Predictor Selection for CNN-based Statistical Downscaling of Monthly Precipitation

Convolutional neural networks(CNNs) have been widely studied and found to obtain favorable results in statistical downscaling to derive high-resolution climate variables from large-scale coarse general circulation models(GCMs).However, there is a lack of research exploring the predictor selection for CNN modeling. This paper presents an effective and efficient greedy elimination algorithm to address this problem. The algorithm has three main steps: predictor importance attribution, predictor removal, and CNN retraining, which are performed sequentially and iteratively. The importance of individual predictors is measured by a gradient-based importance metric computed by a CNN backpropagation technique, which was initially proposed for CNN interpretation. The algorithm is tested on the CNN-based statistical downscaling of monthly precipitation with 20 candidate predictors and compared with a correlation analysisbased approach. Linear models are implemented as benchmarks. The experiments illustrate that the predictor selection solution can reduce the number of input predictors by more than half, improve the accuracy of both linear and CNN models,and outperform the correlation analysis method. Although the RMSE(root-mean-square error) is reduced by only 0.8%,only 9 out of 20 predictors are used to build the CNN, and the FLOPs(Floating Point Operations) decrease by 20.4%. The results imply that the algorithm can find subset predictors that correlate more to the monthly precipitation of the target area and seasons in a nonlinear way. It is worth mentioning that the algorithm is compatible with other CNN models with stacked variables as input and has the potential for nonlinear correlation predictor selection.

An anisotropic Chebyshev descriptor and its optimization for deformable shape correspondence

Shape descriptors have recently gained popularity in shape matching,statistical shape modeling,etc.Their discriminative ability and efficiency play a decisive role in these tasks.In this paper,we first propose a novel handcrafted anisotropic spectral descriptor using Chebyshev polynomials,called the anisotropic Chebyshev descriptor(ACD);it can effectively capture shape features in multiple directions.The ACD inherits many good characteristics of spectral descriptors,such as being intrinsic,robust to changes in surface discretization,etc.Furthermore,due to the orthogonality of Chebyshev polynomials,the ACD is compact and can disambiguate intrinsic symmetry sinces everal directions are considered.To improve the ACD’s discrimination ability,we construct a Chebyshev spectral manifold convolutional neural network(CSMCNN)that optimizes the ACD and produces a learned ACD.Our experimental results show that the ACD outperforms existing state-of-the-art handcrafted descriptors.The combination of the ACD and the CSMCNN is better than other state-of-the-art learned descriptors in terms of discrimination,efficiency,and robustness to changes in shape resolution and discretization.

Initial 24-h perfusion index of ICU admission is associated with acutekidney injury in perioperative critically ill patients: A retrospective cohortanalysis

Background:The relationship between perfusion index(PI)and organ dysfunction in patients in the intensivecare unit(ICU)is not clear.This study aimed to explore the relationship between PI and renal function in theperioperative critical care setting and evaluate the predictive efficiency of PI on patients with acute kidney injury(AKI)in the ICU.Methods:This retrospective analysis involved 12,979 patients who had undergone an operation and were admitted to the ICU in Peking Union Medical College Hospital from January 2014 to December 2019.The distributionof average PI in the first 24 h after ICU admission and its correlation with AKI was calculated by Cox regression.Receiver operating characteristic(ROC)curves were generated to compare the ability of PI,mean arterial pressure(MAP),creatinine,blood urea nitrogen(BUN),and central venous pressure(CVP)to discriminate AKI in thefirst 48 h in all perioperative critically ill patients.Results:Average PI in the first 24 h served as an independent protective factor of AKI(Odds ratio[OR]=0.786,95%confidence interval[CI]:0.704–0.873,P<0.0001).With a decrease in PI by one unit,the incidence of AKIincreased 1.74 times.Among the variables explored for the prediction of AKI(PI,MAP,creatine,BUN,and CVP),PI yielded the highest area under the ROC curve,with a sensitivity of 64.34%and specificity of 70.14%.A cut-offvalue of PI≤2.12 could be used to predict AKI according to the Youden index.Moreover,patients in the low PIgroup(PI≤2.12)exhibited a marked creatine elevation at 24–48 h with a slower decrease compared with thosein the high PI group(PI>2.12).Conclusions:As a local blood flow indicator,the initial 24-h average PI for perioperative critically ill patients canpredict AKI during their first 120 h in the ICU.

Citrus Pan-Genome to Breeding Database(CPBD):A comprehensive genome database for citrus breeding

Dear Editor,The genus Citrus comprises more than 30 species worldwide(Swingle,1943).However,citrus genetics and traditional breeding are hindered due to asexual reproduction,long generation time,and empirical utilization of the germplasm.The whole-genome information of different citrus species accelerates the genetic studies and improves the breeding efficiency by high-density markers.

