Default Parameters Research Articles

Development of a Harmonized Protocol for Measurement of Foliar Wash-off Parameters: First Results.

The foliar wash-off coefficient is a parameter used by environmental fate models to estimate the amount of chemical removed from leaf surfaces by rainfall. In the European Union it is used by FOCUS surface water models to estimate soil loadings following rainfall after leaf surfaces have been treated with plant protection products. Currently, a default value of 0.5/cm is assumed for this parameter, although there is provision to provide experimental data to replace this default. The European Food Safety Authority proposed to increase the default parameter value to 1.0/cm. This increases the need for experimental refinement studies. However, no guidance for a harmonized protocol exists to estimate this parameter. We describe the results of a ring-test conducted to start a process of developing a harmonized experimental protocol to measure the foliar wash-off parameters, conducted by several laboratories across Europe. The proposed design uses whole plants (rather than individual leaves) to retain as much realism as possible. The extent of wash-off is then determined by comparison of compound residues in two sets of plants (with and without a defined rainfall event) measured using a fully validated crop residue method. This initial ring test used tebuconazole (Folicur EW 250) sprayed at 100 g ai/ha onto tomato plants at BBCH25. Each laboratory measured the residues before and after a rainfall event of 20 mm/h for 1 h and calculated the percentage of wash-off from these data. There was good agreement across the eight participating laboratories with a mean percentage of wash-off of 66.8% and a 95% confidence interval of ±11.8%. Determination of robust wash-off parameters was therefore considered feasible using the proposed test design. Environ Toxicol Chem 2023;42:535-541. © 2022 SETAC.

Improvement of local ozone phytotoxicity modelling for autochthonous grape cultivars

The grapevine is a key crop for Mediterranean environments and is both sensitive to climate warming and air pollutants, of which ozone is the most damaging to crop yield and quality. Ambient ozone effects on the grapevine have been noticed since the late fifties but risk assessments are still impaired by the lack of information concerning differences in cultivar sensitivity, and adaptation capacity to environmental factors including drought conditions. This study develops a specific parametrization for autochthonous grape cultivars within a leaf-level stomatal flux model, the DO3SE model, coupled with a meteorological and atmospheric chemical transport modelling system, the WRF-CHIMERE, by using a renowned wine producing area, the Douro wine region of Portugal, as case study. The DO3SE model parametrization introduced in this study included phenology, photosynthetic active radiation, air temperature, air vapour pressure deficit, and leaf water potential as a proxy of soil water content. The modelling experiments, which included simulations with the current default Convention on Long-Range Transboundary Air Pollution DO3SE parametrization and with the proposed parametrization, covered a reference grapevine growing season (from April to September 2017), during which a measuring campaign was carried out. Simulation results show that the proposed parametrization succeeded to replicate the observed grapevine leaf-level stomatal flux gradient in the region. Both field and modified DO3SE model values indicate that considerable areas in the Douro wine region of Portugal can exceed critical phytotoxic ozone dose (POD) values, although with a lower and different spatial extent when compared to the default DO3SE parametrization for the grapevine. However, under irrigated conditions, the POD values increase, and the values are close to those obtained with the default parametrization. Overall, the research results indicate that air quality management, in particular the reduction of ozone levels in the ambient air, must also be considered to define sustainable grape and wine production strategies in the context of climate and wine production management change.

A Robust Artificial Intelligence Approach with Explainability for Measurement and Verification of Energy Efficient Infrastructure for Net Zero Carbon Emissions

Rapid urbanization across the world has led to an exponential increase in demand for utilities, electricity, gas and water. The building infrastructure sector is one of the largest global consumers of electricity and thereby one of the largest emitters of greenhouse gas emissions. Reducing building energy consumption directly contributes to achieving energy sustainability, emissions reduction, and addressing the challenges of a warming planet, while also supporting the rapid urbanization of human society. Energy Conservation Measures (ECM) that are digitalized using advanced sensor technologies are a formal approach that is widely adopted to reduce the energy consumption of building infrastructure. Measurement and Verification (M&V) protocols are a repeatable and transparent methodology to evaluate and formally report on energy savings. As savings cannot be directly measured, they are determined by comparing pre-retrofit and post-retrofit usage of an ECM initiative. Given the computational nature of M&V, artificial intelligence (AI) algorithms can be leveraged to improve the accuracy, efficiency, and consistency of M&V protocols. However, AI has been limited to a singular performance metric based on default parameters in recent M&V research. In this paper, we address this gap by proposing a comprehensive AI approach for M&V protocols in energy-efficient infrastructure. The novelty of the framework lies in its use of all relevant data (pre and post-ECM) to build robust and explainable predictive AI models for energy savings estimation. The framework was implemented and evaluated in a multi-campus tertiary education institution setting, comprising 200 buildings of diverse sensor technologies and operational functions. The results of this empirical evaluation confirm the validity and contribution of the proposed framework for robust and explainable M&V for energy-efficient building infrastructure and net zero carbon emissions.

