Search

Clinical trials are crucial for drug development but are time consuming, expensive, and often burdensome on patients. More importantly, clinical trials face uncertain outcomes due to issues with efficacy, safety, or problems with patient recruitment. If we were better at predicting the results of clinical trials, we could avoid having to run trials that will inevitably fail more resources could be devoted to trials that are likely to succeed. In this paper, we propose Hierarchical INteraction Network (HINT) for more general, clinical trial outcome predictions for all diseases based on a comprehensive and diverse set of web data including molecule information of the drugs, target disease information, trial protocol and biomedical knowledge. HINT first encode these multi-modal data into latent embeddings, where an imputation module is designed to handle missing data. Next, these embeddings will be fed into the knowledge embedding module to generate knowledge embeddings that are pretrained using external knowledge on pharmaco-kinetic properties and trial risk from the web. Then the interaction graph module will connect all the embedding via domain knowledge to fully capture various trial components and their complex relations as well as their influences on trial outcomes. Finally, HINT learns a dynamic attentive graph neural network to predict trial outcome. Comprehensive experimental results show that HINT achieves strong predictive performance, obtaining 0.772, 0.607, 0.623, 0.703 on PR-AUC for Phase I, II, III, and indication outcome prediction, respectively. It also consistently outperforms the best baseline method by up to 12.4\% on PR-AUC.

The Period--Luminosity relation (PLR) of Mira variable stars is an important tool to determine astronomical distances. The common approach of estimating the PLR is a two-step procedure that first estimates the Mira periods and then runs a linear regression of magnitude on log period. When the light curves are sparse and noisy, the accuracy of period estimation decreases and can suffer from aliasing effects. Some methods improve accuracy by incorporating complex model structures at the expense of significant computational costs. Another drawback of existing methods is that they only provide point estimation without proper estimation of uncertainty. To overcome these challenges, we develop a hierarchical Bayesian model that simultaneously models the quasi-periodic variations for a collection of Mira light curves while estimating their common PLR. By borrowing strengths through the PLR, our method automatically reduces the aliasing effect, improves the accuracy of period estimation, and is capable of characterizing the estimation uncertainty. We develop a scalable stochastic variational inference algorithm for computation that can effectively deal with the multimodal posterior of period. The effectiveness of the proposed method is demonstrated through simulations, and an application to observations of Miras in the Local Group galaxy M33. Without using ad-hoc period correction tricks, our method achieves a distance estimate of M33 that is consistent with published work. Our method also shows superior robustness to downsampling of the light curves.

We present an empirical model for the number of globular clusters (GCs) in galaxies based on recent data showing a tight relationship between dark matter halo virial masses and GC numbers. While a simple base model forming GCs in low-mass haloes reproduces this relation, we show that a second formation pathway for GCs is needed to account for observed younger GC populations. We confirm previous works that reported the observed linear correlation as being a consequence of hierarchical merging and its insensitivity to the exact GC formation processes at higher virial masses, even for a dual formation scenario. We find that the scatter of the linear relation is strongly correlated with the relative amount of smooth accretion: the more dark matter is smoothly accreted, the fewer GCs a halo has compared to other haloes of the same mass. This scatter is smaller than that introduced by halo mass measurements, indicating that the number of GCs in a galaxy is a good tracer for its dark matter mass. Smooth accretion is also the reason for a lower average dark matter mass per GC in low-mass haloes. Finally, we successfully reproduce the observed general trend of GCs being old and the tendency of more massive haloes hosting older GC systems. Including the second GC formation mechanism through gas-rich mergers leads to a more realistic variety of GC age distributions and also introduces an age inversion in the halo virial mass range $\log M_\mathrm{vir}/\mathrm{M}_\odot = 11{-}13$., Comment: 18 pages, 19 figures, accepted for publication in MNRAS

We derive a distance of $D = 16.6 \pm 0.3$~Mpc ($\mu=31.10\pm0.04$~mag) to the archetypal narrow-line Seyfert 1 galaxy NGC 4051 based on Cepheid Period--Luminosity relations and new Hubble Space Telescope multiband imaging. We identify 419 Cepheid candidates and estimate the distance at both optical and near-infrared wavelengths using subsamples of precisely-photometered variables (123 and 47 in the optical and near-infrared subsamples, respectively). We compare our independent photometric procedures and distance-estimation methods to those used by the SH0ES team and find agreement to 0.01~mag. The distance we obtain suggests an Eddington ratio $\dot{m} \approx 0.2$ for NGC 4051, typical of narrow-line Seyfert 1 galaxies, unlike the seemingly-odd value implied by previous distance estimates. We derive a peculiar velocity of $-490\pm34$~km~s$^{-1}$ for NGC 4051, consistent with the overall motion of the Ursa Major Cluster in which it resides. We also revisit the energetics of the NGC 4051 nucleus, including its outflow and mass accretion rates., Comment: 15 pages, 12 figures, 6 tables, accepted for publication in ApJ

