Search NASA⌕ Search

SEARCH · Search NASA

Results for “Domain knowledge”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 505 records · Page 28

Retrieving and Ranking Unannotated Images through Collaboratively Mining Online Search Results

We present a new image search and ranking algorithm for retrieving unannotated images by collaboratively mining online search results, which consist of online image and text search results. The online image search results are leveraged as reference examples to perform content-based image search over unannotated images. The online text search results are utilized to estimate individual reference images relevance to the search query as not all the online image search results are closely related to the query. Overall, the key contribution of our method lies in its capability to deal with unreliable online image search results through jointly mining visual and textual aspects of online search results. Through such collaborative mining, our algorithm infers the relevance of an online search result image to a text query. Once we estimate a query relevance score for each online image search result, we can selectively use query specific online search result images as reference examples for retrieving and ranking unannotated images. To explore the performance of our algorithm, we tested our algorithm both on the standard public image datasets and several modest sized personal photo collections. We also compared the performance of our method with that of two peer methods. The results are very positive, indicating that our algorithm is superior to existing content-based image search algorithms for retrieving and ranking unannotated images. Overall, the main advantage of our algorithm comes from its collaborative mining over online search results both in the visual and textual domains.

99 GENERAL AND MISCELLANEOUS↗

Regulating Expression of Cell and Tissue-Specific Genes by Modifying Transcription

Transcriptional regulation is the primary step to control gene expression, therefore function. Such regulation is achieved primarily via a combination of the activities of the promoter cis regulatory DNA elements and trans regulatory proteins that function through binding to these DNA elements. Rice bZIP transcription factors RF2a, RF2b and RLP1 play key roles in regulating the activity of a vascular tissue specific promoter isolated from Rice Tungro Bacilliform Virus (RTBV), through their interactions with the Box II essential cis element located in the promoter (Dai et al., 2006., Dai et al., 2004., Yin et al., 1997). RF2a, RF2b and RLP1 possess multiple regulatory domains. Functional characterization reveals that those domains can activate or repress the activity of the RTBV promoter. It is equally as important to recognize that these proteins control plant development by regulating differentiation and/or function of the vascular tissues. Studies of transcriptional regulation of the RTBV promoter by this group of bZIP proteins will not only provide insights about gene expression in the vascular tissue, but also insights about general mechanisms of transcription activation and repression. The knowledge gained from this research will also enable us to develop a well-described set of tools that can be used to control expression of multiple genes in transgenic plants. We have proposed characterize the function domains of RF2a, RF2b and RLP1 and explore the biological function of the transcription repressor RLP1.

59 BASIC BIOLOGICAL SCIENCES↗

Thermal Safety Analyses for the Production of Plutonium-238 at the High Flux Isotope Reactor

There has been a considerable effort over the previous few years to demonstrate and optimize the production of plutonium-238 ( 238 Pu) at the High Flux Isotope Reactor (HFIR). This effort has involved resources from multiple divisions and facilities at the Oak Ridge National Laboratory (ORNL) to demonstrate the fabrication, irradiation, and chemical processing of targets containing neptunium-237 ( 237 Np) dioxide (NpO 2 )/aluminum (Al) cermet pellets. A critical preliminary step to irradiation at the HFIR is to demonstrate the safety of the target under irradiation via documented experiment safety analyses. The steady-state thermal safety analyses of the target are simulated in a finite element model with the COMSOL Multiphysics code that determines, among other crucial parameters, the limiting maximum temperature in the target. Safety analysis efforts for this model discussed in the present report include: (1) initial modeling of single and reduced-length pellet capsules in order to generate an experimental knowledge base that incorporate initial non-linear contact heat transfer and fission gas equations, (2) modeling efforts for prototypical designs of partially loaded and fully loaded targets using limited available knowledge of fabrication and irradiation characteristics, and (3) the most recent and comprehensive modeling effort of a fully coupled thermo-mechanical approach over the entire fully loaded target domain incorporating burn-up dependent irradiation behavior and measured target and pellet properties, hereafter referred to as the production model. These models are used to conservatively determine several important steady-state parameters including target stresses and temperatures, the limiting condition of which is the maximum temperature with respect to the melting point. The single pellet model results provide a basis for the safety of the irradiations, followed by parametric analyses in the initial prototypical designs that were necessary due to the limiting fabrication and irradiation data available. The calculated parameters in the final production target model are the most accurate and comprehensive, while still conservative. Over 210 permutations in irradiation time and position were evaluated, and are supported by the most recent inputs and highest fidelity methodology. The results of these analyses show that the models presented in this report provide a robust and reliable basis for previous, current and future experiment safety analyses. In addition, they reveal an evolving knowledge of the steady-state behavior of the NpO 2 /Al pellets under irradiation for a variety of target encapsulations and potential conditions.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Combined blockade of Tim-3 and MEK inhibitor enhances the efficacy against melanoma

