Fast and accurate identification of the peptides with anticancer activity potential from large-scale proteins is currently a challenging task. In this study, we propose a new machine learning predictor, namely, ACPred-Fuse, that can automatically and accurately predict protein sequences with or without anticancer activity in peptide form. Specifically, we establish a feature representation learning model that can explore class and probabilistic information embedded in anticancer peptides (ACPs) by integrating a total of 29 different sequence-based feature descriptors. In order to make full use of various multiview information, we further fused the class and probabilistic features with handcrafted sequential features and then optimized the representation ability of the multiview features, which are ultimately used as input for training our prediction model. By comparing the multiview features and existing feature descriptors, we demonstrate that the fused multiview features have more discriminative ability to capture the characteristics of ACPs. In addition, the information from different views is complementary for the performance improvement. Finally, our benchmarking comparison results showed that the proposed ACPred-Fuse is more precise and promising in the identification of ACPs than existing predictors. To facilitate the use of the proposed predictor, we built a web server, which is now freely available via http://server.malab.cn/ACPred-Fuse.

The detectability of peptides is fundamentally important in shotgun proteomics experiments. At present, there are many computational methods to predict the detectability of peptides based on sequential composition or physicochemical properties, but they all have various shortcomings. Here, we present PepFormer, a novel end-to-end Siamese network coupled with a hybrid architecture of a Transformer and gated recurrent units that is able to predict the peptide detectability based on peptide sequences only. Specially, we, for the first time, use contrastive learning and construct a new loss function for model training, greatly improving the generalization ability of our predictive model. Comparative results demonstrate that our model performs significantly better than state-of-the-art methods on benchmark data sets in two species (Homo sapiens and Mus musculus). To make the model more interpretable, we further investigate the embedded representations of peptide sequences automatically learnt from our model, and the visualization results indicate that our model can efficiently capture high-latent discriminative information, improving the predictive performance. In addition, our model shows a strong ability of cross-species transfer learning and adaptability, demonstrating that it has great potential in robust prediction of peptides detectability on different species. The source code of our proposed method can be found via .

Electromyography (EMG) signal can be defined as a measure of electrical activity produced by skeletal muscles. It can be used in handling electronic devices or prosthesis. If we are able recognize the hand gesture captured using EMG signal with greater reliability and classification rate, it could serve a good purpose for handling the prosthesis and to provide the good quality of life to amputees and disabled people. In this paper, we have worked on recognizing the 9 classes of individual and combined finger movement captured using 2 channel EMG sensor. We have used two different classification techniques such as Artificial Neural Network (ANN), and k-nearest neighbors (KNN), to classify the test samples. Seven time domain features a) Mean absolute value, b) root mean square, c) variance, d) waveform length, e) number of zero crossing, f) complexity, g) mobility have been used to uniquely represent the EMG channel data. Tuning parameters like number of hidden layers, learning constant and number of neighbors have been determined from the experimental results to achieve the better classification results. Classification accuracy has been selected as a metric to evaluate the performance of each classifier.

Anticancer peptide (ACP) is a class of anti-cancer peptide which can inhibit and kill tumor cells. Identification of ACPs is of great significance for the development of new anti-cancer drugs. However, most of computational methods make predictions based on machine learning using hand-crafted features. In this paper, we propose a new graph learning based computational model, named ACP-GCN, to automatically and accurately predict ACPs based on graph convolution networks. In this model, we for the first time take the ACP prediction as a graph classification task, where each peptide sample is represented as a graph. The experimental results show that the proposed model outperforms most of state-of-the-art methods. Specifically, the specificity and accuracy of our predictor have reached 91.1% and 85.6%, which are 2.8-22.8% and 3.2-15% higher than that of other existing predictors, respectively, demonstrating that the proposed method can effectively distinguish ACPs from non-ACPs. The excellent predictive ability will rapidly push forward their applications in cancer therapy. INDEX TERMS anti-cancer peptides, graph convolution networks, machine learning, prediction methods