Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance model”

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 1,387 records · Page 77

Exploiting Internal Parallelism for Address Translation in Solid-State Drives

Solid-state Drives (SSDs) have changed the landscape of storage systems and present a promising storage solution for data-intensive applications due to their low latency, high bandwidth, and low power consumption compared to traditional hard disk drives. SSDs achieve these desirable characteristics using internal parallelism—parallel access to multiple internal flash memory chips—and a Flash Translation Layer (FTL) that determines where data are stored on those chips so that they do not wear out prematurely. However, current state-of-the-art cache-based FTLs like the Demand-based Flash Translation Layer (DFTL) do not allow IO schedulers to take full advantage of internal parallelism, because they impose a tight coupling between the logical-to-physical address translation and the data access. In this study to address this limitation, we introduce a new FTL design called Parallel-DFTL that works with the DFTL to decouple address translation operations from data accesses. Parallel-DFTL separates address translation and data access operations into different queues, allowing the SSD to use concurrent flash accesses for both types of operations. We also present a Parallel-LRU cache replacement algorithm to improve the concurrency of address translation operations. To compare Parallel-DFTL against existing FTL approaches, we present a Parallel-DFTL performance model and compare its predictions against those for DFTL and an ideal page-mapping approach. We also implemented the Parallel-DFTL approach in an SSD simulator using real device parameters, and used trace-driven simulation to evaluate Parallel-DFTL’s efficacy. Our evaluation results show that Parallel-DFTL improved the overall performance by up to 32% for the real IO workloads we tested, and by up to two orders of magnitude with synthetic test workloads. Finally, we also found that Parallel-DFTL is able to achieve reasonable performance with a very small cache size and that it provides the best benefit for those workloads with large request size or with high write ratio.

97 MATHEMATICS AND COMPUTING↗

Coated Particle Fuel and Deep Burn Program Monthly Highlights April 2011

The baseline change proposal BCP-FCRD-11026 submitted to change the due date for M21AF080202 'Demonstrate fabrication of Transuranic kernels of Plutonium-239/3.5at%Neptunium-237 using newly installed glove box facilities in ORNL 7930 hot cell complex' from 4/25/11 to 3/30/12 was approved this month. During FY 2011 the CP & DB Program will report Highlights on a monthly basis, but will no longer produce Quarterly Progress Reports. Technical details that were previously included in the quarterly reports will be included in the appropriate Milestone Reports that are submitted to FCRD Program Management. These reports will also be uploaded to the Deep Burn website. The Monthly Highlights report for March 2011, ORNL/TM-2011/96, was distributed to program participants on April 8, 2011. As reported previously, the final Quarterly for FY 2010, Deep Burn Program Quarterly Report for July - September 2010, ORNL/TM-2010/301, was announced to program participants and posted to the website on December 28, 2010. This report discusses the following: (1) Thermochemical Data and Model Development - (a) Thermochemical Modeling, (b) Thermomechanical Behavior, (c) Actinide and Fission Product Transport, (d) Radiation Damage and Properties; (2) TRU (transuranic elements) TRISO (tri-structural isotropic) Development - (a) TRU Kernel Development, (b) Coating Development; (3) Advanced TRISO Applications - Metal Matrix Fuels for LWR; (4) LWR Fully Ceramic Fuel - (a) FCM Fabrication Development, (b) FCM Irradiation Testing; (5) Fuel Performance and Analytical Analysis - Fuel Performance Modeling; and (6) ZrC Properties and Handbook - Properties of ZrC.

Snead, Lance Lewis↗

Electron microscopy analysis of high burnup commercial fuel

Modern electron microscopy analysis can provide detailed insight into the microstructure and chemistry of high burnup UO{sub 2} fuel. This more complete understanding of fuel evolution and performance in reactor can provide critical insight necessary to validate fuel performance models and optimize reactor operations. Preliminary results from the investigation of a polished cross-section of high burnup UO{sub 2} fuel (seven cycles) from the H.B. Robinson pressurized water reactor will be discussed. The analysis focuses on scanning electron microscopy (SEM) of the complete cross-section, while scanning transmission electron microscopy (S/TEM) was performed on targeted features identified by SEM where specimens were fabricated by traditional focused ion beam lift-out techniques. Analysis of the complete cross-section allowed for investigation of the microstructural variations from centerline to fuel/clad interface, pellet/pellet interface, and cladding. Secondary electron imaging and backscatter electron imaging was performed to identify variations in the distribution of fission product secondary phases and bubbles/voids. The analysis of the pellet/pellet interface is a unique feature of this sample where initial analysis shows segregation of fission product phases at the pellet/pellet interface and also indicates an interface microstructure lacking of bubbles and voids. X-ray spectrum imaging was performed by both SEM and STEM techniques to identify the composition of the secondary phases present in the fuel. Multivariate statistical analysis (MVSA) of the X-ray spectrum images has been applied to improve the sensitivity of the analysis and highlight minor variations in the fuel chemistry and secondary phases comprised of fission product as well as the distribution of sub-surface Xe gas bubbles in the high burnup structure. The results presented demonstrate the level of microstructural understanding that is attainable with a combination of SEM techniques along with S/TEM techniques.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