Insights into the role of the mitogen-activated protein kinase (MAPK) pathway and immune checkpoints have led combined targeted therapy and immunotherapy to be a promising regimen. Trametinib, as a mitogen-activated extracellular signal-regulated kinase (MEK) inhibitor, has demonstrated effectiveness in patients with advanced melanoma. T cell immunoglobulin- and mucin-domain-containing molecule-3 (Tim-3), an immune checkpoint molecule, participates in multiple negative regulation of antitumor immunity. We for the first time to our knowledge reported the combination of trametinib and anti-Tim-3 monoclonal antibody (mAb) in treating B16-F10 melanoma mice. We discovered that trametinib remarkably promoted apoptosis and inhibited cell proliferation while inhibition of MEK improved the expression of Tim-3 and caused the decrease of CD8{sup +} T cells; to the contrary, anti-Tim-3 mAb enhanced antitumor immunity by stimulating CD8{sup +} T cells, thus the combined therapy produced potent antitumor effect cooperatively. Taken together, our study provides compelling evidence for combining trametinib and anti-Tim-3 mAb as a potential valuable regimen in treating melanoma. - Highlights: • The potent antitumor efficacy of combined trametinib and anti-Tim-3 mAb. • MEK inhibitor involves in up-regulation of Tim-3. • The combined medication provides new insights into melanoma treatment.

60 APPLIED LIFE SCIENCES↗

32 examples of LLM applications in materials science and chemistry: towards automation, assistants, agents, and accelerated scientific discovery

Abstract Large language models (LLMs) are reshaping many aspects of materials science and chemistry research, enabling advances in molecular property prediction, materials design, scientific automation, knowledge extraction, and more. Recent developments demonstrate that the latest class of models are able to integrate structured and unstructured data, assist in hypothesis generation, and streamline research workflows. To explore the frontier of LLM capabilities across the research lifecycle, we review applications of LLMs through 32 total projects developed during the second annual LLM hackathon for applications in materials science and chemistry, a global hybrid event. These projects spanned seven key research areas: (1) molecular and material property prediction, (2) molecular and material design, (3) automation and novel interfaces, (4) scientific communication and education, (5) research data management and automation, (6) hypothesis generation and evaluation, and (7) knowledge extraction and reasoning from the scientific literature. Collectively, these applications illustrate how LLMs serve as versatile predictive models, platforms for rapid prototyping of domain-specific tools, and much more. In particular, improvements in both open source and proprietary LLM performance through the addition of reasoning, additional training data, and new techniques have expanded effectiveness, particularly in low-data environments and interdisciplinary research. As LLMs continue to improve, their integration into scientific workflows presents both new opportunities and new challenges, requiring ongoing exploration, continued refinement, and further research to address reliability, interpretability, and reproducibility.

