Similarity to molecules in the training set is a good discriminator for prediction accuracy in QSAR

被引:226
作者
Sheridan, RP
Feuston, BP
Maiorov, VN
Kearsley, SK
机构
[1] Merck Res Labs, Mol Syst Dept, Rahway, NJ 07065 USA
[2] Merck Res Labs, Mol Syst Dept, West Point, PA 19486 USA
来源
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES | 2004年 / 44卷 / 06期
关键词
D O I
10.1021/ci049782w
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
How well can a QSAR model predict the activity of a molecule not in the training set used to create the model? A set of retrospective cross-validation experiments using 20 diverse in-house activity sets were done to find a good discriminator of prediction accuracy Lis measured by root-mean-square difference between observed and predicted activity. Among the measures we tested, two seem useful: the similarity of the molecule to be predicted to the nearest molecule in the training set and/or the number of neighbors in the training set, where neighbors are those more similar than a user-chosen cutoff. The molecules with the highest similarity and/or the most neighbors are the best-predicted. This trend holds true for narrow training sets and, to a lesser degree, for many diverse training sets and does not depend on which QSAR method or descriptor is used. One may define the similarity using a different descriptor than that used for the QSAR model. The similarity dependence for diverse training sets is somewhat unexpected. It appears to be greater for those data sets where the association of similar activities vs similar structures (as encoded in the Patterson plot) is stronger. We propose a way to estimate the reliability of the prediction of an arbitrary chemical structure on a given QSAR model, given the training set from which the model was derived.
引用
收藏
页码:1912 / 1928
页数:17
相关论文
共 20 条
[1]   Selected concepts and investigations in compound classification, molecular descriptor analysis, and virtual screening [J].
Bajorath, J .
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES, 2001, 41 (02) :233-245
[2]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[3]   SAMPLE-DISTANCE PARTIAL LEAST-SQUARES - PLS OPTIMIZED FOR MANY VARIABLES, WITH APPLICATION TO COMFA [J].
BUSH, BL ;
NACHBAR, RB .
JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN, 1993, 7 (05) :587-619
[4]   Unsupervised data base clustering based on Daylight's fingerprint and Tanimoto similarity: A fast and automated way to cluster small and large data sets [J].
Butina, D .
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES, 1999, 39 (04) :747-750
[5]   ATOM PAIRS AS MOLECULAR-FEATURES IN STRUCTURE ACTIVITY STUDIES - DEFINITION AND APPLICATIONS [J].
CARHART, RE ;
SMITH, DH ;
VENKATARAGHAVAN, R .
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES, 1985, 25 (02) :64-73
[6]  
Cherkassky V, 1997, IEEE Trans Neural Netw, V8, P1564, DOI 10.1109/TNN.1997.641482
[7]  
Gasteiger J., 1990, TetrahedronComput. Methodol, V3, P537, DOI DOI 10.1016/0898-5529(90)90156-3
[8]   The problem of overfitting [J].
Hawkins, DM .
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES, 2004, 44 (01) :1-12
[9]  
Joachims T., 1999, ADV KERNEL METHODS S
[10]  
Kier L.H., 1986, Molecular Connectivity in Structure-Activity Analysis