Accurate prediction of the transmission of epidemic diseases such as COVID-19 is crucial for implementing effective mitigation measures. In this work, we develop a tensor method to predict the evolution of epidemic trends for many regions simultaneously. We construct a 3-way spatio-temporal tensor (location, attribute, time) of case counts and propose a nonnegative tensor factorization with latent epidemiological model regularization named STELAR. Unlike standard tensor factorization methods which cannot predict slabs ahead, STELAR enables long-term prediction by incorporating latent temporal regularization through a system of discrete-time difference equations of a widely adopted epidemiological model. We use latent instead of location/attribute-level epidemiological dynamics to capture common epidemic profile sub-types and improve collaborative learning and prediction. We conduct experiments using both county- and state-level COVID-19 data and show that our model can identify interesting latent patterns of the epidemic. Finally, we evaluate the predictive ability of our method and show superior performance compared to the baselines, achieving up to 21% lower root mean square error and 25% lower mean absolute error for county-level prediction., Comment: AAAI 2021

Graphite, with many industrial applications, is one of the widely sought-after allotropes of carbon. The sp2 hybridized and thermodynamically stable form of carbon forms a layered structure with strong in-plane carbon bonds and weak inter-layer van der Waals bonding. Graphite is also a high-temperature ceramic, and shaping them into complex geometries is challenging, given its limited sintering behavior even at high temperatures. Although the geometric design of the graphite structure in many of the applications could dictate its precision performance, conventional synthesis methods for formulating complex geometric graphite shapes are limited due to the intrinsic brittleness and difficulties of high-temperature processing. Here, we report the development of colloidal graphite ink from commercial graphite powders with reproducible rheological behavior that allows the fabrication of any complex architectures with tunable geometry and directionality via 3D printing at room temperature. The method is enabled via using small amounts of clay, another layered material, as an additive, allowing the proper design of the graphene ink and subsequent binding of graphite platelets during printing. Sheared layers of clay are easily able to flow, adapt, and interface with graphite layers forming strong binding between the layers and between particles that make the larger structures. The direct ink printing of complex 3D architectures of graphite without further heat treatments could lead to easy shape engineering and related applications of graphite at various length scales, including complex graphite molds or crucibles. The 3D printed complex graphitic structures exhibit excellent thermal, electrical, and mechanical properties, and the clay additive does not seem to alter these properties due to the excellent inter-layer dispersion and mixing within the graphite material., Comment: 19 pages, 4 figures

To test the possibility of differentiating chest x-ray images of COVID-19 against other pneumonia and healthy patients using deep neural networks. We construct the X-ray imaging data from two publicly available sources, which include 5508 chest x-ray images across 2874 patients with four classes: normal, bacterial pneumonia, non-COVID-19 viral pneumonia, and COVID-19. To identify COVID-19, we propose a Focal Loss Based Neural Ensemble Network (FLANNEL), a flexible module to ensemble several convolutional neural network (CNN) models and fuse with a focal loss for accurate COVID-19 detection on class imbalance data. FLANNEL consistently outperforms baseline models on COVID-19 identification task in all metrics. Compared with the best baseline, FLANNEL shows a higher macro-F1 score with 6% relative increase on Covid-19 identification task where it achieves 0.7833(0.07) in Precision, 0.8609(0.03) in Recall, and 0.8168(0.03) F1 score.

