GraphProt

GraphProt models RNA-binding protein (RBP) binding preferences from high-throughput experimental data (e.g., CLIP-seq, RNAcompete) by integrating sequence and structural information to predict RBP-RNA interactions.

Key Features:

• Learning Framework: Employs a learning-based approach to derive binding preferences from empirical datasets.
• Input Data: Uses high-throughput experimental data such as CLIP-seq and RNAcompete to train models.
• Sequence and Structural Integration: Integrates sequence-specific and structural information to model RBP binding preferences.
• Benchmarking and Validation: Has been benchmarked against existing literature with predictions aligning to known biological interactions.
• Biological Relevance: Predicted binding affinities correlate with experimental measurements and successfully identified Ago2 targets showing altered expression upon Ago2 knockdown.

Scientific Applications:

• RBP binding site prediction: Predicts RBP binding sites across transcripts and tissues to map interaction landscapes.
• Regulatory role exploration: Facilitates analysis of RBP regulatory roles in diverse biological contexts.
• Functional genomics support: Identifies candidate targets for experimental validation in functional studies.

Methodology:

GraphProt trains models on high-throughput experimental data (CLIP-seq, RNAcompete) using a learning-based framework that integrates sequence and structural information to capture sequence-specific and structural binding preferences.