Computer Science↗

Resonant seismic emission of subsurface objects

Numerical modeling results and field data indicate that some contrasting subsurface objects (such as tunnels, caves, pipes, filled pits, and fluid-filled fractures) are capable of generating durable resonant oscillations after trapping seismic energy. These oscillations consist of surface types of circumferential waves that repeatedly propagate around the object. The resonant emission of such trapped energy occurs primarily in the form of shear body waves that can be detected by remotely placed receivers. Resonant emission reveals itself in the form of sharp resonant peaks for the late parts of the records, when all strong direct and primary reflected waves are gone. These peaks were observed in field data for a buried barrel filled with water, in 2D finite-difference modeling results, and in the exact canonical solution for a fluid-filled sphere. A computed animation for the diffraction of a plane wave upon a low-velocity elastic sphere confirms the generation of resonances by durable surface waves. Resonant emission has characteristic quasi-hyperbolic traveltime patterns on shot gathers. The inversion of these patterns can be performed in the frequency domain after muting the strong direct and primary scattered waves. Subsurface objects can be detected and imaged at a single resonance frequency without an accurate knowledge of source trigger time. The imaging of subsurface objects requires information about the shear velocity distribution in an embedding medium, which can be done interactively during inversion.

54 ENVIRONMENTAL SCIENCES↗

Chloroplast Dynamics and Photosynthetic Efficiency: Final Technical Report

This project investigated the mechanism by which chloroplasts position themselves to maximize solar energy utilization, to enhance gas exchange, to minimize environmental stress, and to promote efficient exchange of metabolites with other compartments within the plant cell. Chloroplasts move within leaf cells to optimize light levels, moving toward levels of light useful for photosynthesis while moving away from excess light. Plastids sometimes extend their reach by sending out projections (stromules) that can connect anchor chloroplasts in position within the cell or provide close contacts with plasma membrane, mitochondria, peroxisomes, endoplasmic reticulum, and the nucleus. The intracellular location of chloroplasts in relation to other organelles with which they share biosynthetic pathways, such as peroxisomes and mitochondria in photorespiration, affects metabolite flow. This work contributed to the knowledge of the mechanisms of organelle movement and anchoring in specific locations in plant cells and how proteins traffic within the cell. We identified two domains on 12 of the 13 Arabidopsis myosins that were similar to the vacuole-binding (V) domain characterized in yeast and to the DIL domain characterized in yeast and mouse as required for secretory vesicle or melanosome movement, respectively. Because all of the Arabidopsis regions with homology to the V domain contain the amino acid sequence PAL, we refer to this region as the Arabidopsis PAL domain. We have used the yeast Myo2p tail structural information to model the 12 myosin XI tail domains containing the homologous PAL and DIL domains. Eight YFP::DIL domain fusions labeled peroxisomes; none labeled mitochondria or chloroplasts. Six myosin XI Vacuole domains labeled mitochondria and seven labeled Golgi bodies. The Arabidopsis myosin XI-F PAL domain and the homologous myosin XI-F PAL domain from N. benthamiana labels chloroplasts and stromules in N. benthamiana leaves. Using an Arabidopsis line containing hotoconvertible GFP, we observed transfer of protein from one plastid to another and within a stromule from single plastids. We provided time-lapse movies demonstrating movement of both the photoconvertible GFP and standard GFP between plastids. We previously demonstrated the lack of a plastid network within plant cells. We provided protocols explaining how to use fluorescent protein technology to track plastids and stromules within plant cells. We demonstrated that standard GFP unexpectedly could be photoconverted to a red form under certain conditions, allowing the use of GFP lines for studies that require photoconversion.