PV Fleet Performance Data Initiative: Performance Index-Based Analysis

In this analysis, we report on 250 PV systems throughout the US, comprising 157 MWdc of system capacity and over 10,000 monthly PI values based on high-frequency (subhourly) energy data and satellite-based resource data. The distribution of PI values is analyzed, and multiple causes of underperformance are assessed, including first-year startup issues, snowfall and inverter downtime. An initial distribution of raw monthly PI values was collected with mean measured / modeled performance of PI = 0.935. After correcting for the three identifiable loss factors mentioned above, an average monthly performance of PI = 0.994 was obtained, with a distribution closely following a Gumbel Extreme Value distribution. In particular, inverter availability was found to contribute a system energy loss of 2.3% on average, except in the first six months of operation when availability losses are closer to 8%. Other startup issues beyond inverter downtime such as partial string outage contributed an additional 5% underperformance in the first year of operation across the fleet. Winter performance was also found to be 5%-10% below summer performance on average, likely due to snowfall. A simple linear relationship was found between snow loss and monthly snowfall accumulation in cm, indicating between 6% - 40% loss in months with nonzero snowfall, depending on climate.

14 SOLAR ENERGY↗

Coated Particle Fuel and Deep Burn Program Monthly Highlights March 2011

During FY 2011 the CP & DB Program will report Highlights on a monthly basis, but will no longer produce Quarterly Progress Reports. Technical details that were previously included in the quarterly reports will be included in the appropriate Milestone Reports that are submitted to FCRD Program Management. These reports will also be uploaded to the Deep Burn website. The Monthly Highlights report for February 2011, ORNL/TM-2011/71, was distributed to program participants on March 8, 2011. As reported previously, the final Quarterly for FY 2010, Deep Burn Program Quarterly Report for July - September 2010, ORNL/TM-2010/301, was announced to program participants and posted to the website on December 28, 2010. This report discusses the following: (1) Thermochemical Data and Model Development - (a) Thermochemical Modeling, (b) Thermomechanical Behavior, (c) Actinide and Fission Product Transport, (d) Radiation Damage and Properties; (2) TRU (transuranic elements) TRISO (tri-structural isotropic) Development - (a) TRU Kernel Development, (b) Coating Development; (3) Advanced TRISO Applications - Metal Matrix Fuels for LWR; (4) LWR Fully Ceramic Fuel - (a) FCM Fabrication Development, (b) FCM Irradiation Testing; and (5) Fuel Performance and Analytical Analysis - Fuel Performance Modeling.

Snead, Lance Lewis↗

Cost and Performance Assumptions for Modeling Electricity Generation Technologies

The goal of this project was to compare and contrast utility scale power plant characteristics used in data sets that support energy market models. Characteristics include both technology cost and technology performance projections to the year 2050. Cost parameters include installed capital costs and operation and maintenance (O&M) costs. Performance parameters include plant size, heat rate,capacity factor or availability factor, and plant lifetime. Conventional, renewable, and emerging electricity generating technologies were considered. Six data sets, each associated with a different model, were selected. Two of the data sets represent modeled results, not direct model inputs. These two data sets include cost and performance improvements that result from increased deployment aswell as resulting capacity factors estimated from particular model runs; other data sets represent model input data. For the technologies contained in each data set, the levelized cost of energy (LCOE) was also evaluated, according to published cost, performance, and fuel assumptions.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Improvement of snowpack simulations in a regional climate model

