Estimation of trapezoidal-shaped overlapping nuclear pulse parameters based on a deep learning CNN-LSTM model.

The Long Short-Term Memory neural network (LSTM) has excellent learning ability for the time series of the nuclear pulse signal. It can accurately estimate the parameters (such as amplitude, time constant, etc.) of the digitally shaped nuclear pulse signal (especially the overlapping pulse signal). However, due to the large number of pulse sequences, the direct use of these sequences as samples to train the LSTM increases the complexity of the network, resulting in a lower training efficiency of the model. The convolution neural network (CNN) can effectively extract the sequence samples by using its unique convolution kernel structure, thus greatly reducing the number of sequence samples. Therefore, the CNN-LSTM deep neural network is used to estimate the parameters of overlapping pulse signals after digital trapezoidal shaping of exponential signals. Firstly, the estimation of the trapezoidal overlapping nuclear pulse is considered to be obtained after the superposition of multiple exponential nuclear pulses followed by trapezoidal shaping. Then, a data set containing multiple samples is set up; each sample is composed of the sequence of sampling values of the trapezoidal overlapping nuclear pulse and the set of shaping parameters of the exponential pulse before digital shaping. Secondly, the CNN is used to extract the abstract features of the training set in these samples, and then these abstract features are applied to the training of the LSTM model. In the training process, the pulse parameter set estimated by the present neural network is calculated by forward propagation. Thirdly, the loss function is used to calculate the loss value between the estimated pulse parameter set and the actual pulse parameter set. Finally, a gradient-based optimization algorithm is applied to update the weight by getting back the loss value together with the gradient of the loss function to the network, so as to realize the purpose of training the network. After model training was completed, the sampled values of the trapezoidal overlapping nuclear pulse were used as input to the CNN-LSTM model to obtain the required parameter set from the output of the CNN-LSTM model. The experimental results show that this method can effectively overcome the shortcomings of local convergence of traditional methods and greatly save the time of model training. At the same time, it can accurately estimate multiple trapezoidal overlapping pulses due to the wide width of the flat top, thus realizing the optimal estimation of nuclear pulse parameters in a global sense, which is a good pulse parameter estimation method.

An essential role of acetyl coenzyme A in the catalytic cycle of insect arylalkylamine N-acetyltransferase.

Acetyl coenzyme A (Ac-CoA)-dependent N-acetylation is performed by arylalkylamine N-acetyltransferase (AANAT) and is important in many biofunctions. AANAT catalyzes N-acetylation through an ordered sequential mechanism in which cofactor (Ac-CoA) binds first, with substrate binding afterward. No ternary structure containing AANAT, cofactor, and substrate was determined, meaning the details of substrate binding and product release remain unclear. Here, two ternary complexes of dopamine N-acetyltransferase (Dat) before and after N-acetylation were solved at 1.28 Å and 1.36 Å resolution, respectively. Combined with the structures of Dat in apo form and Ac-CoA bound form, we addressed each stage in the catalytic cycle. Isothermal titration calorimetry (ITC), crystallography, and nuclear magnetic resonance spectroscopy (NMR) were utilized to analyze the product release. Our data revealed that Ac-CoA regulates the conformational properties of Dat to form the catalytic site and substrate binding pocket, while the release of products is facilitated by the binding of new Ac-CoA.

Rising time restoration for nuclear pulse using a mathematic model.

The rising time of a nuclear pulse is slowed before being digitized because of the effect of distributed capacitance and resistance. This results in the waveform distortion of a shaped pulse. In this study, the effect of distributed capacitance and resistance is equivalent to the result of RC network. The mathematical model of the network is established to restore the rising time of the input nuclear pulse. Experimental results show that the leading edge of the nuclear pulse becomes steep after rising time restoration, and the shape of the shaped pulse is also improved. The energy spectrum obtained with rising time restoration is compared with that without rising time restoration. The comparison result indicates that using rising time restoration can extend the measurement range of pulse amplitude without affecting the energy resolution of the system.

[Decomposition of X-Ray Fluorescence Overlapping Peaks Based on Statistical and Genetic Algorithms].

In fluorescence analysis, the phenomenon of overlapping often occurs among adjacent peaks. In the view of the random physical properties of formation process of X fluorescence spectra, Gaussian Mixture Statistics Model (GMSM) and Genetic Algorithms were used for the decomposition of overlapping peaks. First, the GMSM was proposed to describe the overlapping peaks, and the local convergence problem of expectation maximization (EM) was analyzed. Secondly, the GMSM parameters were regarded as individual genes, and the log-likelihood function of overlapping peaks random data was set as fitness function. A fast algorithm for the objective function value was proposed. Finally, the population search technology of Genetic Algorithm was used to find the global optimal solution, and to realize the decomposition of overlapping peaks. All measured data were regarded as "useful" data. The "useful" degree was reflected by their probability. The GMSM method can achieve the "best match" effect in the maximum global probability with zero loss of original data, which can fit the random of radiation measurement process. The decomposition experiments of four serious overlapping peaks show high precision of the peak position, peak area and standard deviation. The maximum error was 0.7 channel, 2.3% and 2.17%, respectively, which is especially suitable for the condition of serious overlap and can be widely used for the decomposition of other energy spectrum.