59 BASIC BIOLOGICAL SCIENCES↗

Oklahoma Mesonet Soil Moisture (OKMSOIL) Value-Added Product Report

Land surface and subsurface state characteristics (e.g., soil moisture) are critical measurements for ongoing weather and climate analyses of land-atmospheric interactions and processes. Knowledge of the variation in soil moisture across the U.S. Department of Energy Atmospheric Radiation Measurement (ARM) user facility’s Southern Great Plains (SGP) observatory domain is anticipated as a key quantity for assessing the potential impact of these interactions. This Oklahoma Mesonet Soil Moisture (OKMSOIL) value-added product (VAP) has been developed in response to growing demand for delivering a long-term measurement record of soil moisture (over an extended area) for such land-atmospheric studies. The Oklahoma Mesonet (OKM) started as a joint project between Oklahoma State University and the University of Oklahoma (Brock 1995). The OKM currently operates over 100 surface Mesonet stations across the state of Oklahoma, with an operational data record that extends into the early 1990s. This dense network of atmospheric and surface state instrumentation enables representation of land surface/subsurface conditions over this region and extended record. Algorithms introduced by previous Oklahoma Mesonet activities (e.g., Scott et al. 2013) are implemented in this documented OKMSOIL VAP to derive a key soil moisture quantity, the volumetric water content (volume of water per volume of soil in units, cm 3 /cm 3 ), from measurements of soil matric potential. These algorithms use a soil hydraulic parameter database (Meso-Soil Database) provided by the Oklahoma Mesonet scientists to estimate this quantity and several additional Mesonet station/sample details.

54 ENVIRONMENTAL SCIENCES↗

Oklahoma Mesonet Soil Moisture (OKMSOIL) Value-Added Product Report

Land surface and subsurface state characteristics (e.g., soil moisture) are critical measurements for ongoing weather and climate analyses of land-atmospheric interactions and processes. Knowledge of the variation in soil moisture across the U.S. Department of Energy Atmospheric Radiation Measurement (ARM) user facility’s Southern Great Plains (SGP) observatory domain is anticipated as a key quantity for assessing the potential impact of these interactions. This Oklahoma Mesonet Soil Moisture (OKMSOIL) value-added product (VAP) has been developed in response to growing demand for delivering a long-term measurement record of soil moisture (over an extended area) for such land-atmospheric studies. The Oklahoma Mesonet (OKM) started as a joint project between Oklahoma State University and the University of Oklahoma (Brock 1995). The OKM currently operates over 100 surface Mesonet stations across the state of Oklahoma, with an operational data record that extends into the early 1990s. This dense network of atmospheric and surface state instrumentation enables representation of land surface/subsurface conditions over this region and extended record. Algorithms introduced by previous Oklahoma Mesonet activities (e.g., Scott et al. 2013) are implemented in this documented OKMSOIL VAP to derive a key soil moisture quantity, the volumetric water content (volume of water per volume of soil in units, cm 3 /cm 3 ), from measurements of soil matric potential. These algorithms use a soil hydraulic parameter database (Meso-Soil Database) provided by the Oklahoma Mesonet scientists to estimate this quantity and several additional Mesonet station/sample details.