Successful health risk prediction demands accuracy and reliability of the model. Existing predictive models mainly depend on mining electronic health records (EHR) with advanced deep learning techniques to improve model accuracy. However, they all ignore the importance of publicly available online health data, especially socioeconomic status, environmental factors, and detailed demographic information for each location, which are all strong predictive signals and can definitely augment precision medicine. To achieve model reliability, the model needs to provide accurate prediction and uncertainty score of the prediction. However, existing uncertainty estimation approaches often failed in handling high-dimensional data, which are present in multi-sourced data. To fill the gap, we propose UNcertaInTy-based hEalth risk prediction (UNITE) model. Building upon an adaptive multimodal deep kernel and a stochastic variational inference module, UNITE provides accurate disease risk prediction and uncertainty estimation leveraging multi-sourced health data including EHR data, patient demographics, and public health data collected from the web. We evaluate UNITE on real-world disease risk prediction tasks: nonalcoholic fatty liver disease (NASH) and Alzheimer's disease (AD). UNITE achieves up to 0.841 in F1 score for AD detection, up to 0.609 in PR-AUC for NASH detection, and outperforms various state-of-the-art baselines by up to $19\%$ over the best baseline. We also show UNITE can model meaningful uncertainties and can provide evidence-based clinical support by clustering similar patients.

Experimental evidence for ferromagnetic ordering in isotropic atomically thin two-dimensional crystals has been missing until a bilayer Cr2Ge2Te6, and a three-atom thick monolayer CrI3 are shown to retain ferromagnetic ordering at finite temperatures. Here, we demonstrate successful isolation of a non-van der Waals type ultra-thin nanosheet of FeS2 derived from naturally occurring pyrite mineral (FeS2) by means of liquid-phase exfoliation. Structural characterizations imply that (111) oriented sheets are predominant and is supported theoretically by means of density functional theory surface energy calculations. Spin-polarized density theory calculations further predicted that (111) oriented three-atom thick pyrite sheet has a stable ferromagnetic ground state different from its diamagnetic bulk counterpart. This theoretical finding is evaluated experimentally employing low temperature superconducting quantum interference device measurements and observed an anomalous ferromagnetic kind of behavior., Comment: 21 pages, 5 figures

The efficacy of a drug depends on its binding affinity to the therapeutic target and pharmacokinetics. Deep learning (DL) has demonstrated remarkable progress in predicting drug efficacy. We develop MolDesigner, a human-in-the-loop web user-interface (UI), to assist drug developers leverage DL predictions to design more effective drugs. A developer can draw a drug molecule in the interface. In the backend, more than 17 state-of-the-art DL models generate predictions on important indices that are crucial for a drug's efficacy. Based on these predictions, drug developers can edit the drug molecule and reiterate until satisfaction. MolDesigner can make predictions in real-time with a latency of less than a second., Comment: NeurIPS 2020 Demonstration Track

Molecule optimization is a fundamental task for accelerating drug discovery, with the goal of generating new valid molecules that maximize multiple drug properties while maintaining similarity to the input molecule. Existing generative models and reinforcement learning approaches made initial success, but still face difficulties in simultaneously optimizing multiple drug properties. To address such challenges, we propose the MultI-constraint MOlecule SAmpling (MIMOSA) approach, a sampling framework to use input molecule as an initial guess and sample molecules from the target distribution. MIMOSA first pretrains two property agnostic graph neural networks (GNNs) for molecule topology and substructure-type prediction, where a substructure can be either atom or single ring. For each iteration, MIMOSA uses the GNNs' prediction and employs three basic substructure operations (add, replace, delete) to generate new molecules and associated weights. The weights can encode multiple constraints including similarity and drug property constraints, upon which we select promising molecules for next iteration. MIMOSA enables flexible encoding of multiple property- and similarity-constraints and can efficiently generate new molecules that satisfy various property constraints and achieved up to 49.6% relative improvement over the best baseline in terms of success rate. The code repository (including readme file, data preprocessing and model construction, evaluation) is available https://github.com/futianfan/MIMOSA., Comment: Accepted by AAAI 2021

Thanks to the increasing availability of drug-drug interactions (DDI) datasets and large biomedical knowledge graphs (KGs), accurate detection of adverse DDI using machine learning models becomes possible. However, it remains largely an open problem how to effectively utilize large and noisy biomedical KG for DDI detection. Due to its sheer size and amount of noise in KGs, it is often less beneficial to directly integrate KGs with other smaller but higher quality data (e.g., experimental data). Most of the existing approaches ignore KGs altogether. Some try to directly integrate KGs with other data via graph neural networks with limited success. Furthermore, most previous works focus on binary DDI prediction whereas the multi-typed DDI pharmacological effect prediction is a more meaningful but harder task. To fill the gaps, we propose a new method SumGNN: knowledge summarization graph neural network, which is enabled by a subgraph extraction module that can efficiently anchor on relevant subgraphs from a KG, a self-attention based subgraph summarization scheme to generate a reasoning path within the subgraph, and a multi-channel knowledge and data integration module that utilizes massive external biomedical knowledge for significantly improved multi-typed DDI predictions. SumGNN outperforms the best baseline by up to 5.54\%, and the performance gain is particularly significant in low data relation types. In addition, SumGNN provides interpretable prediction via the generated reasoning paths for each prediction., Comment: Published in Bioinformatics 2021

