Integrated convolution and self-attention for improving peptide toxicity prediction

Bioinformatics, Год журнала: 2024, Номер 40(5)

Опубликована: Май 1, 2024

Abstract Motivation Peptides are promising agents for the treatment of a variety diseases due to their specificity and efficacy. However, development peptide-based drugs is often hindered by potential toxicity peptides, which poses significant barrier clinical application. Traditional experimental methods evaluating peptide time-consuming costly, making process inefficient. Therefore, there an urgent need computational tools specifically designed predict accurately rapidly, facilitating identification safe candidates drug development. Results We provide here novel approach, CAPTP, leverages power convolutional self-attention enhance prediction from amino acid sequences. CAPTP demonstrates outstanding performance, achieving Matthews correlation coefficient approximately 0.82 in both cross-validation settings on independent test datasets. This performance surpasses that existing state-of-the-art predictors. Importantly, maintains its robustness generalizability even when dealing with data imbalances. Further analysis reveals certain sequential patterns, particularly head central regions crucial determining toxicity. insight can significantly inform guide design safer drugs. Availability implementation The source code freely available at https://github.com/jiaoshihu/CAPTP.

Язык: Английский

---

ACVPred: Enhanced prediction of anti-coronavirus peptides by transfer learning combined with data augmentation

Future Generation Computer Systems, Год журнала: 2024, Номер 160, С. 305 - 315

Опубликована: Июнь 7, 2024

Язык: Английский

---

Prediction of cancer drug combinations based on multidrug learning and cancer expression information injection

Future Generation Computer Systems, Год журнала: 2024, Номер 160, С. 798 - 807

Опубликована: Июнь 25, 2024

Язык: Английский

---

SpaCcLink: exploring downstream signaling regulations with graph attention network for systematic inference of spatial cell–cell communication

BMC Biology, Год журнала: 2025, Номер 23(1)

Опубликована: Фев. 12, 2025

Cellular communication is vital for the proper functioning of multicellular organisms. A comprehensive analysis cellular demands consideration not only binding between ligands and receptors but also a series downstream signal transduction reactions within cells. Thanks to advancements in spatial transcriptomics technology, we are now able better decipher process microenvironment. Nevertheless, majority existing cell–cell algorithms fail take into account signals In this study, put forward SpaCcLink, method that takes influence individual cells systematically investigates patterns as well networks. Analyses conducted on real datasets derived from humans mice have demonstrated SpaCcLink can help identifying more relevant receptors, thereby enabling us decode genes signaling pathways influenced by communication. Comparisons with other methods suggest identify closely associated biological processes discover reliable ligand-receptor relationships. By means profound all-encompassing comprehension mechanisms underlying be achieved, which turn promotes deepens our understanding intricate complexity

Язык: Английский

---

Moss-m7G: A Motif-Based Interpretable Deep Learning Method for RNA N7-Methlguanosine Site Prediction

Journal of Chemical Information and Modeling, Год журнала: 2024, Номер 64(15), С. 6230 - 6240

Опубликована: Июль 16, 2024

N-7methylguanosine (m7G) modification plays a crucial role in various biological processes and is closely associated with the development progression of many cancers. Accurate identification m7G sites essential for understanding their regulatory mechanisms advancing cancer therapy. Previous studies often suffered from insufficient research data, underutilization motif information, lack interpretability. In this work, we designed novel motif-based interpretable method site prediction, called Moss-m7G. This approach enables analysis RNA sequences motif-centric perspective. Our proposed word-detection module motif-embedding within Moss-m7G extract information sequences, transforming raw base-level into motif-level generating embeddings these sequences. Compared base contain richer contextual which further analyzed integrated through Transformer model. We constructed comprehensive data set to implement training testing process address insufficiency noted prior research. experimental results affirm effectiveness superiority predicting sites. Moreover, introduction enhances interpretability model, providing insights predictive mechanisms.

Язык: Английский

---

MFD–GDrug: multimodal feature fusion-based deep learning for GPCR–drug interaction prediction

Methods, Год журнала: 2024, Номер 223, С. 75 - 82

Опубликована: Янв. 28, 2024

Язык: Английский

---

TPpred-SC: multi-functional therapeutic peptide prediction based on multi-label supervised contrastive learning

Science China Information Sciences, Год журнала: 2024, Номер 67(11)

Опубликована: Окт. 22, 2024

Язык: Английский

---

Excavation of gene markers associated with pancreatic ductal adenocarcinoma based on interrelationships of gene expression

IET Systems Biology, Год журнала: 2024, Номер unknown

Опубликована: Март 26, 2024

Pancreatic ductal adenocarcinoma (PDAC) accounts for 95% of all pancreatic cancer cases, posing grave challenges to its diagnosis and treatment. Timely is pivotal improving patient survival, necessitating the discovery precise biomarkers. An innovative approach was introduced identify gene markers precision PDAC detection. The core idea our method discover pairs that display consistent opposite relative expression differential co-expression patterns between normal samples. Reversal pair analysis partial correlation were performed determine reversal (RDC) pairs. Using incremental feature selection, authors refined selected set constructed a machine-learning model recognition. As result, identified 10 RDC And could achieve remarkable accuracy 96.1% during cross-validation, surpassing expression-based models. experiment on independent validation data confirmed model's performance. Enrichment revealed involvement these genes in essential biological processes shed light their potential roles pathogenesis. Overall, findings highlight as effective diagnostic early detection, bringing hope prognosis survival.

Язык: Английский

---

CFCN: An HLA-peptide Prediction Model based on Taylor Extension Theory and Multi-view Learning

Current Bioinformatics, Год журнала: 2024, Номер 19(10), С. 977 - 990

Опубликована: Март 29, 2024

Background: With the increasing development of biotechnology, many cancer solutions have been proposed nowadays. In recent years, Neo-peptides-based methods made significant contributions, with an essential prerequisite bindings between peptides and HLA molecules. However, binding is hard to predict, accuracy expected improve further. Methods: Therefore, we propose Crossed Feature Correction Network (CFCN) deep learning method, which can automatically extract adaptively learn discriminative features in HLA-peptide binding, order make more accurate predictions on tasks. fancy structure encoding feature extracting process for peptides, as well fusion fine-grained coarse-grained level, it shows advantages given Results: The experiment illustrates that CFCN achieves better performances overall, compared other models aspects. Conclusion: addition, also consider use multi-view process, find out further relations among features. Eventually, encapsulate our model a useful tool research

Язык: Английский

---

Advancing cancer driver gene detection via Schur complement graph augmentation and independent subspace feature extraction

Computers in Biology and Medicine, Год журнала: 2024, Номер 174, С. 108484 - 108484

Опубликована: Апрель 16, 2024

Язык: Английский

---