic phase (LCP) media, which can also be used to nucleate and grow the crystals. However, despite the fact that the LCP is widely used with HVIs, the potential impact of the injection process on the LCP structure has not been reported and hence is not yet well understood. The self-assembled structure of the LCP can be affected by pressure, dehydration and temperature changes, all of which occur during continuous flow injection. These changes to the LCP structure may in turn impact the results of X-ray diffraction measurements from membrane protein crystals. To investigate the influence of HVIs on the structure of the LCP we conducted a study of the phase changes in monoolein/water and monoolein/buffer mixtures during continuous flow injection, at both atmospheric pressure and under vacuum. The reservoir pressure in the HVI was tracked to determine if there is any correlation with the phase behaviour of the LCP. The results indicated that, even though the reservoir pressure underwent (at times) significant variation, this did not appear to correlate with observed phase changes in the sample stream or correspond to shifts in the LCP lattice parameter. During vacuum injection, there was a three-way coexistence of the gyroid cubic phase, diamond cubic phase and lamellar phase. During injection at atmospheric pressure, the coexistence of a cubic phase and lamellar phase in the monoolein/water mixtures was also observed. The degree to which the lamellar phase is formed was found to be strongly dependent on the co-flowing gas conditions used to stabilize the LCP stream. A combination of laboratory-based optical polarization microscopy and simulation studies was used to investigate these observations.