pNPs-CapsNet: Predicting Neuropeptides Using Protein Language Models and FastText Encoding-Based Weighted Multi-View Feature Integration with Deep Capsule Neural Network

ACS Omega, Journal Year: 2025, Volume and Issue: unknown

Published: March 18, 2025

Neuropeptides (NPs) are critical signaling molecules that essential in numerous physiological processes and possess significant therapeutic potential. Computational prediction of NPs has emerged as a promising alternative to traditional experimental methods, often labor-intensive, time-consuming, expensive. Recent advancements computational peptide models provide cost-effective approach identifying NPs, characterized by high selectivity toward target cells minimal side effects. In this study, we propose novel deep capsule neural network-based model, namely pNPs-CapsNet, predict non-NPs accurately. Input samples numerically encoded using pretrained protein language models, including ESM, ProtBERT-BFD, ProtT5, extract attention mechanism-based contextual semantic features. A differential evolution-based weighted feature integration method is utilized construct multiview vector. Additionally, two-tier selection strategy, comprising MRMD SHAP analysis, developed identify select optimal Finally, the network (CapsNet) trained selected set. The proposed pNPs-CapsNet model achieved remarkable predictive accuracy 98.10% an AUC 0.98. To validate generalization capability independent reported 95.21% 0.96. outperforms existing state-of-the-art demonstrating 4% 2.5% improved for training data sets, respectively. demonstrated efficacy consistency underline its potential valuable robust tool advancing drug discovery academic research.

Language: Английский

---

An optimized transformer model for efficient detection of thoracic diseases in chest X-rays with multi-scale feature fusion

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(5), P. e0323239 - e0323239

Published: May 7, 2025

This study presents the development and application of an optimized Detection Transformer (DETR) model, known as CD-DETR, for detection thoracic diseases from chest X-ray (CXR) images. The CD-DETR model addresses challenges detecting minor pathologies in CXRs, particularly regions with uneven medical resource distribution. In central western China, due to a shortage radiologists, CXRs township hospitals are concentrated diagnosis. requires processing large number short period time obtain results. integrates multi-scale feature fusion approach, leveraging Efficient Channel Attention (ECA-Net) Spatial Upsampling (SAU) enhance representation improve accuracy. It also introduces dedicated Chest Diseases Intersection over Union (CDIoU) loss function optimize small targets reduce class imbalance. Experimental results on NIH dataset demonstrate that achieves precision 88.3% recall 86.6%, outperforming other DETR variants by average 5% CNN-based models like YOLOv7 6–8% these metrics, showing its potential practical imaging diagnostics.

Language: Английский

---

Illicit uptake of peptide-based antibiotics through bacterial peptide transporters: an approach towards overcoming drug resistance

Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 193, P. 110444 - 110444

Published: May 27, 2025

Language: Английский

---