Anticancer peptides (ACPs) have emerged as promising therapeutic candidates for cancer treatment due to their high efficacy and low propensity for inducing drug resistance. However, existing ACP identification methods primarily rely on peptide sequence features while neglecting spatial structural characteristics. Moreover, few approaches can simultaneously predict the functional activity of ACPs. To address these limitations, this study proposes Multi-ACPNet, a novel dual-function predictor capable of both ACP identification and activity type classification. This model innovatively integrates sequence and structural features through a multi-stage framework. It employs a hybrid Bidirectional Long Short-Term Memory (BiLSTM) and causal convolutional network to capture both long-range dependencies and local sequence patterns, followed by a multi-scale Graph Convolutional Network (GCN) that dynamically fuses local and long-range structural dependencies using residual connections and adaptive weighting. Experimental results demonstrate that Multi-ACPNet achieves outstanding performance, with Accuracy of 0.8140, 0.9536, and 0.8770 on three benchmark datasets for ACP identification. For functional prediction, it attains an AUC of 0.9033, F1-score of 0.8472, and Hamming loss of 0.1303, significantly outperforming state-of-the-art predictors.

Somatic variants play a crucial role in the occurrence and progression of cancer. However, in the absence of matched normal controls, distinguishing between germline and somatic variants becomes challenging in tumor samples. The existing tumor-only genomic analysis methods either suffer from limited performance or insufficient interpretability due to an excess of features. Therefore, there is an urgent need for an alternative approach that can address these issues and have practical implications. Here, we presented OncoTOP, a computational method for genomic analysis without matched normal samples, which can accurately distinguish somatic mutations from germline variants. Reference sample analysis revealed a 0% false positive rate and 99.7% reproducibility for variant calling. Assessing 2864 tumor samples across 18 cancer types yielded a 99.8% overall positive percent agreement and a 99.9% positive predictive value. OncoTOP can also accurately detect clinically actionable variants and subclonal mutations associated with drug resistance. For the prediction of mutation origins, the positive percent agreement stood at 97.4% for predicting somatic mutations and 95.7% for germline mutations. High consistency of tumor mutational burden (TMB) was observed between the results generated by OncoTOP and tumor-normal paired analysis. In a cohort of 97 lung cancer patients treated with immunotherapy, TMB-high patients had prolonged PFS (P = .02), proving the reliability of our approach in estimating TMB to predict therapy response. Furthermore, microsatellite instability status showed a strong concordance (97%) with polymerase chain reaction results, and leukocyte antigens class I subtypes and homozygosity achieved an impressive concordance rate of 99.3% and 99.9% respectively, compared to its tumor-normal paired analysis. Thus, OncoTOP exhibited high reliability in variant calling, mutation origin prediction, and biomarker estimation. Its application will promise substantial advantages for clinical genomic testing.

MOTIVATION: Proteins play pivotal roles in biological systems, and precise prediction of their functions is indispensable for practical applications. Despite the surge in protein sequence data facilitated by high-throughput techniques, unraveling the exact functionalities of proteins still demands considerable time and resources. Currently, numerous methods rely on protein sequences for prediction, while methods targeting protein structures are scarce, often employing convolutional neural networks (CNN) or graph convolutional networks (GCNs) individually. RESULTS: To address these challenges, our approach starts from protein structures and proposes a method that combines CNN and GCN into a unified framework called the two-model adaptive weight fusion network (TAWFN) for protein function prediction. First, amino acid contact maps and sequences are extracted from the protein structure. Then, the sequence is used to generate one-hot encoded features and deep semantic features. These features, along with the constructed graph, are fed into the adaptive graph convolutional networks (AGCN) module and the multi-layer convolutional neural network (MCNN) module as needed, resulting in preliminary classification outcomes. Finally, the preliminary classification results are inputted into the adaptive weight computation network, where adaptive weights are calculated to fuse the initial predictions from both networks, yielding the final prediction result. To evaluate the effectiveness of our method, experiments were conducted on the PDBset and AFset datasets. For molecular function, biological process, and cellular component tasks, TAWFN achieved area under the precision-recall curve (AUPR) values of 0.718, 0.385, and 0.488 respectively, with corresponding Fmax scores of 0.762, 0.628, and 0.693, and Smin scores of 0.326, 0.483, and 0.454. The experimental results demonstrate that TAWFN exhibits promising performance, outperforming existing methods. AVAILABILITY AND IMPLEMENTATION: The TAWFN source code can be found at: https://github.com/ss0830/TAWFN.

