ѧѧ
ACTASCIENTIAECIRCUMSTANTIAE
19982µ182ڿƼڿ
໯QSARо
»**
(йѧԺӦûѧо130022)

ժҪֱӦûķӽṹʽĽṹо45໯Ժͷӽṹ֮ԣöԪع編ӦѧģͣԤ5໯.ȡĽṹϺõطӳ໯Ľṹԣ編ýڶԪع.
ؼʡ໯ṹ붾ԣָˮԲ».
QUANTITATIVE STRUCTUREACTIVITY RELATIONSHIP FOR
TOXICITY OF PHENOLS USING REGRESSION ANALYSIS AND
NEURAL NETWORK

Guo Ming, Xu Lu
( Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022)

ABSTRACTQuantitative structure-toxicity models were developed that link molecular structures of a set of 50 alkylated and/or halogenated phenols with their polar narcosis toxicity, expressed as the negative logarithm of the IDC50(50% growth inhibitory concentration) value in milimoles per liter. Regression alalysis and fully connected, feed-forward neural networks were used to develop the models. Two neural network training algorithms (back-propagation and a quasi-Newton method) were employed.The results demonstrated that the property of compounds can be described better by selective parameters of this paper. Superior results have been achieved by quasi-Newton neural network to regression analysis.
Keywordsphenols, quantitative structure-toxicity relationship, topological indices, partition coefficient, Van der Waals volumes.


1ǰ
ųǧֻѧʵŷŵ̬ϵͳУѧʵĻΣԽԽҪ.ѧ߹ЩȾлϣȷЩлĶЧӦ.Ȼлʵⶨл󼰿ܴڵһЩǱڵлδʵⶨԿɿع㻷дڵлĶʮֱҪ.
ṹ-/о(QSAR/QSPR)ڻѧоõ˷ǳЧӦã1ݣִرڻѧУ2ݣڻȾﶾԵоҪʵüֵ.
ҪӦʵ½Amָ3ݣˮԲlgPԷ໯ԽطȫȫɻķӽṹʽȽöԪعý.

2
вļɱʵƵĶԪͳƷ(MSAPP)֧.Բļһ򵥽.
2.1ˮԲlgPļ
/ˮеķϵĶlgPΪˮԲQSARойΪӦõһlgPӳ˻֬ԵĴС./ˮϵѱ֤ܽϺõģĤ͸ϸĤת˹lgPΪǹؼӰ֮һӾʵlgPĻ.ĿǰΪֹѾ˶ּlgPֵķ48.нṹƬǽӦñȽϹ㷺һַ .ַıʽΪ
lgP=a+bjfi
(1)
Уa,bΪfiΪеiṹƬϵĹbiΪiṹƬڻгֵĴ.ǵиṹƬ֮໥ãӽһЩ׼ȷȣ
lgP=a+bifi+cjFj
(2)
Уabifiͬʽ(1)FjΪjУӣcjΪFjִ.ṹƬ乱׳ȡͨۺϷ9ȷһЩƬϼ乱׳ȻͨͳƷչã8ݣڴ˻ϣͳƷӽṹУK߽ṹϽΪӵлlgPļ㾫.
lgP=a+bifi+cjFj+K
(3)
ɴƳлķӽṹʽlgPֵĳCACP.ڱоĻΪԻȾ±໯ṹԽΪ򵥣CACP໯lgPֵʱȡK=0ʽ(2)lgPֵ.50໯ķӽṹʽSMILESṹ뷨1011ݻϣиĽڻӵ¼룬MSAPPеCACP㻯ˮԲlgP.CACPвɣṹƬ乱׳ݿ⣬ṹģ飬ƬֽlgPģ飬ģ.
2.2ļ
ڱоУǲõ(V)Ϊӵķ».ڻ͸ϸĤת˹УӵĴСҲҪ.»ͨɻżӺͷ.Уǲף12ݵķ㻯ӵķ».