58 GEOSCIENCES↗

DHS Internship Summary-Crystal Assembly at Different Length Scales

I was part of a project in which in situ atomic force microscopy (AFM) was used to monitor growth and dissolution of atomic and colloidal crystals. At both length scales, the chemical environment of the system greatly altered crystal growth and dissolution. Calcium phosphate was used as a model system for atomic crystals. A dissolution-reprecipitation reaction was observed in this first system, involving the conversion of brushite (DCPD) to octacalcium phosphate (OCP). In the second system, polymeric colloidal crystals were dissolved in an ionic solvent, revealing the underlying structure of the crystal. The dissolved crystal was then regrown through an evaporative step method. Recently, we have also found that colloids can be reversibly deposited in situ onto an ITO (indium tin oxide) substrate via an electrochemistry setup. The overall goal of this project was to develop an understanding of the mechanisms that control crystallization and order, so that these might be controlled during material synthesis. Controlled assembly of materials over a range of length scales from molecules to nanoparticles to colloids is critical for designing new materials. In particular, developing materials for sensor applications with tailorable properties and long range order is important. In this work, we examine two of these length scales: small molecule crystallization of calcium phosphate (whose crystal phases include DCPD, OCP, and HAP) and colloidal crystallization of Poly(methyl methacrylate) beads. Atomic Force Microscopy is ideal for this line of work because it allows for the possibility of observing non-conducting samples in fluid during growth with high resolution ({approx} 10 nm). In fact, during atomic crystal growth one can observe changes in atomic steps, and with colloidal crystals, one can monitor the individual building blocks of the crystal. Colloids and atoms crystallize under the influence of different forces acting at different length scales as seen in Table 1. In particular, molecular crystals, which are typically dominated by ionic and covalent bonding, are an order of magnitude more strongly bonded than colloidal crystals. In molecular crystals, ordering is driven by the interaction potentials between molecules. By contrast, colloidal assembly is a competition between the repulsive electrostatic forces that prevent aggregation in solution (due to surface charge), and short-range van der Waals and entropic forces that leads to ordering. Understanding atomic crystallization is fundamentally important for fabrication of tailorable crystalline materials, for example for biological or chemical sensors. The transformation of brushite to OCP not only serves as a model system for atomic crystal growth (applicable to many other crystal growth processes), but is also important in bone cements. Colloidal crystals have unique optical properties which respond to chemical and mechanical stimuli, making them very important for sensing applications. The mechanism of colloidal crystal assembly is thus fundamentally important. Our in situ dissolution and regrowth experiments are one good method of analyzing how these crystals pack under different conditions and how defect sites are formed and filled. In these experiments, a silica additive was used to strengthen the colloidal crystal during initial assembly (ex situ) and to increase domain size and long range order. Reversible electrodeposition of colloids onto a conductive substrate (ITO in our case) is another system which can further our knowledge of colloidal assembly. This experiment holds promise of allowing in situ observation of colloidal crystal growth and the influence of certain additives on crystal order. The ultimate goal would be to achieve long range order in these crystals by changing the surface charge or the growth environment.

36 MATERIALS SCIENCE↗

Exploring Causal Physical Mechanisms via Non-Gaussian Linear Models and Deep Kernel Learning: Applications for Ferroelectric Domain Structures

Rapid emergence of multimodal imaging in scanning probe, electron, and optical microscopies has brought forth the challenge of understanding the information contained in these complex data sets, targeting the intrinsic correlations between different channels, and further exploring the underpinning causal physical mechanisms. Here, we develop such an analysis framework for Piezoresponse Force Microscopy. We argue that under certain conditions, we can bootstrap experimental observations with the prior knowledge of materials structure to get information on certain nonobserved properties, and demonstrate linear causal analysis for PFM observables. We further demonstrate that the strength of individual causal links between complex descriptors can be ascertained using the deep kernel learning (DKL) model. In this DKL analysis, we use the prior information on domain structure within the image to predict the physical properties. This analysis demonstrates the correlative relationships between morphology, piezoresponse, elastic property, etc., at nanoscale. The prediction of morphology and other physical parameters illustrates a mutual interaction between surface condition and physical properties in ferroelectric materials. Overall, this analysis is universal and can be extended to explore the correlative relationships of other multichannel data sets, and allow for high-fidelity reconstruction of underpinning functionalities and physical mechanisms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Interconnected ionic domains enhance conductivity in microphase separated block copolymer electrolytes

Block copolymer electrolytes (BCEs) represent an attractive choice as solid-state ionic conductors for electrochemical technologies used in energy storage and conversion, water treatment, sensors, and data storage and processing. Unlocking the maximum ionic conductivity of BCEs requires an intimate understanding as to how the microphase separated structure influences transport properties. However, elucidating such knowledge remains elusive due to the challenging task of precisely engineering BCEs with a defined structure in bulk materials. In this work, we examined BCEs in a thin film format because it was amenable to attaining BCEs with a desired nanostructure. Specifically, we systematically investigated anion-conducting BCEs with different degrees of connectivity of the ionic domains. For the first time, we demonstrate that increasing terminal defects in the ionic domain from 1 terminal defect per mu m(2) to 20 terminal defects per mu m(2) ( a relatively small amount of defects) decreased ionic conductivity by 67% compared to the maximum value attained. Conversely, maximizing ionic domain connectivity increased the ionic conductivity by two-fold over a non-ordered BCE film. These experiments highlight that microphase separation alone was insufficient for ameliorating ionic conductivity in BCEs. Rather, microphase separation coupled with complete ionic domain connectivity realized BCEs with significantly enhanced ionic conductivity.