To improve simulations of regional-scale snow processes and related cold-season hydroclimate, the Community Land Model version 3 (CLM3), developed by the National Center for Atmospheric Research (NCAR), was coupled with the Pennsylvania State University/NCAR fifth-generation Mesoscale Model (MM5). CLM3 physically describes the mass and heat transfer within the snowpack using five snow layers that include liquid water and solid ice. The coupled MM5–CLM3 model performance was evaluated for the snowmelt season in the Columbia River Basin in the Pacific Northwestern United States using gridded temperature and precipitation observations, along with station observations. The results from MM5–CLM3 show a significant improvement in the SWE simulation, which has been underestimated in the original version of MM5 coupled with the Noah land-surface model. One important cause for the underestimated SWE in Noah is its unrealistic land-surface structure configuration where vegetation, snow and the topsoil layer are blended when snow is present. This study demonstrates the importance of the sheltering effects of the forest canopy on snow surface energy budgets, which is included in CLM3. Such effects are further seen in the simulations of surface air temperature and precipitation in regional weather and climate models such as MM5. In addition, the snow-season surface albedo overestimated by MM5–Noah is now more accurately predicted by MM5–CLM3 using a more realistic albedo algorithm that intensifies the solar radiation absorption on the land surface, reducing the strong near-surface cold bias in MM5–Noah. The cold bias is further alleviated due to a slower snowmelt rate in MM5–CLM3 during the early snowmelt stage, which is closer to observations than the comparable components of MM5–Noah. In addition, the over-predicted precipitation in the Pacific Northwest as shown in MM5–Noah is significantly decreased in MM5 CLM3 due to the lower evaporation resulting from the longer snow duration.

54 ENVIRONMENTAL SCIENCES↗

Energy balance in the DSSAT-CSM-CROPGRO model

One potential way to improve crop growth models is for the models to predict energy balance and evapotranspiration (ET) from first principles, thus serving as a check on “engineered” ET methodology. In this paper, we present new implementations and the results of an energy balance model (EBL) developed by Jagtap and Jones (1989) and then implemented in DSSAT’s CROPGRO (CG-EBL) model by Pickering et al. (1995) as a linked energy balance-photosynthesis model that has not been field-tested until now. The energy balance code computes evapotranspiration and other energy balance components, as well as a canopy air temperature, based on three sources (sunlit leaves, shaded leaves, soil surface). Model performance was evaluated with measured biomass and energy fluxes from two sites in Nebraska, namely, the US-Ne2 irrigated maize-soybean rotation field and the US-Ne3 rainfed maize-soybean rotation field, which are part of the Ameriflux eddy covariance network (htt ps://ameriflux.lbl.gov/sites). After implementing new aerodynamic resistances and the stomatal conductance model of the Ball–Berry–Leuning, crop growth, evapotranspiration and soil temperature were simulated well by the EBL model. The EBL improved ET predictions slightly over the often-used FAO56 method [Penman–Monteith (Allen et al., 1998)] for 4 of the 5 years evaluated for both irrigated and rainfed conditions. Further, a significant improvement was achieved using EBL for the simulation of soil temperature at the various depths compared to STEMP, the original subroutine in DSSAT for simulating soil temperature. Compared to the other available DSSAT methods, the EBL explicitly simulates the impacts of crop morphology, physiology and management on the crop’s environment and energy and mass exchange, which in turn directly affect the water use and irrigation requirements, phenology, photosynthesis, growth, sterility, and yield of the crop.

54 ENVIRONMENTAL SCIENCES↗

A Morphological Model to Separate Resolved–Unresolved Sources in the DESI Legacy Surveys: Application in the LS4 Alert Stream

Separating resolved and unresolved sources in large imaging surveys is a fundamental step to enable downstream science, such as searching for extragalactic transients in wide-field time-domain surveys. Here we present our method to effectively separate point sources from the resolved, extended sources in the Dark Energy Spectroscopic Instrument (DESI) Legacy Surveys (LS). We develop a supervised machine learning model based on the Gradient Boosting algorithm XGBoost. The features input to the model are purely morphological and are derived from the tabulated LS data products. We train the model using ∼2 × 10 5 LS sources in the COSMOS field with HST morphological labels and evaluate the model performance on LS sources with spectroscopic classification from the DESI Data Release 1 (∼2 × 10 7 objects) and the Sloan Digital Sky Survey Data Release 17 (∼3 × 10 6 objects), as well as on ∼2 × 10 8 Gaia stars. A significant fraction of LS sources are not observed in every LS filter, and we therefore build a “Hybrid” model as a linear combination of two XGBoost models, each containing features combining aperture flux measurements from the “blue” (gr) and “red” (iz) filters. The Hybrid model shows a reasonable balance between sensitivity and robustness, and achieves higher accuracy and flexibility compared to the LS morphological typing. With the Hybrid model, we provide classification scores for ∼3 × 10 9 LS sources, making this the largest ever machine learning catalog separating resolved and unresolved sources. The catalog has been incorporated into the real-time pipeline of the La Silla Schmidt Southern Survey (LS4), enabling the identification of extragalactic transients within the LS4 alert stream.

