im6A-TS-CNN

im6A-TS-CNN predicts N6-methyladenosine (m6A) sites across multiple tissues in Homo sapiens, Mus musculus, and Rattus norvegicus using convolutional neural networks applied to one-hot encoded nucleotide sequences to enable tissue-specific m6A mapping.

Key Features:

• Convolutional neural network (CNN): Employs CNNs to learn sequence patterns predictive of m6A sites.
• One-hot encoding: Represents nucleotide sequences using one-hot encoding for model input.
• Tissue coverage: Models developed for brain, liver, kidney, heart, and testis tissues.
• Cross-species support: Applicable to Homo sapiens, Mus musculus, and Rattus norvegicus.
• Spatial specificity: Captures spatial specificity of m6A modifications across different tissues.
• Performance evaluation: Assessed using 5-fold cross-validation and independent dataset analyses with superior results relative to existing methods.

Scientific Applications:

• m6A site identification: Accurate detection of N6-methyladenosine (m6A) sites in RNA sequences.
• Tissue-specific epitranscriptomics: Comparative analysis of m6A modification patterns across brain, liver, kidney, heart, and testis.
• Functional and disease studies: Supporting elucidation of m6A functional roles and implications in disease contexts.
• Cross-species investigations: Enabling comparative epitranscriptomic studies among human, mouse, and rat.

Methodology:

Samples were encoded with one-hot encoding and analyzed using convolutional neural networks (CNNs), with performance evaluated by 5-fold cross-validation and independent dataset analyses.