MuLiMs-MCoMPAs

MuLiMs-MCoMPAs computes three-dimensional (3D) protein structural descriptors using tensor algebra to generate two-linear and three-linear descriptors that capture geometrical relationships among amino acids as part of the ToMoCoMD-CAMPS suite.

Key Features:

• Tensor-Based Descriptors: Computes 3D protein structural descriptors using tensor algebra, focusing on interactions at two-linear and three-linear levels.
• Novel Representations and Metrics: Produces novel 3D structural representations and (dis)similarity metrics using multimetrics and weighting schemes based on amino acid properties.
• Matrix Normalization and Transformations: Applies matrix normalization procedures including simple-stochastic and mutual probability transformations.
• Topological and Geometrical Cutoffs: Integrates topological and geometrical cutoffs to calculate amino acid interactions and support group-based molecular dynamics (MD) calculations.
• Aggregation Operators: Uses aggregation operators to merge amino-acid and group MD outputs.
• Parallelization and Batch Computing: Implemented in Java 1.8 with the Chemistry Development Kit (CDK) and Jmol libraries, employing a divide-and-conquer strategy for parallelization and including modules for data preprocessing and batch computing.

Scientific Applications:

• Informational Entropy Analysis: The three-linear descriptor family exhibits higher informational entropy than descriptors from existing tools, indicating richer structural information capture.
• Orthogonal Information Capture: Generated indices provide additional orthogonal information beyond current computational approaches for protein structure analysis.
• QSAR Integration: Integrated into a QSAR-based expert system via ProStAF to predict SCOP protein structural classes and folding rates.

Methodology:

Computes 3D descriptors via tensor algebra (two-linear and three-linear forms), uses multimetrics and weighting schemes based on amino acid properties, applies matrix normalization (simple-stochastic and mutual probability transformations), employs topological and geometrical cutoffs for interaction and group-based MD calculations, uses aggregation operators to merge amino-acid and group MDs, and is implemented in Java 1.8 using the Chemistry Development Kit (CDK) and Jmol libraries with a divide-and-conquer parallelization strategy and modules for data preprocessing and batch computing.