Thermodynamic property measurements and modelling of CO2 + difluoromethane (R32): Density, heat capacity, and vapour-liquid equilibrium

Accurate thermodynamic property data and models are demanded to reduce the design margins of industrial processes based on these fluids. In this work, measurements of density (ρ), heat capacity (cp), and vapour-liquid equilibrium (VLE) for CO2 + difluoromethane (R32) have been performed by the vibrating tube densimeter and the magnetic suspension balance densimeter, the visual cell and gas chromatograph, and the differential scanning calorimeter, respectively. Experiments were conducted over the temperatures ranging from (208.3 to 334.3) K and pressures reaching 10.04 MPa at CO2 concentrations of 0.7, 0.8 and 0.9. The measured data, together with the results reported in literature where applicable, were subsequently applied to regress the binary interaction parameters utilised in the mixture functions of Helmholtz energy model from REFPROP 10 software package. Noticeable improvements have been achieved for the model's ability to represent thermodynamic property data. The most significant achievement exists in density description: compared with the default parameters, the root-mean-square (RMS) deviation has been decreased by half.

A Benchmark Framework for Multiregion Analysis of Vesselness Filters.

Vessel enhancement (aka vesselness) filters, are part of angiographic image processing for more than twenty years. Their popularity comes from their ability to enhance tubular structures while filtering out other structures, especially as a preliminary step of vessel segmentation. Choosing the right vesselness filter among the many available can be difficult, and their parametrization requires an accurate understanding of their underlying concepts and a genuine expertise. In particular, using default parameters is often not enough to reach satisfactory results on specific data. Currently, only few benchmarks are available to help the users choosing the best filter and its parameters for a given application. In this article, we present a generic framework to compare vesselness filters. We use this framework to compare seven gold standard filters. Our experiments are performed on three public datasets: the hepatic Ircad dataset (CT images), the Bullit dataset (brain MRA images) and the synthetic VascuSynth dataset. We analyse the results of these seven filters both quantitatively and qualitatively. In particular, we assess their performances in key areas: the organ of interest, the whole vascular network neighbourhood and the vessel neighbourhood split into several classes, based on their diameters. We also focus on the vessels bifurcations, which are often missed by vesselness filters. We provide the code of the benchmark, which includes up-to-date C++ implementations of the seven filters, as well as the experimental setup (parameter optimization, result analysis, etc.). An online demonstrator is also provided to help the community apply and visually compare these vesselness filters.

Reparameterized semi-empirical methods for computing anharmonic vibrational frequencies of multiply-bonded hydrocarbons

Abstract Reparameterized semi-empirical methods can reproduce gas-phase experimental vibrational frequencies to within 24 cm−1 or better for a 100-fold decrease in computational cost in the anharmonic fundamental vibrational frequencies. To achieve such accuracy and efficiency, the default parameters in the PM6 semi-empirical model are herein optimized to reproduce the experimental and high-level theoretical vibrational spectra of three small hydrocarbon molecules, C2H2, c-C3H2, and C2H4, with the hope that these same parameters will be applicable to large polycyclic aromatic hydrocarbons (PAHs). This massive cost reduction allows for the computation of explicit anharmonic frequencies and the inclusion of resonance corrections that have been shown to be essential for accurate predictions of anharmonic frequencies. Such accurate predictions are necessary to help to disentangle the heretofore unidentified infrared spectral features observed around diverse astronomical bodies and hypothesized to be caused by PAHs, especially with the upcoming influx of observational data from the James Webb Space Telescope. The optimized PM6 parameters presented herein represent a substantial step in this direction with those obtained for ethylene (C2H4) yielding a 37% reduction in the mean absolute error of the fundamental frequencies compared to the default PM6 parameters.

Evaluating the Hydrus-1D Model Optimized by Remote Sensing Data for Soil Moisture Simulations in the Maize Root Zone