astrostatistics↗

PERSIANN-CNN: Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Convolutional Neural Networks

Abstract Accurate and timely precipitation estimates are critical for monitoring and forecasting natural disasters such as floods. Despite having high-resolution satellite information, precipitation estimation from remotely sensed data still suffers from methodological limitations. State-of-the-art deep learning algorithms, renowned for their skill in learning accurate patterns within large and complex datasets, appear well suited to the task of precipitation estimation, given the ample amount of high-resolution satellite data. In this study, the effectiveness of applying convolutional neural networks (CNNs) together with the infrared (IR) and water vapor (WV) channels from geostationary satellites for estimating precipitation rate is explored. The proposed model performances are evaluated during summer 2012 and 2013 over central CONUS at the spatial resolution of 0.08° and at an hourly time scale. Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks (PERSIANN)–Cloud Classification System (CCS), which is an operational satellite-based product, and PERSIANN–Stacked Denoising Autoencoder (PERSIANN-SDAE) are employed as baseline models. Results demonstrate that the proposed model (PERSIANN-CNN) provides more accurate rainfall estimates compared to the baseline models at various temporal and spatial scales. Specifically, PERSIANN-CNN outperforms PERSIANN-CCS (and PERSIANN-SDAE) by 54% (and 23%) in the critical success index (CSI), demonstrating the detection skills of the model. Furthermore, the root-mean-square error (RMSE) of the rainfall estimates with respect to the National Centers for Environmental Prediction (NCEP) Stage IV gauge–radar data, for PERSIANN-CNN was lower than that of PERSIANN-CCS (PERSIANN-SDAE) by 37% (14%), showing the estimation accuracy of the proposed model.

54 ENVIRONMENTAL SCIENCES↗

AGR-3/4 Irradiation Test Predictions using PARFUME

PARFUME, a fuel performance modeling code used for high temperature gas reactors, was used to model the AGR-3/4 irradiation test using as-run physics and thermal hydraulics data. The AGR-3/4 test is the combined third and fourth planned irradiations of the Advanced Gas Reactor (AGR) Fuel Development and Qualification Program. The AGR-3/4 test train consists of twelve separate and independently controlled and monitored capsules. Each capsule contains four compacts filled with both uranium oxycarbide (UCO) unaltered “driver” fuel particles and UCO designed-to-fail (DTF) fuel particles. The DTF fraction was specified to be 1×10-2. This report documents the calculations performed to predict failure probability of TRISO-coated fuel particles during the AGR-3/4 experiment. In addition, this report documents the calculated source term from both the driver fuel and DTF particles. The calculations include the modeling of the AGR-3/4 irradiation that occurred from December 2011 to April 2014 in the Advanced Test Reactor (ATR) over a total of ten ATR cycles including seven normal cycles, one low power cycle, one unplanned outage cycle, and one Power Axial Locator Mechanism cycle. Results show that failure probabilities are predicted to be low, resulting in zero fuel particle failures per capsule. The primary fuel particle failure mechanism occurred as a result of localized stresses induced by the calculated IPyC cracking. Assuming 1,872 driver fuel particles per compact, failure probability calculated by PARFUME leads to no predicted particle failure in the AGR-3/4 driver fuel. In addition, the release fraction of fission products Ag, Cs, and Sr were calculated to vary depending on capsule location and irradiation temperature. The maximum release fraction of Ag occurs in Capsule 7 reaching up to 56% for the driver fuel and 100% for the DTF fuel. The release fraction of the other two fission products, Cs and Sr, are much smaller and in most cases less than 1% for the driver fuel. The notable exception occurs in Capsule 7 where the release fraction for Cs and Sr reach up to 0.73% and 2.4%, respectively, for the driver fuel. For the DTF fuel in Capsule 7, the release fraction for Cs and Sr are estimated to be 100% and 5%, respectively.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Impacts of cloud superparameterization on projected daily rainfall intensity climate changes in multiple versions of the Community Earth System Model

