01567nam a2200205 a 450000100080000000500110000800800410001910000160006024500910007626001640016730000140033152008430034565300300118865300230121865300250124165300140126665300240128065300400130470000170134415799852020-01-15       2009    bl uuuu     u00u1 u    #d1 aHIGA, R. H.  aUsing Markov random fields to predict protein interface region.h[electronic resource]  aIn: Encontro dos alunos e docentes do departamento de engenharia de computação e automação industrial, 2., 2009, Campinas. Anais... Campinas: UNICAMPc2009  ap. 22-25.  aProtein interface region prediction is very helpful for molecular biologists since it is able to identify specific amino acids potentially inside the interface region for further experimental analysis. This can save specialists time as well as financial budget. Usually, this problem is modeled as a classification problem with contextual information taken into consideration in a post-processing procedure. We claim that this problem can be modeled as a contextual classification problem and propose to use Markov Random Fields - MRF for that. Obtained results show that the performance for this approach is compatible to other methods and deserves further research. Currently, we are evaluating this approach considering different parameters for the MRF models and establishing a suitable criterion for comparison with previous results.  aContextual classification  aInterface proteica  aMarkov random fields  aModelagem  aPattern recognition  aProtein interface region prediction1 aTOZZI, C. L.