SIP

SIP predicts proteins that interact with S-Adenosyl methionine (SAM) and localizes SAM interaction sites using primary protein sequences.

Key Features:

• Machine learning models: Uses classifiers trained on whole protein sequence features to predict SAM-interacting proteins.
• Interaction site prediction: Predicts interaction sites using overlapping sequence windows combined with multiple instance learning (MIL) to accommodate imprecisely annotated SAM-binding residues.
• Simulation and evaluation: Employs simulation-based evaluation studies to assess predictive performance for SAM interaction and site prediction.
• Amino acid sub-sequence analysis: Applies data mining to identify characteristic amino acid sub-sequences and their relative positions associated with SAM interaction sites.

Scientific Applications:

• Protein functional annotation: Identification of SAM-interacting proteins and binding sites to support annotation of protein function.
• Drug discovery and therapeutic research: Prioritization of SAM-interacting proteins and sites for mechanistic studies and therapeutic targeting in contexts where SAM is implicated, including cancer, Alzheimer's disease, epilepsy, and other neurological disorders.
• Methylation-related studies: Mapping SAM-binding sites to inform investigations of SAM-dependent methylation processes and enzyme interactions.

Methodology:

Inputs are primary protein sequences; machine learning classifiers trained on whole-sequence features predict SAM-interacting proteins; overlapping sequence windows together with multiple instance learning (MIL) are used for interaction-site prediction; simulation studies evaluate performance; data mining identifies characteristic amino acid sub-sequences and their relative positions.