DNA-protein interactions exert the fundamental structure of many pivotal biological processes, such as DNA replication, transcription, and gene regulation. However, accurate and efficient computational methods for identifying these interactions are still lacking. In this study, we propose a method ESM-DBP through refining the DNA-binding protein sequence repertory and domain-adaptive pretraining based the general protein language model. Our method considers the lacking exploration of general language model for DNA-binding protein domain-specific knowledge, so we screen out 170,264 DNA-binding protein sequences to construct the domain-adaptive language model. Experimental results on four downstream tasks show that ESM-DBP provides a better feature representation of DNA-binding protein compared to the original language model, resulting in improved prediction performance and outperforming the state-of-the-art methods. Moreover, ESM-DBP can still perform well even for those sequences with only a few homologous sequences. ChIP-seq on two predicted cases further support the validity of the proposed method.

With a noisy environment caused by fluorescence and additive white noise as well as complicated spectra fingerprints, the identification of complex mixture materials remains a significant challenge in Raman spectroscopy application. This paper proposes a new scheme based on a constant wavelet transform (CWT) and a deep network for classifying complex mixture. The scheme first transforms the noisy Raman spectra to a two-dimensional scale map using CWT. Then, the multi-label deep neural network model (MLDNN) is simulated for material classification. The proposed model accelerates the feature extraction and expands the feature graph using the global averaging pooling layer. The Sigmoid function is implemented in the last layer of the model. The MLDNN model was trained, validated, and tested with data collected from the samples prepared from palm oil substances. During the training and validating process, data augmentation is applied to overcome the imbalance of data and enrich Raman spectra's diversity. From the test results, MLDNN successfully surpassed VGG16, VGG19, ResNet50, DenseNet121, InceptionResNetV2, MobileNet50 in average precision, F1 macro averaging and F1 micro averaging, which were 0.9900, 0.9874 and 0.9874, respectively. In Hamming loss, one error, coverage, ranking loss, the MLDNN model reached 0.014, 0.020, 1.738 and 0.000, respectively. The values of other models in Hamming loss, one error, coverage, ranking loss are higher than those of MLDNN. In the comparison of AUC value, MLDNN outperforms other models on each label. The average detection time of the MLDNN is 5.3123 s, which is much faster than those of VGG16, VGG19, ResNet50, DenseNet121, InceptionResNetV2 and MobileNetV2, which are 7.1245, 7.5046, 8.6300, 12.3294, 16.1131 and 6.6451, respectively.

Hybrid beamforming can provide rapid data transmission rates while reducing the complexity and cost of massive multiple-input multiple-output (MIMO) systems. However, channel state information (CSI) is imperfect in realistic downlink channels, introducing challenges to hybrid beamforming (HBF) design. This paper proposes an unsupervised deep learning neural network (USDNN) for hybrid beamforming to prevent the labeling overhead of supervised learning and improve the achievable sum rate based on imperfect CSI. The simulation results show that our proposed method is 74% better than MO and 120% better than orthogonal match pursuit (OMP) systems; our proposed USDNN can achieve near-optimal performance.

Because it is relatively difficult in practice to classify the Raman spectrum under baseline noise and additive white Gaussian noise environments, this paper proposes a new framework based on a wavelet transform and deep neural network for identification of noisy Raman spectra. The framework consists of two main engines. Wavelet transform is proposed as the framework front end for transforming the 1-D noise Raman spectrum to two-dimensional data. The two-dimensional data are fed to the framework back end, which is a classifier. The optimum classifier is chosen by implementing several traditional machine learning (ML) and deep learning (DL) algorithms, and we investigate their classification accuracy and robustness performances. The four chosen MLs are naive Bayes (NB), a support vector machine (SVM), a random forest (RF) and a k-nearest neighbor (KNN), and a deep convolution neural network (DCNN) was chosen as a DL classifier. Noise-free, Gaussian noise, baseline noise, and mixed-noise Raman spectra were applied to train and validate the ML and DCNN models. The optimum back-end classifier was obtained by testing the ML and DCNN models with several noisy Raman spectra (10-30 dB noise power). Based on the simulation, the accuracy of the DCNN classifier is 9% higher than that of the NB classifier, 3.5% higher than the RF classifier, 1% higher than the KNN classifier, and 0.5% higher than the SVM classifier. In terms of robustness to mixed noise scenarios, the framework with the DCNN back end showed superior performance compared with the other ML back ends. The DCNN back end achieved 90% accuracy at 3 dB SNR, while the NB, SVM, RF, and K-NN back ends required 27 dB, 22 dB, 27 dB, and 23 dB SNR, respectively. In addition, in the low-noise test dataset, the F-measure score of the DCNN back end exceeded 99.1%, and the F-measure scores of the other ML engines were below 98.7%.