The Hydrus-1D model is widely used for soil water content (SWC) simulations, wherein the exact configuration of soil hydraulic parameters is key to accuracy. To assess the feasibility of using “low-cost” multi-source remote sensing data to optimize the parameters of the Hydrus-1D model, five types of soil hydrodynamic parameter acquisition methods were designed for comparative evaluation, including the use of default parameters for soil texture types (DSHP), predictions from three and five soil mechanical composition parameters (NNP3/NNP5), inverse solutions from measured historical data (ISHD), and innovative introduction of historical remote sensing data (ERA-5 land reanalysis information and MODIS LAI products) instead of ground measured data for the inverse solution (ISRS). Two spring maize crops were planted in Beijing, China, in 2021 and 2022. Meteorological, soil, and crop data were collected as real measurements of the true values during the growth period. The boundary flux characteristics of the model simulation results were analyzed. The accuracy differences in the five approaches were compared from three perspectives: overall root zone, growth stage, and soil depth. The results showed that (1) evapotranspiration was the main pathway for soil water depletion in the root zone of maize; the actual total evapotranspiration accounted for 68.26 and 69.43% of the total precipitation in 2012 and 2022, respectively. (2) The accuracy of the SWC simulations in the root zone was acceptable for different approaches in the following order: NNP5 (root mean squared error (RMSE) = 5.47%) > ISRS (RMSE = 5.48%) > NNP3 (RMSE = 5.66%) > ISHD (RMSE = 5.68%) > DSHP (RMSE = 6.57%). The ISRS approach based on remote sensing data almost achieved the best performance while effectively reducing the workload and cost. (3) The accuracy of the SWC simulation at different growth stages was ranked as follows: seedling stage (mean absolute error (MAE) = 3.29%) > tassel stage (MAE = 4.68%) > anthesis maturity stage (MAE = 5.52%). (4) All approaches’ simulation errors exhibited a decreasing trend with increasing soil depth. The ISHD approach, based on the measured data, achieved the best performance at a depth of 60 cm (MAE = 2.8%). The Hydrus-1D model optimized using multi-source remote sensing data can effectively simulate SWC in the maize root zone with low working cost, which is significant for applications in areas where it is difficult to obtain field soil hydrodynamic property parameters to simulate SWC at a global scale.

Parameterizing echo state networks for multi-step time series prediction

Prediction of multi-dimensional time-series data, which may represent such diverse phenomena as climate changes or financial markets, remains a challenging task in view of inherent nonlinearities and non-periodic behavior. In contrast to other recurrent neural networks, echo state networks (ESNs) are attractive for (online) learning due to lower requirements w.r.t.training data and computational power. However, the randomly-generated reservoir renders the choice of suitable hyper-parameters as an open research topic. We systematically derive and exemplarily demonstrate design guidelines for the hyper-parameter optimization of ESNs. For the evaluation, we focus on the prediction of chaotic time series, an especially challenging problem in machine learning. Our findings demonstrate the power of a hyper-parameter-tuned ESN when auto-regressively predicting time series over several hundred steps. We found that ESNs’ performance improved by 85.1%-99.8% over an already wisely chosen default parameter initialization. In addition, the fluctuation range is considerably reduced such that significantly worse performance becomes very unlikely across random reservoir seeds. Moreover, we report individual findings per hyper-parameter partly contradicting common knowledge to further, help researchers when training new models.

Pre-Conditioning Easix Is Associated with Survival after Autologous Hematopoietic Stem Cell Transplant in Mature T-Cell Lymphomas

Introduction: Mature T-cell lymphomas (MTCLs) comprise a group of rare, heterogeneous entities with poor prognosis. Optimal use of autologous (auto) and allogeneic (allo) hematopoietic stem cell transplantation (HSCT) remains an area of active study (Marchi & O'Connor 2019). Prognostic scores help guide clinical decision making and are useful to assess baseline risk in retrospective studies. The Endothelial Activation and Stress Index (EASIX) has shown promise in predicting post-alloHSCT survival (Shouval et al. 2019, Luft et al. 2020). However, we have little data on the association between EASIX and post-HSCT survival specific to the MTCL population. Here, we assess the relationship between EASIX prior to conditioning chemotherapy (EASIX-pre) and post-HSCT survival. Methods: Subjects with MTCL who underwent auto- or alloHSCT were identified from a Montefiore Medical Center database. Data were collected on the following by manual chart review: date of birth, sex, race, ethnicity, MTCL subtype, date of stem cell infusion (D0), stem cell source, conditioning intensity, and survival. To be included in the auto- or alloHSCT cohort, a subject's MTCL subtype had to have auto- or alloHSCT as a recommendation in the NCCN Guidelines Version 2.2022 (Horwitz et al. 2022). Conditioning intensity was determined by the Center for International Blood and Bone Marrow Transplant Research operational definition (Giralt et al. 2009). EASIX prior to conditioning (EASIX-pre) was calculated as (lactate dehydrogenase [U/L] x creatinine [mg/dL])/platelets [109 cells/L] at D-21. To estimate lab values at D-21 we used linear interpolation on data available from D-81 to D-9. Missing numerical variables were imputed using scikit-learn's IterativeImputer with the RandomForestRegressor estimator and default parameters, a method similar to missForest. The Nelson-Aalen estimate of the cumulative hazard and event status at last follow up were included in imputation (White & Royston 2009). Missing categorical variables were placed into an unknown category. Univariable Cox models were made for each variable. Variables whose Cox models had likelihood ratio test p < 0.20 were used for adjustment in multivariable models. Survival was measured from D0 to last follow up. Events included both death and discharge to hospice with unknown date of death. Data were analyzed using Python, lifelines, and sci-kit learn (Davidson-Pilon 2019, Pedregosa et al. 2011). Results: 23 subjects underwent autoHSCT 9/19/2006-3/24/2021. Median follow up among censored subjects was 2.30 years. 9 subjects had peripheral T-cell lymphoma, NOS (PTCL, NOS), 7 had angioimmunoblastic T-cell lymphoma, 4 had ALK-negative anaplastic large cell lymphoma, and 3 had other subtypes. EASIX 10th, 50th, and 90th percentiles were 0.55, 1.51, and 3.68. The univariable model for age at D0 had likelihood ratio test p < 0.20 and was used for adjustment in the multivariable Cox model; other potential adjustment variables did not meet this criterion. EASIX-pre adjusted for age at D0 had HR 1.37 (95% CI 1.06-1.76, p = 0.0155, Fig. 1a). 32 subjects underwent alloHSCT 4/22/2008-4/5/2021. Median follow up among censored subjects was 3.36 years. 16 subjects had adult T-cell leukemia/lymphoma, 6 had hepatosplenic T-cell lymphoma, 4 had mycosis fungoides/Sezary syndrome, 4 had PTCL, NOS, and 2 had other subtypes. EASIX 10th, 50th, and 90th percentiles were 0.55, 1.16, and 2.82. The univariable models for age at D0, race, and conditioning intensity had p < 0.20 and were included for adjustment in a multivariable Cox model. In the multivariable model, EASIX-pre had HR 1.12 (95% CI 1.00-1.26, p = 0.0584, Fig. 1b). Conclusion: In our MTCL cohort, at an alpha level of 0.05, EASIX-pre was independently associated with post-autoHSCT but not post-alloHSCT survival. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Accurate Detection of Clinically Actionable Copy Number Variants in Diverse Hematological Neoplasms By Routine Targeted Sequencing: A Comparative Performance Study

