NIFPTMLaprendizaje automático por teoría de perturbaciones con fusión de información de redes biomoleculares en química médica, cromosómica, y nanoinformática

Abstract

Complex network theory allows the study of biomolecular systems. Since graphs can represent networks, in a protein network, as an example, the nodes are the amino acids and the axes are the sequences and/or spatial interactions/proximities between the amino acids. Numerical parameters/indexes extracted from these Networks Invariants (NI) are used to quantify the structure of these systems. These parameters can be correlated with biological properties using Machine Learning (ML) techniques, allowing predictive models to be found. Additionally, it is necessary to use Information Fusion (IF) techniques from various sources to obtain an enriched dataset. Perturbation Theory (PT) operators process the information by quantifying the perturbations/deviations in the structural variables with respect to expected values for different subsets of categorical variables. It is proposed to use the NIFPTML strategy combining the above-mentioned phases (NI+IF+PT+PT+ML) in an innovative way necessary to study problems involving one or more of these systems at a time. NIFPTML is applied to several complex problems with different systems (drugs, proteins, genes, chromosomes, nanoparticles). GOIN (Gen Orientation Inversion Network) complex networks and their numerical parameters are defined for the first time. This allowed to exemplify the use of NIFPTML in problems involving chromosomes, making a direct incursion into the application of ML in the new area known as Chromosomics.