{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"omics_type":["Unknown"],"volume":["5(1)"],"submitter":["Tahir M"],"pubmed_abstract":["<h4>Motivation</h4>N6-methyladenosine ( m6A ) is the most abundant internal modification in eukaryotic mRNA and plays essential roles in post-transcriptional gene regulation. While several deep learning approaches have been proposed to predict m6A sites, most suffer from limited chromosome-level generalizability due to evaluation on randomly split datasets.<h4>Results</h4>In this study, we propose two novel hybrid deep learning models-Hybrid Model and Hybrid Deep Model-that integrate local sequence features (<i>k</i>-mers) and contextual embeddings via convolutional neural networks to improve predictive performance and generalization. We evaluate these models using both a Random-Split strategy and a more biologically realistic Leave-One-Chromosome-Out setting to ensure robustness across ge"],"journal":["Bioinformatics advances"],"pagination":["vbaf170"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC12288952"],"repository":["biostudies-literature"],"pubmed_title":["Hybrid representation learning for human m&lt;sup&gt;6&lt;/sup&gt;A modifications with chromosome-level generalizability."],"pmcid":["PMC12288952"],"pubmed_authors":["Liu Q","Tahir M","Ramanna S"],"additional_accession":[]},"is_claimable":false,"name":"Hybrid representation learning for human m&lt;sup&gt;6&lt;/sup&gt;A modifications with chromosome-level generalizability.","description":"<h4>Motivation</h4>N6-methyladenosine ( m6A ) is the most abundant internal modification in eukaryotic mRNA and plays essential roles in post-transcriptional gene regulation. While several deep learning approaches have been proposed to predict m6A sites, most suffer from limited chromosome-level generalizability due to evaluation on randomly split datasets.<h4>Results</h4>In this study, we propose two novel hybrid deep learning models-Hybrid Model and Hybrid Deep Model-that integrate local sequence features (<i>k</i>-mers) and contextual embeddings via convolutional neural networks to improve predictive performance and generalization. We evaluate these models using both a Random-Split strategy and a more biologically realistic Leave-One-Chromosome-Out setting to ensure robustness across ge","dates":{"release":"2025-01-01T00:00:00Z","publication":"2025","modification":"2026-03-17T15:35:37.102Z","creation":"2025-08-17T03:06:09.621Z"},"accession":"S-EPMC12288952","cross_references":{"pubmed":["40708868"],"doi":["10.1093/bioadv/vbaf170"]}}