Abstract Changes in the character of rainfall are assessed using a holistic set of statistics based on rainfall frequency and amount distributions in climate change experiments with three conventional and superparameterized versions of the Community Atmosphere Model (CAM and SPCAM). Previous work has shown that high‐order statistics of present‐day rainfall intensity are significantly improved with superparameterization, especially in regions of tropical convection. Globally, the two modeling approaches project a similar future increase in mean rainfall, especially across the Inter‐Tropical Convergence Zone (ITCZ) and at high latitudes, but over land, SPCAM predicts a smaller mean change than CAM. Changes in high‐order statistics are similar at high latitudes in the two models but diverge at lower latitudes. In the tropics, SPCAM projects a large intensification of moderate and extreme rain rates in regions of organized convection associated with the Madden Julian Oscillation, ITCZ, monsoons, and tropical waves. In contrast, this signal is missing in all versions of CAM, which are found to be prone to predicting increases in the amount but not intensity of moderate rates. Predictions from SPCAM exhibit a scale‐insensitive behavior with little dependence on horizontal resolution for extreme rates, while lower resolution (∼2°) versions of CAM are not able to capture the response simulated with higher resolution (∼1°). Moderate rain rates analyzed by the “amount mode” and “amount median” are found to be especially telling as a diagnostic for evaluating climate model performance and tracing future changes in rainfall statistics to tropical wave modes in SPCAM.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Rapid Calculation of Max-Min Fair Rates for Multi-Commodity Flows in Fat-Tree Networks

Max-min fairness is often used in the performance modeling of interconnection networks. Existing methods to compute max-min fair rates for multi-commodity flows have high complexity and are computationally infeasible for large networks. In this paper, we show that by considering topological features, this problem can be solved efficiently for the fat-tree topology that is widely used in data centers and high performance compute clusters. Several efficient new algorithms are developed for this problem, including a parallel algorithm that can take advantage of multi-core and shared-memory architectures. Using these algorithms, we demonstrate that it is possible to find the max-min fair rate allocation for multi-commodity flows in fat-tree networks that support tens of thousands of nodes. We evaluate the run-time performance of the proposed algorithms and show improvement in orders of magnitude over the previously best known method. Finally, we further demonstrate a new application of max-min fair rate allocation that is only computationally feasible using our new algorithms.

97 MATHEMATICS AND COMPUTING↗

Cryogenic Control Circuitry for Superconducting Qubits

Superconducting qubits have reached the point where system designers are worried about the heat that control wiring brings into the cryostat. To continue scaling cryogenic quantum systems, control solutions that work inside the cold space must be explored. One possibility is to use control electronics that is native to superconductivity, so called single-flux-quantum (SFQ) circuitry, to form an interface between qubits and whatever other electronics is needed to control eventual quantum systems. To begin exploring the utility of SFQ as control circuitry, we performed modeling and experiments on qubit readout using ballistic fluxons which are SFQ in the limit of ballistic fluxon transport. Our modeling results show that a flavor of qubit, the fluxonium, can be read out using ballistic fluxons. We designed test samples to prove some of the key concepts needed for such a readout but were ultimately unable to getting a working demonstration. The lack of testing success was due to challenges in fabrication and running short of time to perform testing rather than a fundamental problem with our analysis.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Tropical dry forest response to nutrient fertilization: a model validation and sensitivity analysis

Abstract. Soil nutrients, especially nitrogen (N) and phosphorus (P), regulate plant growth and hence influence carbon fluxes between the land surface and atmosphere. However, how forests adjust biomass partitioning to leaves, wood, and fine roots in response to N and/or P fertilization remains puzzling. Recent work in tropical forests suggests that trees increase fine root production under P fertilization, but it is unclear whether mechanistic models can reproduce this dynamic. In order to better understand mechanisms governing nutrient effects on plant allocation and improve models, we used the nutrient-enabled ED2 model to simulate a fertilization experiment being conducted in a secondary tropical dry forest in Costa Rica. We evaluated how different allocation parameterizations affected model performance. These parameterizations prescribed a linear relationship between relative allocation to fine roots and soil P concentrations. The slope of the linear relationship was allowed to be positive, negative, or zero. Some parameterizations realistically simulated leaf, wood, and fine root production, and these parameterizations all assumed a positive relationship between relative allocation to fine roots and soil P concentration. Model simulations of a 30-year timeframe indicated strong sensitivity to parameterization and fertilization treatment. Without P fertilization, the simulated aboveground biomass (AGB) accumulation was insensitive to the parameterization. With P fertilization, the model was highly sensitive to the parameterization and the greatest AGB accumulation occurred when relative allocation to fine roots was independent of soil P. Our study demonstrates the need for simultaneous measurements of leaf, wood, and fine root production in nutrient fertilization experiments and for longer-term experiments. Models that do not accurately represent allocation to fine roots may be highly biased in their simulations of AGB, especially on multi-decadal timescales.