Background: Accurate genotyping of Copy Number Variations (CNVs) is an essential part of the clinical evaluation of diverse hematologic malignancies. Conventional karyotyping and FISH are current gold standards to assess CNVs, but both have some limitations. Next-generation targeted sequencing panels are commonly incorporated into current clinical practice in most cancer centers. We examined the utility of CANARy (Copy number ANomaly Assessment and Recovery), a CNV profiling method for three hematological malignancies using targeted panel sequencing data. Patients and Methods: We studied patients with myelodysplastic syndrome (MDS), chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), and plasma cell disorders (PCD) identified from a cohort of 4752 hematological specimens profiled at Stanford University by conventional cytogenetics and next-generation sequencing (NGS) between 2018 and 2021. We curated a cohort of 200 patients diagnosed with MDS (n=72), CLL (n=46), and PCDs including MM, AL amyloidosis, and MGUS (PCD, n=82), along with 50 cytogenetically normal controls. Specimens were bone marrow aspirates, peripheral blood, or FFPE tissue. All samples were characterized by conventional metaphase karyotyping and/or FISH and we focused on the analysis of clinically relevant CNV (-5q, -7/-7q, -20q, +8 for MDS; -11q22, -13q14, -17p13, +12 for CLL; and +1q21, -1p32, -13q14, -17p.13 for PCD). The specimens separately had clinical NGS (164 genes, Heme-STAMP) performed no more than four weeks from standard cytogenetic analysis. We applied multiple CNV enumeration tools to sequencing data for CNV genotyping at these loci, including our recently developed method CANARy (Chabon et al, Nature 2020). CANARy is a read depth-based, genome-wide CNV caller from targeted sequencing data that combines information from on-target and off-target fragments to assign a final Z-score and copy number index to a set of pre-defined genomic regions. We benchmarked CANARy alongside three others publicly available CNV callers: CNVkit (Talevich et al, PLoS Comput Biol. 2016), ichorCNA (Adalsteinsson et al, Nat Commun. 2017), and WISECONDOR (Straver et al, Nucleic Acids Res. 2014). All tools were run using the default parameters and following author recommendations. We calculated the accuracy and other statistics parameters for each CNV caller, considering karyotype/FISH as the true reference. We estimated tumor fraction using two strategies: the variant allele fraction (VAF) of disease-related driver mutations and using the ichorCNA estimate. Results: A total of 211 CNVs were identified by karyotype and/or FISH, with disease-specific distribution summarized in Fig. A. CANARy detected 148/211 (70.1%), CNVkit 88/211 (41.7%), ichorCNA 116/211 (55%) and WISECONDOR 158/211 (74.9%). The overall accuracy for CANARy, CNVkit, ichorCNA and WISECONDOR was 86.3%, 81.8%, 84.3% and 83.1% respectively. CANARy had both the highest overall accuracy (86.3%) and overall F1-score (0.73) (Fig.B). WISECONDOR had the highest number of false positives (83 vs. CANARy = 48). Interestingly, all CNV tools performed better in MDS specimens, with 92% accuracy for CANARy in MDS samples, likely related to the higher neoplastic purity of these samples. The median ichorCNA tumor fraction estimates for MDS, CLL and PCD were 22% (0-99%), 14% (0-70%), and 7% (0-94%), respectively. In addition, CNV calling was inferior in PCD tumors than other tumor types. Within CLL, -11q22 and -13q14 lesions were more challenging to detect with any of the four methods, with CANARy sensitivity of 50% and 31.3% respectively, likely related to the smaller minimal deletion regions of these lesions. The 17p deletion in CLL cases, known for its poor prognostic impact, was best identified by CANARy; with a 100% sensitivity and a 97.6% specificity. The only non-concordant result for -17p was a false positive that was detected by all 4 NGS CNV genotypers but not FISH, possibly suggesting a false negative FISH result. Conclusions: CANARy accurately identifies clinically significant CNVs in common hematologic neoplasms and could thus serve as a complementary approach to the current cytogenetic evaluation. CANARy uses data already generated from targeted sequencing panels and therefore does not impose additional work or cost. Low tumor fraction could be a limitation for CNV calling. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Single-Cell Dissection of Primary Quadruplet- and Triplet-Resistant Multiple Myeloma

