Aplicación de algoritmos evolutivos a la búsqueda

Abstract

The control in the production of proteins in cells is an important problem in molecular biology, which depend  on  a  number  of  applications  in  the  fields  of  medicine ,agriculture  and  livestock. This production is regulated within the cell by a biological switch based on  binding of  a protein on binding sites within certain genes carrying genetic  information. These  binding sites are known  as the  motif of a protein. There are methods in biology to identify these binding sites, but are  very  expensive  procedures  and  time  consuming.  The methods  based  on evolutionary computation algorithms have proven to be more efficient and effective in identifying the positions of the sequences that every other method developed so far. In this work we implemented two  search methods , one based on genetic algorithms (MBMAG) and the other in the estimation of distributions (MBMEDA), which assess the information content of the individuals in the population to discriminate the best  solutions  in  each  generation.  The  results of  these  methods  in  the  search  for  motifs on  DNA bases  were  evaluated using  metrics to measure the performance of different  search  methods.  These  results  demonstrate  that  evolutionary  methods  are  superior  in precision and recall to other methods in the task of  finding the correct sequences of a motif.