3
3.1ѡ
ΪʹõѧģͼкõԣСľƫֺнٵı仯𲽻ع鷨ѡAm3VlgPѡıѧ任õVV2lgP,lg2P,Am3ֵ1.໯λĤȺһãlgBRʾlgBRΪIDC5050%Ũ)ĸֵ.
3.2Ԫع
50໯13ݱΪ飺ѵ45Ԥ⼯5(1)ѡ5б׼ѵе45lgBRԷõĶԪعģΪ
lgBR=0.8515-1.7644V+1.0306V2+2.2687?lgP-1.1649?lg2P+0.2069Am3
(4)

R=0.9702,F=125.2423,RMS=0.1590,n=45
ʽУRΪϵFΪF飬RMSΪƫnΪ.
ɷʽ(4)ԤԼ5RMS=0.2068.Ӽоѡõıܷӳ໯lgBRı仯.
3.3編о
ĿǰӦ㷺㷨Ƿ(Backpropagation,BP)㷨Ƿ-ţ㷨14죬.Ҫ-ţ㷨.

1໯ĽṹlgBR
Table 1 The structure parameters and lgBR of phenols

ŻVV2lgPlg2PAm3lgBR(expt)lgBR(reg)lgBR(neu)
153.882903.051.462.131626.7022-0.431-0.384-0.425
226-60.843701.511.903.610046.60280.3960.4020.405
32-57.563313.151.682.822438.46000.2480.0080.246
44-57.163267.271.682.822432.74960.017-0.04450.023
53-57.163267.271.682.822452.37760.4730.21830.460
64-׻65.034228.901.873.496932.9773-0.192-0.113-0.095
73-׻65.034228.901.873.496933.8849-0.062-0.101-0.101
82-ȱ63.364014.492.124.494438.73520.2770.52710.268
92-屽66.484419.592.305.290038.82170.5040.6790.659
104-ȱ63.364014.492.124.494432.97330.5450.4500.552
113-һ75.265664.072.295.244139.12080.2290.2130.208
122-һ75.265664.072.295.244144.48340.1760.2850.128
134-屽66.484419.592.305.290033.04380.6810.6020.606
1423-׻76.185803.392.285.198450.13960.1220.2980.071
1524-׻76.185803.392.285.198444.12360.1280.2180.106
1625-׻76.185803.392.285.198444.11370.0090.2170.106
1734-׻76.185803.392.285.198444.18190.1220.2180.105
1835-׻76.185803.392.285.198440.82440.1130.1730.164
193--4-66.644440.892.335.428943.92520.8420.7920.703
202--5-׻74.515551.742.526.350444.11370.6400.7100.700
214-ⱽ71.005041.002.657.022533.13540.8540.9521.023
223-ⱽ71.005041.002.657.022534.05381.1180.9651.071
232-85.487306.832.646.969651.34110.8030.4910.619
243-85.487306.832.646.969645.78990.6090.4160.498
254-85.487306.832.646.969650.01090.4730.4720.434
2625-ȱ72.845305.662.777.672944.11371.1281.1941.217
2723-ȱ72.845305.662.777.672950.13961.2711.2741.354
284--2-׻74.515551.742.526.350444.12360.7000.7090.701
294--3-׻74.515551.742.526.350444.18190.7950.7110.702
3024-ȱ72.845305.6652.777.672944.12361.0361.1941.216
313-嶡95.709158.493.099.548153.66520.7300.7810.775
324-嶡95.709158.493.099.548159.43360.9130.8580.874
3335-ȱ72.845305.672.777.692740.83441.5621.1501.381
342-97.209447.843.4912.180166.20681.0941.4221.236
3524-屽79.086253.653.139.796744.26781.4031.4211.502

1

ŻVV2lgPlg2PAm3lgBR(expt)lgBR(reg)lgBR(neu)
36236-׻87.337226.532.697.236159.46250.4180.5980.460
37345-׻87.337626.532.697.236184.47030.9300.9340.949
38246-ȱ82.326776.583.4211.696455.14741.6951.8021.724
394--35-׻85.667337.642.938.584955.24161.2030.9821.213
404--26-ȱ85.447299.993.6113.032155.20181.7791.8971.721
41245-ȱ82.326776.583.4211.696484.39972.1002.1942.078
424--6--׻87.117588.153.3611.289655.20811.2771.5141.167
434--26-׻88.787881.893.129.734455.20811.2781.1211.383
44246-屽91.688405.223.9715.760955.36042.0502.0361.894
452-嶡-4-׻106.8511416.923.7413.987662.51511.2971.3331.406
464--2--5-׻106.1111259.333.6813.542465.91491.8621.3341.570
476-嶡-24-׻118.0013924.003.9115.288170.07251.2451.2941.259
4826-140.5219745.864.9924.900196.58712.1132.0852.60
4924--6-122.4014981.764.9124.108179.14832.2072.1472.163
5026-嶡-4-׻148.6722102.765.1226.214486.23941.7881.9231.794