Background: With the advent of triplet and quadruplet induction therapies, response rates and survival outcomes have improved significantly in multiple myeloma. However, a minority of patients does not achieve deep responses to these anti-myeloma therapies and survival outcomes for these patients remain dismal. The molecular mechanisms of primary therapeutic resistance remain poorly understood and novel therapeutic approaches are needed for these high-risk patients. Methods: We generated single-cell gene expression and chromatin accessibility profiles at the time of diagnosis from 16 patients with multiple myeloma using the 10X Genomics Multiome platform. Viably frozen CD138+ bone marrow plasma cells were thawed and washed before removing dead cells and isolating nuclei according to the manufacturer's protocols. RNA- and ATAC-seq libraries were constructed per the manufacturer's user guides. GEX and ATAC libraries were sequenced separately on an Illumina HiSeq 4000 instrument before demultiplexing, alignment to the reference genome, and post-alignment quality control. The Cell Ranger pipelines with default parameters were used to trim reads, align, count unique molecular identifiers, and call cells, transcripts, and peaks. Downstream processing was carried out using Seurat and Signac. Doublets were identified by DoubletFinder and batch effects were corrected using Harmony. After quality control for feature counts and mitochondrial gene expression, there were a total of 16188 single cells: 6139 cells from 6 patients resistant to quadruplet induction (Dara-Ixa-Rd), 4841 cells from 4 patients resistant to triplet induction (VRd), and 5208 cells from 6 patients responding to quadruplet / triplet induction. Primary resistance was defined as not achieving at least a very good partial response (VGPR+ after a minimum of 4 cycles of therapy). Expression of the up-regulated resistance genes in single cells was measured by calculating a module score, with higher numbers indicating higher expression (Tirosh et al. Science 2016;352(6282):189-196). We used a score cut-off of 0.200 to define increased expression of resistance genes. Results: Half of the patients were female, the median age at diagnosis was 61 years (42-80), and half of them had high-risk cytogenetics. There was differential gene expression between cells from quadruplet-resistant and -responsive patients (1272 genes, FDR<0.05). Likewise, there was differential gene expression triplet-resistant and -responsive patients (1354 genes, FDR<0.05). There were 212 genes up- and 222 down-regulated in both comparisons (shared resistance genes), an overlap unlikely to have arisen by chance (p<2.22x10-16). Functional annotation of these shared differentially expressed genes revealed up-regulation of GTPase (FDR=4.01x10-4), kinase (FDR=4.39x10-3), and transferase (FDR=1.65x10-3) activity in cells from resistant patients. Conversely, there was down-regulation of genes involved in ubiquitin (FDR=6.63x10-13) and ubiquitin-like protein ligase binding (FDR=2.50x10-12) as well as unfolded protein binding (FDR=1.21x10-10). Expectedly, the mean expression of the shared up-regulated resistance genes was higher among quadruplet- (score=0.261) and triplet-resistant patients (score=0.212) compared to responding patients (score=0.096, Figure 1A). This difference in gene expression was driven by larger subpopulations of cells with increased expression of resistance genes: Quadruplet-resistant (39.2% of cells) and triplet-resistant (37.5% of cells), compared to patients responding to induction therapy (17.2% of cells, Figure 1B). The subpopulations of cells with increased expression of resistance genes were characterized by increased accessibility in Spi-B (FDR=2.55x10-2) and RAR-γ (FDR=1.86x10-3) binding sites. The shared up-regulated genes included known drug resistance genes such as IKZF1 and viable therapeutic targets including CCND3, RASA1, AKT3, CDK8, and ATM. Conclusions: Single-cell interrogation of the transcriptome and epigenome of patients with primary resistant multiple myeloma suggests an overexpression of actionable drug targets and increased cell proliferation in cell subpopulations. There is biological rationale for the investigation of novel individualized therapeutic approaches like IKZF1 degradation and targeted kinase inhibition in this patient population. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Excitation functions for fast-neutron induced reactions on zinc

Cross sections of the 64Zn(n, p)64Cu, natZn(n, x)67Cu, 66Zn(n, 2n)65Zn and 70Zn(n, 2n)69mZn reactions have been measured by using the activation technique at 14.0 MeV neutron energy. The neutrons were produced via the 3H(d, n)4He reaction. The present experimental data illustrated satisfactory agreement with most of the previously reported experimental data. Experimental data are compared with evaluated nuclear data of the CENDL-3.2, ENDF/B-VIII.0, JENDL-5, BROND-3.1 and JEFF-3.3 libraries. Besides, these excitation functions were calculated by using theoretical model of the TALYS-1.96 code from thresholds up to 20 MeV. A group set of parameters was obtained which better reproduce the experimental data than the default parameters.

