SPLIF-Enhanced Attention-Driven 3D CNNs for Precise and Reliable Protein–Ligand Interaction Modeling for METTL3

ACS Omega, Journal Year: 2025, Volume and Issue: 10(16), P. 16748 - 16761

Published: April 16, 2025

Structure-based virtual screening (SBVS) is a cornerstone of modern drug discovery pipelines. However, conventional scoring functions often fail to capture the complexities protein-ligand binding interactions. To address this limitation, we developed DeepMETTL3, novel function that integrates 3D convolutional neural networks (CNNs) with multihead attention mechanisms and high-dimensional Structural Protein-Ligand Interaction Fingerprints (SPLIF). This approach enables model intricate interaction patterns while refining prioritizing features for precise classification active inactive compounds. We validated DeepMETTL3 using METTL3 as therapeutic target, employing scaffold-based data-splitting strategy multiple test sets, including challenging sets minimal chemical similarity training data. Our results demonstrate outperforms traditional functions, achieving superior accuracy, robustness, scalability. Key findings include importance an active-to-decoy ratio (1:50) in set enhanced performance optimal placement mechanism after CNN1 improved generalization. represents significant advancement target-specific machine learning SBVS, offering framework can be adapted other biological targets. work underscores potential deep artificial intelligence-based design, balancing computational efficiency predictive power molecular docking screening. The freely available at https://github.com/juniML/DeepMETTL3.

Language: Английский

Pathological α-synuclein dysregulates epitranscriptomic writer METTL3 to drive neuroinflammation in microglia

Cell Reports, Journal Year: 2025, Volume and Issue: 44(5), P. 115618 - 115618

Published: April 23, 2025

Language: Английский

Novel insights into the N 6-methyladenosine modification on circRNA in cancer

Frontiers in Oncology, Journal Year: 2025, Volume and Issue: 15

Published: May 19, 2025

Circular RNAs (circRNAs) are a class of non-coding (ncRNAs) generated through the reverse splicing mRNA precursors (pre-mRNAs). They possess unique loop structure and exhibit remarkable stability. CircRNAs have emerged as promising biomarkers for cancer, with specific circRNAs playing crucial roles in cancer drug discovery, treatment, resistance mechanisms. N6 methyl adenosine (m6A) represents most prevalent RNA modification eukaryotes. In 2017, researchers identified that m6A modifications also occur circRNAs, displaying characteristics. m6A-modified undergo reversible regulation mediated by enzymes involved pathways. These modified interact m6A-binding proteins, thereby influencing processes such alternative splicing, translation degradation. Some enhance their metabolism or facilitate nuclear export to cytoplasm interacting regulation. The study has gained great attention circRNA research due association various diseases. This review summarizes functional mechanisms regulated implications occurrence therapy, primary focus on genesis, regulatory mechanisms, biology diverse types cancers. Additionally, we explore potential application clinical treatment.

Language: Английский