Gravitational waves from neutron star mergers have long been considered a promising way to measure the Hubble constant, $H_0$, which describes the local expansion rate of the universe. While black hole mergers are more abundantly observed, their expected lack of electromagnetic emission and poor gravitational-wave localization make them less well suited for measuring $H_0$. Black hole mergers within the disks of Active Galactic Nuclei (AGN) could be an exception. Accretion from the AGN disk may produce an electromagnetic signal, pointing observers to the host galaxy. Alternatively, the low number density of AGNs could help identify the host galaxy of $1-5\%$ of mergers. Here we show that black hole mergers in AGN disks may be a sensitive way to determine $H_0$ with gravitational waves. If $1\%$ ($10\%$) of LIGO's observations occur in AGN disks with identified host galaxies, we could measure $H_0$ with $12\%$ ($4\%$) uncertainty in five years, possibly comparable to the sensitivity of neutron star mergers and set to considerably improve current gravitational wave measurements., Comment: 5 pages, 2 figures

LTL synthesis is the problem of synthesizing a reactive system from a formal specification in Linear Temporal Logic. The extension of allowing for partial observability, where the system does not have direct access to all relevant information about the environment, allows generalizing this problem to a wider set of real-world applications, but the difficulty of implementing such an extension in practice means that it has remained in the realm of theory. Recently, it has been demonstrated that restricting LTL synthesis to systems with finite executions by using LTL with finite-horizon semantics (LTLf) allows for significantly simpler implementations in practice. With the conceptual simplicity of LTLf, it becomes possible to explore extensions such as partial observability in practice for the first time. Previous work has analyzed the problem of LTLf synthesis under partial observability theoretically and suggested two possible algorithms, one with 3EXPTIME and another with 2EXPTIME complexity. In this work, we first prove a complexity lower bound conjectured in earlier work. Then, we complement the theoretical analysis by showing how the two algorithms can be integrated in practice into an established framework for LTLf synthesis. We furthermore identify a third, MSO-based, approach enabled by this framework. Our experimental evaluation reveals very different results from what the theory seems to suggest, with the 3EXPTIME algorithm often outperforming the 2EXPTIME approach. Furthermore, as long as it is able to overcome an initial memory bottleneck, the MSO-based approach can often outperforms the others., Comment: In Proceedings GandALF 2020, arXiv:2009.09360

GTPases are molecular switches that regulate a wide range of cellular processes, such as organelle biogenesis, position, shape, and signal transduction. These enzymes operate by toggling between an active ("ON") guanosine triphosphate (GTP)-bound state and an inactive ("OFF") guanosine diphosphate (GDP)-bound state; such a toggle is regulated by GEFs (guanine nucleotide exchange factors) and GAPs (GTPase activating proteins). Here we dissect a network motif between monomeric (m) and trimeric (t) GTPases assembled exclusively in eukaryotic cells of multicellular organisms. To this end, we develop a system of ordinary differential equations in which these two classes of GTPases are interlinked conditional to their ON/OFF states within a motif through feedforward and feedback loops. We provide formulas for the steady states of the system and perform local stability analysis to investigate the role of the different connections between the GTPase switches. A feedforward from the active mGTPase to the GEF of the tGTPase was sufficient to provide two locally stable states: one where both active/inactive forms of the mGTPase can be interpreted as having low concentrations and the other where both m- and tGTPase have high concentrations. When a feedback loop from the GEF of the tGTPase to the GAP of the mGTPase was added to the feedforward system, two other locally stable states emerged, both having the tGTPase inactivated and being interpreted as having low active tGTPase concentrations. Finally, the addition of a second feedback loop, from the active tGTPase to the GAP of the mGTPase, gives rise to a family of steady states parametrized by the inactive tGTPase concentrations. Our findings reveal that the coupling of these two different GTPase motifs can dramatically change their steady state behaviors and shed light on how such coupling may impact information processing in eukaryotic cells., Comment: 7 figures