Impact of flexible noise control (FNC) image processing parameters on portable chest radiography.

There is a lack of understanding in the performance of flexible noise control (FNC) processing, which is used in digital radiography on a scanner vendor and has four parameters each involving multiple options. The aim of this study was to investigate the impact of FNC on portable chest imaging. An anthropomorphic chest phantom was imaged using a clinical chest program with 85kV and five radiation dose levels at 40″ source-to-image distance with software-based scatter reduction method. All images were processed without and with FNC. Noise analysis was performed in two regions of interest (ROI) on subtracted noise-only images, and line profiles were generated through a lung-rib interface. In addition, noise power spectra (NPS) analysis was performed in solid water phantoms of 10 and 20cm thicknesses, using the same acquisition program and a range of dose levels. Last, feedback on retrospectively deidentified, reprocessed, and randomized clinical images from 20 portable chest exams was gathered from two thoracic radiologists. Noise reduction performances of FNC were demonstrated, with the level depending on specific FNC parameters, dose levels, ROI placement, and phantom sizes. Higher frequency textural patterns were revealed through the NPS analysis, which varied based on FNC parameters, dose levels, and phantom sizes. Overall, the vendor default parameter FGA0.5 yielded the highest noise reduction and textural artifacts. Radiologist feedback showed consistent preference of no FNC due to the presence of textural artifacts in the FNC-processed images. An algorithm improvement to avoid introducing artifacts would be desired.

Multi‐modelling predictions show high uncertainty of required carbon input changes to reach a 4‰ target

AbstractSoils store vast amounts of carbon (C) on land, and increasing soil organic carbon (SOC) stocks in already managed soils such as croplands may be one way to remove C from the atmosphere, thereby limiting subsequent warming. The main objective of this study was to estimate the amount of additional C input needed to annually increase SOC stocks by 4‰ at 16 long‐term agricultural experiments in Europe, including exogenous organic matter (EOM) additions. We used an ensemble of six SOC models and ran them under two configurations: (1) with default parametrization and (2) with parameters calibrated site‐by‐site to fit the evolution of SOC stocks in the control treatments (without EOM). We compared model simulations and analysed the factors generating variability across models. The calibrated ensemble was able to reproduce the SOC stock evolution in the unfertilised control treatments. We found that, on average, the experimental sites needed an additional 1.5 ± 1.2 Mg C ha−1 year−1 to increase SOC stocks by 4‰ per year over 30 years, compared to the C input in the control treatments (multi‐model median ± median standard deviation across sites). That is, a 119% increase compared to the control. While mean annual temperature, initial SOC stocks and initial C input had a significant effect on the variability of the predicted C input in the default configuration (i.e., the relative standard deviation of the predicted C input from the mean), only water‐related variables (i.e., mean annual precipitation and potential evapotranspiration) explained the divergence between models when calibrated. Our work highlights the challenge of increasing SOC stocks in agriculture and accentuates the need to increasingly lean on multi‐model ensembles when predicting SOC stock trends and related processes. To increase the reliability of SOC models under future climate change, we suggest model developers to better constrain the effect of water‐related variables on SOC decomposition.Highlights The feasibility of the 4‰ target was studied at 16 long‐term agricultural experiments. An ensemble of soil organic carbon models was used to estimate the uncertainty of the predictions. On average across the sites, carbon input had to increase by 119% compared to initial conditions. High uncertainty of the simulations was mainly driven by water‐related variables.

In silico Prediction and Pharmacokinetic Studies on Glucosinolates as a Potential Drug and Key Inhibitor Molecule for Lanosterol-14α- demethylase: A Fungal Membrane Biosynthesis Enzyme.