Arges, Christopher G. (ORCID:0000000317038323)↗

Applications of explainable artificial intelligence in renewable energy research

Researchers in renewable energy are applying deep learning (DL) to a variety of problems from diverse renewable energy domains, such as biofuels, wind, solar, power systems, buildings, vehicles, and transportation systems. Improvements in accuracy may be demonstrated using DL in laboratory settings. However, the lack of interpretability of DL models poses a practical limitation to their utility in advancing scientific knowledge and in the deployment of DL models in safety-critical energy systems. In this article, we discuss explainable artificial intelligence (XAI) as one pathway toward more interpretable DL models. We explore a brief timeline of U.S. national laboratory interest in XAI, an overview and taxonomy of methods in the field of XAI, and a selection of applications across renewable energy research domains. We conclude by highlighting pivotal areas where XAI can accelerate innovation in artificial intelligence for renewable energy research and other essential future directions.

97 MATHEMATICS AND COMPUTING↗

Conforming Morse-Smale Complexes

Morse-Smale (MS) complexes have been gaining popularity as a tool for feature-driven data analysis and visualization. However, the quality of their geometric embedding and the sole dependence on the input scalar field data can limit their applicability when expressing application-dependent features. In this paper we introduce a new combinatorial technique to compute an MS complex that conforms to both an input scalar field and an additional, prior segmentation of the domain. The segmentation constrains the MS complex computation guaranteeing that boundaries in the segmentation are captured as separatrices of the MS complex. We demonstrate the utility and versatility of our approach with two applications. First, we use streamline integration to determine numerically computed basins/mountains and use the resulting segmentation as an input to our algorithm. This strategy enables the incorporation of prior flow path knowledge, effectively resulting in an MS complex that is as geometrically accurate as the employed numerical integration. Our second use case is motivated by the observation that often the data itself does not explicitly contain features known to be present by a domain expert. We introduce edit operations for MS complexes so that a user can directly modify their features while maintaining all the advantages of a robust topology-based representation.

97 MATHEMATICS AND COMPUTING↗

Model Assumptions and Data Characteristics: Impacts on Domain Adaptation in Building Segmentation

Studies on domain adaptation (DA) for remote sensing (RS) imagery analysis lack consistency in selection and description of evaluation scenarios. Without properly characterizing datasets, model assumptions, and evaluation scenarios, it is difficult to objectively compare DA methods and reach conclusions about their suitability across different applications. With this motivation, this work seeks to empirically assess to which extent the interaction between data characteristics and model assumptions influences the effectiveness of DA methods. Using the widely explored task of building footprint segmentation as a case study, we perform a large-scale study across over 200 DA scenarios that include variations across view angles, areas observed, and sensors used for data acquisition. Rather than adopting different model architectures or optimization criteria, we contrast the performances of two DA methods based on adversarial learning that differ only in their assumptions about source and target domains. Informed by metadata and data characteristics unveiled using traditional computer vision (CV) techniques as well as pretrained deep models, we provide a detailed meta-analysis of experiments highlighting the importance of accurately considering data assumptions for DA in RS segmentation tasks. As demonstrated by a “cherry-picking” exercise, different claims regarding which model is best could be made by selecting different subsets of evaluation scenarios. While well-calibrated assumptions can be beneficial, mismatching assumptions can lead to negative biases in DA applications. Furthermore, this study intends to motivate the community toward more consistent evaluation protocols while providing recommendations and insights toward creating novel benchmark datasets, documenting data characteristics, application-specific knowledge, and model assumptions.