Environmental Sciences & Ecology↗

Assessment of an In Silico Mechanistic Model for Proarrhythmia Risk Prediction Under the Ci PA Initiative

The International Council on Harmonization (ICH) S7B and E14 regulatory guidelines are sensitive but not specific for predicting which drugs are pro‐arrhythmic. In response, the Comprehensive In Vitro Proarrhythmia Assay (Ci PA ) was proposed that integrates multi‐ion channel pharmacology data in vitro into a human cardiomyocyte model in silico for proarrhythmia risk assessment. Previously, we reported the model optimization and proarrhythmia metric selection based on Ci PA training drugs. In this study, we report the application of the prespecified model and metric to independent Ci PA validation drugs. Over two validation datasets, the Ci PA model performance meets all pre‐specified measures for ranking and classifying validation drugs, and outperforms alternatives, despite some in vitro data differences between the two datasets due to different experimental conditions and quality control procedures. This suggests that the current Ci PA model/metric may be fit for regulatory use, and standardization of experimental protocols and quality control criteria could increase the model prediction accuracy even further.

59 BASIC BIOLOGICAL SCIENCES↗

Physics vs structure: A systematic benchmark of learning strategies for multi-zone building thermal dynamics

Recent advances in physics-informed and data-driven machine learning promise improved thermal models for advanced building control, yet there is limited quantitative evidence on when added physics structure and architectural complexity are beneficial. Here, this work presents a systematic benchmark of five representative system identification methods for modeling multi-zone building thermal dynamics: linear state-space models, multi-layer perceptrons, neural state-space models, neural ordinary differential equations, and physically-consistent neural networks. The methods are evaluated across multiple data regimes and zone coupling strategies. Using a high-fidelity multi-zone commercial building emulator, we examine short-term and long-term prediction accuracy, computational efficiency, and ease of development. Our results reveal critical trade-offs between prediction performance, model complexity, and physical consistency. We demonstrate that decoupled, nonlinear black-box models consistently outperform coupled physics-constrained architectures in both predictive accuracy and out-of-distribution robustness in majority of the test cases for the building type considered in the study. Our findings quantify the cost of complexity in building thermal modeling and provide concrete, actionable, scenario-based guidelines for selecting model classes for control-oriented applications.

Building thermal modeling↗

Predictive Models of Genetic Redundancy in Arabidopsis thaliana

Abstract Genetic redundancy refers to a situation where an individual with a loss-of-function mutation in one gene (single mutant) does not show an apparent phenotype until one or more paralogs are also knocked out (double/higher-order mutant). Previous studies have identified some characteristics common among redundant gene pairs, but a predictive model of genetic redundancy incorporating a wide variety of features derived from accumulating omics and mutant phenotype data is yet to be established. In addition, the relative importance of these features for genetic redundancy remains largely unclear. Here, we establish machine learning models for predicting whether a gene pair is likely redundant or not in the model plant Arabidopsis thaliana based on six feature categories: functional annotations, evolutionary conservation including duplication patterns and mechanisms, epigenetic marks, protein properties including posttranslational modifications, gene expression, and gene network properties. The definition of redundancy, data transformations, feature subsets, and machine learning algorithms used significantly affected model performance based on holdout, testing phenotype data. Among the most important features in predicting gene pairs as redundant were having a paralog(s) from recent duplication events, annotation as a transcription factor, downregulation during stress conditions, and having similar expression patterns under stress conditions. We also explored the potential reasons underlying mispredictions and limitations of our studies. This genetic redundancy model sheds light on characteristics that may contribute to long-term maintenance of paralogs, and will ultimately allow for more targeted generation of functionally informative double mutants, advancing functional genomic studies.

59 BASIC BIOLOGICAL SCIENCES↗