Glucosinolates (β-thioglucoside-N-hydroxysulfates) are a water-soluble organic anion with sulfur- and nitrogen-containing glycosides which are found in abundance in Cruciferous plants. Ergosterol (ERG13) lanosterol-14α-demethylase protein has been targeted for inhibition studies as a key regulator enzyme of fungal membrane biosynthesis. To understand the molecular mechanism of inhibition of Ergosterol (ERG13) lanosterol- 14α-demethylase by various phytochemicals from brassicales, i.e., glucosinolates and their potential role as putative drug molecules. In this study, in silico analyses were performed to predict the molecular basis of various glucosinolates as a potential inhibitor of lanosterol-14α-demethylase protein, which is a key regulator of fungal membrane biosynthesis and its pharmacodynamics and toxicity profile. 3d structures of various glucosinolates were retrieved from PubChem, and the target protein, lanosterol-14α-demethylase (Pdb ID- 4lxj), was retrieved from the RCSB protein data bank. Molecular docking and interactions were carried out using the PyRx software using the AutoDOCK toolbar with default parameters. Dru- LiTo, ORISIS web servers were used to predict various drug likeliness predictions and Lipinski's Rule of 5, whereas admetSAR was used for prediction of toxicity, and PASS Program was used to study the antifungal and antimicrobial properties of these compounds. This study shows that among the different compounds screened, gluconasturtiin, Glucotropaeolin, and Indolylmethyl-Glucosinolate showed the highest binding energies of -8.7 kcal/mol, -8.5 kcal/mol, and -8.3 kcal/mol with the lanosterol-14α-demethylase, respectively. Further all the compounds follow the Lipinski's rule as well as they are found to be non-carcinogenic and non-cytotoxic in nature. These compounds also show antifungal properties. This study thus reveals that various glucosinolates interact with the ERG13 enzyme at various amino acid positions, which behaves as a catalytic site, thus indicates the probable mechanism of inactivation, and subsequently, these can be used as potential drug molecules. In vitro studies can be taken to further examine the utility of these compounds as antifungal agents.

LBSUN129 Adipose Tissue Analysis Toolkit (at-at): An Imagej Plugin For Automated Analysis Of Adipocyte Number, Size, And Pericellular Fibrosis In Histological Sections

Abstract Obesity-induced pathological expansion of white adipose tissue (WAT) involves adipocyte hypertrophy that occurs in tandem with the development of pericellular fibrosis. Other pathophysiological states, such as cancer and HIV, are also associated with fibrotic transformation of WAT depots. Aging, obesity, inactivity, and estrogen deficiency are also associated with an increase in bone marrow adiposity. Disease-associated changes in WAT morphology are typically assessed using histochemical analysis of WAT sections. However, manual analysis of large histological sections of WAT or bone marrow adipose tissue can be very time-consuming, while digital tools for the analysis of adipocyte size and pericellular fibrosis are limited. In response to this need, we have developed the Adipose Tissue Analysis Toolkit (AT-AT), a plugin for the open-source image analysis software ImageJ. AT-AT employs both local and global differentials in pixel intensity to threshold background apart from adipocyte-free tissue without user input. The AT-AT accurately captures adipocyte perimeter, maximum diameter, minimum diameter, and density per square millimeter of tissue in histological sections stained with common histological dyes, and performs automatic analysis of adipocyte cross-sectional area as well as pericellular fibrosis. AT-AT enables fully-automated analysis of batches of histological images using either default or user-defined adipocyte detection parameters including minimum circularity as well as minimum and maximum cross-sectional area. Likewise, users may set parameters for the thresholding algorithms, including the number of localized thresholds to be calculated in each image. AT-AT's thresholding techniques confer uniquely high resistances to both false-negative and false-positive detection events by comparison to other open-source adipocyte analysis software. They also enable superior measurement of adipocyte morphology, especially in the case of low-magnification images. AT-AT features several quality-of-life features, including automatic summarization of batch-collected data, the implementation of its thresholding algorithms as independent ImageJ commands, and an optional histogram builder. AT-AT can enhance the rigor and reproducibility of adipose tissue research. AT-AT will become available as a plugin for ImageJ in the fall of 2022. Presentation: Sunday, June 12, 2022 12:30 p.m. - 2:30 p.m.

CERAD-NAB and flexible battery based neuropsychological differentiation of Alzheimer’s dementia and depression using machine learning approaches

ABSTRACT Depression (DEP) and dementia of the Alzheimer’s type (DAT) represent the most common neuropsychiatric disorders in elderly patients. Accurate differential diagnosis is indispensable to ensure appropriate treatment. However, DEP can yet mimic cognitive symptoms of DAT and patients with DAT often also present with depressive symptoms, impeding correct diagnosis. Machine learning (ML) approaches could eventually improve this discrimination using neuropsychological test data, but evidence is still missing. We therefore employed Support Vector Machine (SVM), Naïve Bayes (NB), Random Forest (RF) and conventional Logistic Regression (LR) to retrospectively predict the diagnoses of 189 elderly patients (68 DEP and 121 DAT) based on either the well-established Consortium to Establish a Registry for Alzheimer’s Disease neuropsychological assessment battery (CERAD-NAB) or a flexible battery approach (FLEXBAT). The best performing combination consisted of FLEXBAT and NB, correctly classifying 87.0% of patients as either DAT or DEP. However, all accuracies were similar across algorithms and test batteries (83.0% – 87.0%). Accordingly, our study is the first to show that common ML algorithms with their default parameters can accurately differentiate between patients clinically diagnosed with DAT or DEP using neuropsychological test data, but do not necessarily outperform conventional LR.