* Լ

3.3.1ĽṹĽṹеĲÿĽ.оУ4㷴ṹУΪ5㣬3.1ѡVV2lgPlg2PAm35нֵ01֮䣬ԣֵӦб׼.βĲΪҪݾȨӦС⣬Ҫʼʱ䣬ǵоУβ㣬Ϊ3.оõֵΪ໯lgBR.ڱУĽṹ5X3X3X1.
3.3.2ʼȨصáͨʼȨأѡųʼȨأ޹ɿѭĿǰõķòͬĳʼȨأʵѡ.
3.3.3Լļغģ͵ѡѵУӦܱ⡰ѵ.νѵָѵƫŵӶͣԼƫʼ.ѵѧģȥϡ۲еĸ.αѵһЩ15ݣоѡһ·òԼѵѵ.
һ˵ѵ̵ĽչԼƫ½ܱ֮ƽʼһԼƫʼѵƫ½Ӧֹͣѵ.ͨ£ԼľƫɸСֵ.оУѡȡСļСֵӦȨΪѧģ.
3.3.4編о3.2ͬ50໯ﱻΪ飺ѵ45Լ5.ѵԣýΪRMS=0.0885ֵ1(lgBR(neu))ѵѧģȥԤԼе5RMS0.1313.
3.4
öԪع編50໯lgBRѡԽо.ȫֱͨķӽṹʽ.ʵ֤Щܷӳ໯Ľṹ.編ýԶڶԪعý.йУۺϲ˹̶ĳһСֵòԼѵѵ̷ֹѵĵ.òԼѵ̷Ĳã.
ο
1».ѧѧ.ѧ磬1995
2.ѧչ.ѧҵ磬1995
3ҦԪ»Ԭ˳.лṹ/оµָӦ.ѧѧ1993511041
4Ghose A K Crippen G M. Atomic Physicochemical Parameters for Three-Dimensional Structure-Directed Quantitative Structure-Activity Relationships I. Partition Coefficiests as a Measure of Hydrophobicity. J Comput Chem, 1986,7:565
5Leo A. Calculating logPoct from Structure. Chem Rev, 199330:1283
6Bodor N Gabanyi Z, Wong C K. A New Method for the Estimation of Partition Coefficient. J Am Chem Soc, 1989,111:3783
7Rekker R F. The Hydrophobic Fragmental Constent. New York: Elsevier,1977
8Klopman G, Li J Y, Wang S et al. Computer Automated logP Calculation Based on an Extended Group Contribution Approach. J Chem Inf Comput Sci, 1994,34:752
9Homsch C, Leo A. Substituent Constants for Correlation Analysis in Chemistry and Biology.New York: Wiley, 1979
10Weininger D. SMILES, A Chemical Language and Information System. 1. Introduction to Methodology and Encoding Rules. J Chem Inf Comput Sci, 1988,28:31
11Weininger D. SMILES. 2. Algorithm for Generation of Unigue SMILES Notation. J Chem Inf Comput Sci,1989,29:97
12Bondi A. Von der Waals Volumes and Radii. J Phys Chem, 1964,68:441
13Baker L L, Wesley S K, Schultz T W. Third International Workshop on Quantitative Structure-Activity Relationships in Environmental Toxicology. Knoxuille, TN, 1988: 165168
14Xu L, Ball J W, Dixon S L et al.Quantitative Structure-Activity Relationships for Toxicity of Phenols Using Regression Analysis and Computational Neural Networks. Environ Toxi Chem, 1994, 13:841
15Weigend A S, Rumelhart D E, Huberman B A. In: Touretzky D S, Elman J L, Sefnowski T J et al(eds.). Back-propagation, weight-elimination, and the time geries prediction. Connectionists Models, Proceedings of the 1990 Summer School. Morgan Kaufman, San Mateo. CA, 1991:105116