42 ENGINEERING↗

RCSB protein data Bank: Next‐generation advanced search for exploration of experimental structures and computed structure models

Abstract The Protein Data Bank (PDB), established in 1971, is the primary global, open‐access archive for experimentally determined 3D macromolecular structures (proteins, RNA, DNA). The research‐focused RCSB.org web‐portal provides access to these data alongside more than one million machine‐learning‐predicted structure models, greatly expanding the available structural landscape. Rapid growth of both experimental and computational structures has increased the need for powerful yet accessible search tools that serve a broad and diverse scientific community. Herein, we describe a redesigned RCSB Protein Data Bank RCSB.org Advanced Search capability that supports intuitive discovery of 3D structures through a unified interface. This interface integrates annotation‐, sequence‐, and 3D structure‐based searches, embeds an interactive 3D viewer, and incorporates curated biological knowledge, such as catalytic site definitions from Mechanism and Catalytic Site Atlas and ligand‐guided structural motifs, for constructing geometry‐driven queries. A new Chemical Search tool allows definition of chemical queries via an integrated drawing tool or standard identifiers, seamlessly combining them with annotation filters. By allowing query definition directly within spatial and chemical contexts, these search interfaces reduce the need for detailed knowledge of residue numbering, chain identifiers, or external cheminformatics software. This capability enables efficient exploration of structures, chemical diversity, and structure–function relationships across all life domains. The redesigned interfaces can be accessed directly at rcsb.org/search/advanced for Advanced Search and rcsb.org/search/chemical for Chemical Search.

Rose, Yana [Research Collaboratory for Structural ↗

A Study on Efficient Reinforcement Learning Through Knowledge Transfer

Although Reinforcement Learning (RL) algorithms have made impressive progress in learning complex tasks over the past years, there are still prevailing short-comings and challenges. Specifically, the sample-inefficiency and limited adaptation across tasks often make classic RL techniques impractical for real-world applications despite the gained representational power when combining deep neural networks with RL, known as Deep Reinforcement Learning (DRL). Recently, a number of approaches to address those issues have emerged. Many of those solutions are based on smart DRL architectures that enhance single task algorithms with the capability to share knowledge between agents and across tasks by introducing Transfer Learning (TL) capabilities. Here this survey addresses strategies of knowledge transfer from simple parameter sharing to privacy preserving federated learning and aims at providing a general overview of the field of TL in the DRL domain, establishes a classification framework, and briefly describes representative works in the area.

97 MATHEMATICS AND COMPUTING↗

Polycomb-like proteins link the PRC2 complex to CpG islands

The Polycomb repressive complex 2 (PRC2) mainly mediates transcriptional repression1,2 and has essential roles in various biological processes including the maintenance of cell identity and proper differentiation. Polycomb-like (PCL) proteins, such as PHF1, MTF2 and PHF19, are PRC2-associated factors that form sub-complexes with PRC2 core components3, and have been proposed to modulate the enzymatic activity of PRC2 or the recruitment of PRC2 to specific genomic loci4,5,6,7,8,9,10,11,12,13. Mammalian PRC2-binding sites are enriched in CG content, which correlates with CpG islands that display a low level of DNA methylation14. However, the mechanism of PRC2 recruitment to CpG islands is not fully understood. Here we solve the crystal structures of the N-terminal domains of PHF1 and MTF2 with bound CpG-containing DNAs in the presence of H3K36me3-containing histone peptides. We show that the extended homologous regions of both proteins fold into a winged-helix structure, which specifically binds to the unmethylated CpG motif but in a completely different manner from the canonical winged-helix DNA recognition motif. We also show that the PCL extended homologous domains are required for efficient recruitment of PRC2 to CpG island-containing promoters in mouse embryonic stem cells. Our research provides the first, to our knowledge, direct evidence to demonstrate that PCL proteins are crucial for PRC2 recruitment to CpG islands, and further clarifies the roles of these proteins in transcriptional regulation in vivo.

58 GEOSCIENCES↗