Drastic reduction of false positive species in samples of insects by intersecting the default output of two popular metagenomic classifiers.

The use of high-throughput sequencing to recover short DNA reads of many species has been widely applied on biodiversity studies, either as amplicon metabarcoding or shotgun metagenomics. These reads are assigned to taxa using classifiers. However, for different reasons, the results often contain many false positives. Here we focus on the reduction of false positive species attributable to the classifiers. We benchmarked two popular classifiers, BLASTn followed by MEGAN6 (BM) and Kraken2 (K2), to analyse shotgun sequenced artificial single-species samples of insects. To reduce the number of misclassified reads, we combined the output of the two classifiers in two different ways: (1) by keeping only the reads that were attributed to the same species by both classifiers (intersection approach); and (2) by keeping the reads assigned to some species by any classifier (union approach). In addition, we applied an analytical detection limit to further reduce the number of false positives species. As expected, both metagenomic classifiers used with default parameters generated an unacceptably high number of misidentified species (tens with BM, hundreds with K2). The false positive species were not necessarily phylogenetically close, as some of them belonged to different orders of insects. The union approach failed to reduce the number of false positives, but the intersection approach got rid of most of them. The addition of an analytic detection limit of 0.001 further reduced the number to ca. 0.5 false positive species per sample. The misidentification of species by most classifiers hampers the confidence of the DNA-based methods for assessing the biodiversity of biological samples. Our approach to alleviate the problem is straightforward and significantly reduced the number of reported false positive species.

A Model for Gastrointestinal Tract Motility in a 4D Imaging Phantom of Human Anatomy

<h3>Purpose/Objective(s)</h3> Gastrointestinal (GI) tract motility is one of the main sources for intra/inter-fraction variability and uncertainty in validating deformable image registration (DIR) algorithms. We present an implementation of four modes of GI motility in the 4D extended cardiac-torso (XCAT) digital phantom. <h3>Materials/Methods</h3> We consider motility modes that exhibit large amplitude changes in the diameter of the GI tract and may persist over timescales comparable to online adaptive planning and radiotherapy delivery. The following modes are modeled: peristalsis, rhythmic segmentation, high amplitude propagating contractions (HAPC), and tonic contractions. The models are intended for the organ of origin (i.e., peristalsis in esophagus and small intestine; HAPCs in colon) but can be applied to any user-specified GI tract organ. Peristalsis and rhythmic segmentations are modeled by traveling and standing sinusoidal waves. HAPCs and tonic contractions are modeled by traveling and stationary Gaussian waves. Spatial and temporal dispersion are implemented by exponential functions. The model is applied at the control points of the nonuniform rational B-spline (NURBS) surfaces defined in the reference XCAT phantom. The reference NURBS surfaces are dynamically updated at each timepoint based on the user-defined parameters shown in Table 1. GI motility is combined with the respiratory and cardiac motion available in XCAT. Cine MRI acquisitions from 50 fractions of 10 patients treated in a 1.5 T MR-Linac were analyzed to estimate default parameters for each model. <h3>Results</h3> We demonstrate the ability to generate realistic 4D MRI images that simulate four modes of GI motility combined with respiratory and cardiac motion. All modes of motility, except tonic contractions, were observed in the analysis of our cine MRI acquisitions. Peristalsis was the most common. Default parameters estimated from cine MRI data or from literature (if not observed in our measurements) are used as initial values for simulation experiments. <h3>Conclusion</h3> The digital phantom provides realistic models to aid in multimodality imaging research. The addition of GI motility will further contribute to the development, testing, and validation of DIR algorithms for MR-guided radiotherapy.

Individualized Short-Term Electric Load Forecasting Using Data-Driven Meta-Heuristic Method Based on LSTM Network.

Short-term load forecasting is viewed as one promising technology for demand prediction under the most critical inputs for the promising arrangement of power plant units. Thus, it is imperative to present new incentive methods to motivate such power system operations for electricity management. This paper proposes an approach for short-term electric load forecasting using long short-term memory networks and an improved sine cosine algorithm called MetaREC. First, using long short-term memory networks for a special kind of recurrent neural network, the dispatching commands have the characteristics of storing and transmitting both long-term and short-term memories. Next, four important parameters are determined using the sine cosine algorithm base on a logistic chaos operator and multilevel modulation factor to overcome the inaccuracy of long short-term memory networks prediction, in terms of the manual selection of parameter values. Moreover, the performance of the MetaREC method outperforms others with regard to convergence accuracy and convergence speed on a variety of test functions. Finally, our analysis is extended to the scenario of the MetaREC_long short-term memory with back propagation neural network, long short-term memory networks with default parameters, long short-term memory networks with the conventional sine-cosine algorithm, and long short-term memory networks with whale optimization for power load forecasting on a real electric load dataset. Simulation results demonstrate that the multiple forecasts with MetaREC_long short-term memory can effectively incentivize the high accuracy and stability for short-term power load forecasting.