m6A-TSHub

m6A-TSHub provides integrated databases and predictive models for analyzing tissue-specific N6-methyladenosine (m6A) RNA methylation on mRNA and long non-coding RNA (lncRNA) and for assessing the impact of genetic variants on m6A in human tissues and cancers.

Key Features:

• m6A-TSDB: Contains 184,554 functionally annotated m6A sites from 23 human tissues and an additional 499,369 m6A sites across 25 tumor conditions for tissue- and cancer-specific m6A annotation.
• m6A-TSFinder: Predicts tissue-specific m6A methylation sites from RNA sequences using multi-instance deep neural networks with gated attention mechanisms to improve prediction precision.
• m6A-TSVar: Evaluates the impact of genetic variants on tissue-specific m6A RNA modification to identify mutations that may alter m6A methylation patterns across tissues.
• m6A-CAVar: Catalogs 587,983 cancer-associated mutations from The Cancer Genome Atlas (TCGA) across 27 cancer types that have been predicted to affect m6A modifications in primary tissues.

Scientific Applications:

• Tissue-specific m6A mapping: Enables exploration of the distribution and functional implications of m6A modifications across distinct human tissues and tumor conditions.
• m6A site prediction: Supports identification and characterization of potential m6A sites on mRNA and lncRNA within specific tissue contexts.
• Variant impact analysis: Facilitates assessment of how specific genetic mutations may alter m6A methylation and contribute to epitranscriptomic regulation.
• Cancer epitranscriptomics: Provides a resource to study how TCGA-derived somatic mutations influence m6A marks in oncogenesis across 27 cancer types.

Methodology:

m6A-TSFinder employs multi-instance deep neural networks with gated attention to predict tissue-specific m6A sites from RNA sequences; m6A-TSVar assesses effects of genetic variants on m6A, and m6A-CAVar catalogs TCGA-derived mutations predicted to affect m6A.