The use of deep learning methods to decode visual perception images from brain activity recorded by fMRI has received a lot of attention. However, limited fMRI data make the task of visual reconstruction challenging. Inspired by hierarchical encoding of the visual cortex and the theory of brain homology with convolutional neural networks (CNNs), we propose a novel neural decoding model called hierarchical semantic generative adversarial network (HS-GAN). Specifically, we use CNN-based image encoder to extract hierarchical and semantic features of visually stimulus images. Then a neural decoder is used to decode hierarchical and semantic features from fMRI. In order to take full advantage of the information from different visual cortexes, we construct a generator with self-attention modules and skip connections to fuse the image features of different layers. In model training, adversarial learning is introduced to realize more natural image reconstruction. Compared to existing advanced methods, our method significantly improves the naturalness and fidelity of reconstructed images.

P2P online lending platforms have been developing rapidly worldwide in recent years. This study seeks to explain platform users’ behaviour in different cultures. Specifically, this research aims to investigate the sentiment polarity of the posts in P2P online lending platforms and how it influences the quantity and quality of user reviews. Over 3000 online posts were collected from large online platforms in China and U.S. The findings indicate that in U.S. P2P online lending platforms, there are more reviews for negative-sentiment posts than those for positive-sentiment posts. In a collective culture such as China, there are more reviews in P2P online lending platforms for positive-sentiment posts, and platform participants tend to post strong sentiment information.

We propose an end-to-end learning method based on statistical features extracted on set-of-samples level as a step toward solving the writer verification problem which is about deciding whether two handwriting sources are identical given handwriting samples from the two sources. The set-of-samples features are extracted on top of single sample features. Single sample features are traditionally learned using sample-to-sample comparison similarity learning. In this paper, we learn it as a sub-module of an end-to-end two-sets-of-samples comparison similarity learning. We compare human-engineered (GSC features) single sample features and automatically learned features using convolutional neural networks (CNN) and find the latter performs better with abundant training data and data augmentation. The statistical features on sets of samples capture both inner-writer variabilities and intra-writer variabilities. Experiments are conducted on frequently occurring words and digraphs such as "and" and "th" from around 1500 writers. We perform experiments on pre-training using sample-to-sample similarity learning and end-to-end fine-tuning. The results show that two-sets-of-samples comparison gives much better accuracy than sample-to-sample comparison. In addition, the end-to-end training based on parametric statistical features gives better accuracy than standard distribution comparison tests such as the K-S test based on distance space.

This paper presented a novel application of Magneto encephalography (MEG) and diffusion tensor image (DTI) on word recognition, in which the spatiotemporal signature and the neural network of brain activation associated with word recognition were investigated. The word stimuli consisted of matched and mismatched words, which were visually and acoustically presented simultaneously. Twenty participants were recruited to distinguish and gave different reactions to these two types of stimuli. The neural activations caused by their reactions were recorded by MEG system and 3T magnetic DTI scanner. Virtual sensor technique and wavelet beam former source analysis, which were state-of-the-art methods, were used to study the MEG and DTI data. Three responses were evoked in the MEG waveform and M160 was identified in the left temporal-occipital junction. All the results coincided with the previous studies' conclusions, which indicated that the integration of virtual sensor and wavelet beam former were effective techniques in analyzing the MEG and DTI data.