Temporal logic is often used to describe temporal properties in AI applications. The most popular language for doing so is Linear Temporal Logic (LTL). Recently, LTL on finite traces, LTLf, has been investigated in several contexts. In order to reason about LTLf, formulas are typically compiled into deterministic finite automata (DFA), as the intermediate semantic representation. Moreover, due to the fact that DFAs have canonical representation, efficient minimization algorithms can be applied to maximally reduce DFA size, helping to speed up subsequent computations. Here, we present a thorough investigation on two classical minimization algorithms, namely, the Hopcroft and Brzozowski algorithms. More specifically, we show how to apply these algorithms to semi-symbolic (explicit states, symbolic transition functions) automata representation. We then compare the two algorithms in the context of an LTLf-synthesis framework, starting from LTLf formulas. While earlier studies on comparing the two algorithms starting from randomly-generated automata concluded that neither algorithm dominates, our results suggest that starting from LTLf formulas, Hopcroft's algorithm is the best choice in the context of reactive synthesis. Deeper analysis explains why the supposed advantage of Brzozowski's algorithm does not materialize in practice., Comment: In Proceedings GandALF 2021, arXiv:2109.07798

Objective: The COVID-19 pandemic has created many challenges that need immediate attention. Various epidemiological and deep learning models have been developed to predict the COVID-19 outbreak, but all have limitations that affect the accuracy and robustness of the predictions. Our method aims at addressing these limitations and making earlier and more accurate pandemic outbreak predictions by (1) using patients' EHR data from different counties and states that encode local disease status and medical resource utilization condition; (2) considering demographic similarity and geographical proximity between locations; and (3) integrating pandemic transmission dynamics into deep learning models. Materials and Methods: We proposed a spatio-temporal attention network (STAN) for pandemic prediction. It uses an attention-based graph convolutional network to capture geographical and temporal trends and predict the number of cases for a fixed number of days into the future. We also designed a physical law-based loss term for enhancing long-term prediction. STAN was tested using both massive real-world patient data and open source COVID-19 statistics provided by Johns Hopkins university across all U.S. counties. Results: STAN outperforms epidemiological modeling methods such as SIR and SEIR and deep learning models on both long-term and short-term predictions, achieving up to 87% lower mean squared error compared to the best baseline prediction model. Conclusions: By using information from real-world patient data and geographical data, STAN can better capture the disease status and medical resource utilization information and thus provides more accurate pandemic modeling. With pandemic transmission law based regularization, STAN also achieves good long-term prediction performance.

We derive a distance of $15.8\pm0.4$ Mpc to the archetypical Seyfert 1 galaxy NGC 4151 based on the near-infrared Cepheid Period-Luminosity relation and new Hubble Space Telescope multiband imaging. This distance determination, based on measurements of 35 long-period ($P > 25$d) Cepheids, will support the absolute calibration of the supermassive black hole mass in this system, as well as studies of the dynamics of the feedback or feeding of its active galactic nucleus., Comment: 18 pages. 11 figures. 7 tables. Accepted for publication in ApJ

Clinical trials play important roles in drug development but often suffer from expensive, inaccurate and insufficient patient recruitment. The availability of massive electronic health records (EHR) data and trial eligibility criteria (EC) bring a new opportunity to data driven patient recruitment. One key task named patient-trial matching is to find qualified patients for clinical trials given structured EHR and unstructured EC text (both inclusion and exclusion criteria). How to match complex EC text with longitudinal patient EHRs? How to embed many-to-many relationships between patients and trials? How to explicitly handle the difference between inclusion and exclusion criteria? In this paper, we proposed CrOss-Modal PseudO-SiamEse network (COMPOSE) to address these challenges for patient-trial matching. One path of the network encodes EC using convolutional highway network. The other path processes EHR with multi-granularity memory network that encodes structured patient records into multiple levels based on medical ontology. Using the EC embedding as query, COMPOSE performs attentional record alignment and thus enables dynamic patient-trial matching. COMPOSE also introduces a composite loss term to maximize the similarity between patient records and inclusion criteria while minimize the similarity to the exclusion criteria. Experiment results show COMPOSE can reach 98.0% AUC on patient-criteria matching and 83.7% accuracy on patient-trial matching, which leads 24.3% improvement over the best baseline on real-world patient-trial matching tasks., Comment: Accepted by KDD'20

