Clasificación vía aprendizaje automático de conformaciones moleculares en estructuras teloméricas

Abstract

A great number of studies have been published on the area of small molecules bound to telomeric sequences that fold on a non-cannonical DNA structure known as G-quadruplex structure. The study of these structures has been driven for its potential as targets for drug development. Furthermore, G-quadruplex structures have been identified in other sequences of importance through all the genome. Within the different studies two main models have been used: the human telomeric sequence and the oxytricha, tetrahymena telomeric sequence. Both cases are used as models for computational studies and in vitro essays. Here we study the trajectory generated from an atomistic molecular dynamic simulation to obtain the amount of flexibility and mobility of these model structures. We used data mining and classification tools (i.e. k nearest neighbor method) to identify automatically subpopulations of structures within the simulation. We found the most populated conformations and we discuss the structural findings.