01605nam a2200229 a 450000100080000000500110000800800410001910000190006024501320007926001610021130000190037250000200039152007970041165000190120865000130122765300220124065300200126265300290128265300280131165300150133970000210135410318142020-01-15       2008    bl uuuu     u00u1 u    #d1 aMANCINI, A. L.  aPredicting the position of hydrogen atoms in proteins through the pattern recognition of hydrogen bonds.h[electronic resource]  aIn: INTERNATIONAL CONFERENCE OF BRAZILIAN ASSOCIATION FOR BIOINFORMATICS AND COMPUTATIONAL BIOLOGY, 4., 2008, Salvador. Proceedings... Salvador: AB³Cc2008  aNão paginado.  aX-Meeting 2008.  aThe recognition of patterns of hydrogens bond can generate knowledge and be applied to estimate the position of hydrogen atoms in a protein. Tradicional methods of prediction are based on physicalchemical knowledge. This research uses an innovative approach, based on computational methods. Data on hydrogen bond is obtained from the 3D coordinates of atoms from the resolved protein structure. Information about hydrogen bond patterns, once extracted from PDB database, are grouped by similarity, spatially aligned and used as input for the training of a neural network. The present stage of this reseach uses these intelligent techniques to predict the position of atom of hydrogen contained in groups hydroxyl (OH) belonging to the side chains of aminoacids serine, threonine and tyrosine.  aBioinformatics  aProteins  aBase de dados PDB  aBioinformática  aLigação de hidrogênio  aMétodos computacionais  aProteínas1 aROMERO, R. A. F.