TY - JOUR
T1 - MIA-QSAR coupled to principal component analysis-adaptive neuro-fuzzy inference systems (PCA-ANFIS) for the modeling of the anti-HIV reverse transcriptase activities of TIBO derivatives
AU - Goodarzi, Mohammad
AU - Freitas, Matheus P.
N1 - Funding Information:
CNPq is gratefully acknowledged for the fellowship (to M.P.F.), as is FAPEMIG for the financial support of this research.
PY - 2010/4
Y1 - 2010/4
N2 - The activities of a series of HIV reverse transcriptase inhibitor TIBO derivatives were recently modeled by using genetic function approximation (GFA) and artificial neural networks (ANN) on topological, structural, electronic, spatial and physicochemical descriptors. The prediction results were found to be superior to those previously established. In the present work, the multivariate image analysis applied to quantitative structure-activity relationship (MIA-QSAR) method coupled to principal component analysis-adaptive neuro-fuzzy inference systems (PCA-ANFIS), which accounts for non-linearities, was applied on the same set of compounds previously reported. Additionally, partial least squares (PLS) and multilinear partial least squares (N-PLS) regressions were used for comparison with the MIA-QSAR/PCA-ANFIS model. The ANFIS procedure was capable of accurately correlating the inputs (PCA scores) with the bioactivities. The predictive performance of the MIA-QSAR/PCA-ANFIS model was significantly better than the MIA-QSAR/PLS and N-PLS models, as well as than the reported models based on CoMFA, CoMSIA, OCWLGI and classical descriptors, suggesting that the present methodology may be useful to solve other QSAR problems, specially those involving non-linearities.
AB - The activities of a series of HIV reverse transcriptase inhibitor TIBO derivatives were recently modeled by using genetic function approximation (GFA) and artificial neural networks (ANN) on topological, structural, electronic, spatial and physicochemical descriptors. The prediction results were found to be superior to those previously established. In the present work, the multivariate image analysis applied to quantitative structure-activity relationship (MIA-QSAR) method coupled to principal component analysis-adaptive neuro-fuzzy inference systems (PCA-ANFIS), which accounts for non-linearities, was applied on the same set of compounds previously reported. Additionally, partial least squares (PLS) and multilinear partial least squares (N-PLS) regressions were used for comparison with the MIA-QSAR/PCA-ANFIS model. The ANFIS procedure was capable of accurately correlating the inputs (PCA scores) with the bioactivities. The predictive performance of the MIA-QSAR/PCA-ANFIS model was significantly better than the MIA-QSAR/PLS and N-PLS models, as well as than the reported models based on CoMFA, CoMSIA, OCWLGI and classical descriptors, suggesting that the present methodology may be useful to solve other QSAR problems, specially those involving non-linearities.
KW - Anti-HIV reverse transcriptase activities
KW - MIA-QSAR
KW - PCA-ANFIS
KW - TIBO derivatives
UR - http://www.scopus.com/inward/record.url?scp=77649190356&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77649190356&partnerID=8YFLogxK
U2 - 10.1016/j.ejmech.2009.12.028
DO - 10.1016/j.ejmech.2009.12.028
M3 - Article
C2 - 20060625
AN - SCOPUS:77649190356
SN - 0223-5234
VL - 45
SP - 1352
EP - 1358
JO - European Journal of Medicinal Chemistry
JF - European Journal of Medicinal Chemistry
IS - 4
ER -