In this article we propose a compartmental model for the dynamics of Coronavirus Disease 2019 (COVID-19). We take into account the presence of asymptomatic infections and the main policies that have been adopted so far to contain the epidemic: isolation (or social distancing) of a portion of the population, quarantine for confirmed cases and testing. We model isolation by separating the population in two groups: one composed by key-workers that keep working during the pandemic and have a usual contact rate, and a second group consisting of people that are enforced/recommended to stay at home. We refer to quarantine as strict isolation, and it is applied to confirmed infected cases. In the proposed model, the proportion of people in isolation, the level of contact reduction and the testing rate are control parameters that can vary in time, representing policies that evolve in different stages. We obtain an explicit expression for the basic reproduction number $\mathcal{R}_0$ in terms of the parameters of the disease and of the control policies. In this way we can quantify the effect that isolation and testing have in the evolution of the epidemic. We present a series of simulations to illustrate different realistic scenarios. From the expression of $\mathcal{R}_0$ and the simulations we conclude that isolation (social distancing) and testing among asymptomatic cases are fundamental actions to control the epidemic, {and the stricter these measures are and the sooner they are implemented,} the more lives can be saved. Additionally, we show that people that remain in isolation significantly reduce their probability of contagion, so risk groups should be recommended to maintain a low contact rate during the course of the epidemic.

The accuracy of the Hubble constant measured with extragalactic Cepheids depends on robust photometry and background estimation in the presence of stellar crowding. The conventional approach accounts for crowding by sampling backgrounds near Cepheids and assuming they match those at their positions. We show a direct consequence of crowding by unresolved sources at Cepheid sites is a reduction in the fractional amplitudes of their light curves. We use a simple analytical expression to infer crowding directly from the light curve amplitudes of >200 Cepheids in 3 SNe~Ia hosts and NGC 4258 as observed by HST -- the first near-infrared amplitudes measured beyond the Magellanic Clouds. Where local crowding is minimal, we find near-infrared amplitudes match Milky Way Cepheids at the same periods. At greater stellar densities we find that the empirically measured amplitudes match the values predicted (with no free parameters) from crowding assessed in the conventional way from local regions, confirming their accuracy for estimating the background at the Cepheid locations. Extragalactic Cepheid amplitudes would need to be ~20% smaller than measured to indicate additional, unrecognized crowding as a primary source of the present discrepancy in H_0. Rather we find the amplitude data constrains a systematic mis-estimate of Cepheid backgrounds to be 0.029 +/- 0.037 mag, more than 5x smaller than the size of the present ~0.2 mag tension in H_0. We conclude that systematic errors in Cepheid backgrounds do not provide a plausible resolution to the Hubble tension., Comment: submitted to ApJ, comments welcome, 25 pages, 7 figures

We report here time domain infrared spectroscopy and optical photometry of the HD145263 silica-rich circumstellar disk system taken from 2003 through 2014. We find an F4V host star surrounded by a stable, massive 1e22 - 1e23 kg (M_Moon to M_Mars) dust disk. No disk gas was detected, and the primary star was seen rotating with a rapid ~1.75 day period. After resolving a problem with previously reported observations, we find the silica, Mg-olivine, and Fe-pyroxene mineralogy of the dust disk to be stable throughout, and very unusual compared to the ferromagnesian silicates typically found in primordial and debris disks. By comparison with mid-infrared spectral features of primitive solar system dust, we explore the possibility that HD 145263's circumstellar dust mineralogy occurred with preferential destruction of Fe-bearing olivines, metal sulfides, and water ice in an initially comet-like mineral mix and their replacement by Fe-bearing pyroxenes, amorphous pyroxene, and silica. We reject models based on vaporizing optical stellar megaflares, aqueous alteration, or giant hypervelocity impacts as unable to produce the observed mineralogy. Scenarios involving unusually high Si abundances are at odds with the normal stellar absorption near-infrared feature strengths for Mg, Fe, and Si. Models involving intense space weathering of a thin surface patina via moderate (T < 1300 K) heating and energetic ion sputtering due to a stellar superflare from the F4V primary are consistent with the observations. The space weathered patina should be reddened, contain copious amounts of nanophase Fe, and should be transient on timescales of decades unless replenished., Comment: 41 Pages, 5 Figures, 5 Tables, Accepted for publication in the Astrophysical Journal