Diseño de un algoritmo para la detección y validación de patrones de biomarcación en conjuntos de mediciones de espectrometría de masas aplicados al estudio del cáncer

Abstract

This Graduation Project shows the results of the design, implementation and testing of an algorithm for extracting biomarkers zones in datasets of mass spectrometry measurements. The algorithm designed is structured by three main stages, processing of measurements, biomarkers zones selection and validation of results. The major contribution of this work resides in the computational efficiency of the algorithm for its heuristic nature, which in combination with reduced dimensional measurements through a statistical filter formed by the t-Student test, Mann-Whitney and chi2 decrease the number of components on rate of two hundred by one, allowing clearly identify which are the genetic mutations that differentiate cancerous of the healthy biological states in the human body systems. The tests of this algorithm were performed in Matlab. The final results are exported in plain text files, which in turn can be easily read in computing translation platforms of format to their chemical equivalents for their analysis and synthesis of drugs that help in the treatment of cancer.