CROSSalive

CROSSalive predicts RNA secondary-structure profiles in vivo, identifying double-stranded and single-stranded regions and modeling the impact of N6 adenosine methylation (m6A) to provide biologically relevant structural profiles of RNAs regardless of sequence length.

Key Features:

• In vivo prediction: Predicts RNA secondary-structure profiles that distinguish double-stranded and single-stranded regions in living cells.
• m6A analysis: Models the effects of N6 adenosine methylation (m6A) on RNA folding and stability.
• Training and accuracy: Trained on icSHAPE data in the presence (m6a+) and absence (m6a-) of methylation and reports 0.88 accuracy on test datasets.
• Protein-interaction integration: Leverages predictions of protein–RNA interactions to model how m6A impacts RNA structure.
• Large RNA capability: Operates without sequence length constraints and was applied to murine long non-coding RNA Xist (~17,900 nt), yielding a Pearson’s correlation of 0.45 with SHAPE-map data.

Scientific Applications:

• Functional genomics: Assessing how RNA structural features influence gene expression and regulation.
• Epitranscriptomics: Investigating the structural consequences of chemical modifications such as m6A on RNA function and stability.
• Non-coding RNA research: Analyzing large non-coding RNAs, including murine Xist (~17,900 nt), to study structure–function relationships.

Methodology:

Trained on icSHAPE datasets with and without methylation (m6a+ and m6a-), integrates predictions of protein–RNA interactions to model m6A-driven structural changes, and outputs RNA secondary-structure profiles distinguishing single- and double-stranded regions.