Hypocoagulation induced by broadspectrum antibiotics in extensive burn patients

Background:Patients with extensive burns usually develop pro-coagulation soon after the injury if there is no sepsis occurred.We describe the case of an extensive burn adult suffering from hypocoagulation not related to sepsis,but secondary to antibiotic treatment.Case presentation:Here,we report a case of an adult male patient suffering from flame burns of 45%total body surface area(40%full thickness)combined with inhalation injury.Hypocoagulopathy with soaring prolonged activated partial thromboplastin time value occurred on third week post-burn while systemic infection had been under control by application of broad-spectrum antibiotics.Investigations showed that not the infection but vitamin K-related coagulation factor deficiency were responsible for unexpected bleeding.However,supplemental vitamin K was not the key as we expected,which prompted us trying to decode the underlying cause of coagulation disturbance in this patient and pick out the most effective treatment for live-saving.After the withdrawal of highly suspected broad-spectrum antibiotic,Meropenem^(■),disturbed vitamin K related coagulation factors gradually restored to their optimal levels so as to maintain normal coagulation status.Therefore,surgical procedures without further risk of bleeding could be carried out in time for wound recovery.The patient was discharged on post-burn day 67 and transferred to a secondary hospital for his rehabilitation.Conclusion:Hypocoagulopathy may be devoted to different reasons other than sepsis in extensive burns.Early recognition of the cause for coagulation disturbance is critical to make appropriate treatment and save patients’lives.This case illustrated the importance of unveiling the mist cause for coagulation disturbance occurred in extensive burn patient,which paved the way for optimal life-saving treatments.And we also recommend burn surgeons to be alerted to antibiotic-induced vitamin K deficiency-related coagulopathy among critical burn patients.

The study of the correlated X-ray scattering(CXS)and its data analysis

Background At present,it is insufficient to understand the basic data characteristics of the correlated X-ray scattering.And there is a great challenge about how to master the nature of the data.So it is difficult to use and analyze the experimental data more effectively.In addition,there are many reasons,for the experimental artifacts such as whether the shutter is on or off,whether there is the beam line or not,the swaying of the nozzle and the shadow of the detector.So it is rather challenging to analyze the scattering patterns.Purpose The purpose of this paper was to develop a method to filter the invalid scattering data and provide the theoretical and experiment fundamentals for studying the X-ray scattering data of the complex biological sample further.Methods The heliummolecules were scattered by the X-ray free-electron laser in Spring8 in Japan.Andmillions of scattering patterns were obtained from the X-ray free-electron laser experiment.Through the analysis of the scattering data,the sum,mean,median and variance of the scattering intensity were obtained.Then different clusters were obtained with the densitybased spatial clustering of applications with noise(DBSCAN)algorithm.Results Based on the DBSCAN,some of the scattering patterns with high artifacts were removed and different clusters were clarified.So the experimental scattering data could be analyzed more effectively.Conclusion The theoretical and experiment fundamentals for comprehensively studying the X-ray scattering data of the complex biological sample were provided.After the data filtering,the angular autocorrelation of different clusters with Kam’s method will be computed and analyzed effectively.

Field-portable ratiometric fluorescence imaging of dual-color label-free carbon dots for uranyl ions detection with cellphone-based optical platform

Under the public spotlight,uranyl(UO22+)ions has attracted considerable attention for the extreme radioactive and chemical toxicity to ourselves and our environment.Herein,we present a simple and effective ratiometric fluorescence imaging method for the visualizing and quantitative detection UO22+ions by cellphone-based optical platform.The sensing solution was prepared by mixing label-free red carbon dots(r-CDs)and blue carbon dots(b-CDs)together with a fixed photoluminescence intensity ratio of 4:1.When UO22+ions were added,the fluorescence of r-CDs can be selectively quenched,while the fluorescence of b-CDs remains stable without spectral cha nges.With the gradually increase the amounts of UO22+ions,the different response of dual-color CDs resulted in a signification color evolution from deep red to dark purple under the ultraviolet(UV)light illumination.Then,a cellphone-based optical platform was constructed for directly imaging the color change of the samples,and the built-in Colorpicker APP quickly output the red,green and blue(RGB)channel values of these images within one second.Interesting,there was a linear relationship between the ratio of red and blue(R/B)channel values and UO22+ions concentration from 0μmol/L to 30.0μmol/L(R^2=0.92804)with the detection limit of^8.15μmol/L(signal-to-noise ratio of 3).In addition,the optical platform has also been applied to the quantification of UO22+ions in tap water and river water sample.With the advantage of low-cost,portable,easy to operation,we anticipate that this method would greatly improve the accessibility of UO22+ions detection even in resource-limited areas.