CN103714220A - Method for predicting elimination speed of persistent organic pollutants on coastal zones - Google Patents

Abstract

The invention relates to a method for predicting the elimination speed of persistent organic pollutants on coastal zones. The method comprises the following steps that a QSAR model is established through a partial least square method and the elimination speed of the persistent organic pollutants is obtained according to the QSAR model. The model of the method is explicit in application domain and the imitative effect, the robustness and the predictive ability are good. On the basis that a compound molecular structure is obtained, the elimination speed of the POPs in a mussel body can be rapidly and efficiently predicted only through descriptors of calculation representational structure characteristics with the application of the established QSAR model. The method for predicting the elimination speed of the persistent organic pollutants on the coastal zones is low in cost, convenient to use, rapid in use, capable of saving a large amount of labor cost and time which are needed for a test and providing important data supports for ecological risk assessment and management of the persistent organic pollutants and important in significance.

Description

Prediction coastal zone persistence organic pollutant is eliminated the method for speed

Technical field

The present invention relates to a kind ofly by setting up quantitative structure activity relationship model (QSAR) prediction persistence organic pollutant, in mussel (Elliptio complanata) body, eliminate the method for speed, belong to D-M (Determiner-Measure) construction and the activity relationship technical field of ecological risk assessment.

Background technology

Palycyclic aromatic (PAHs), PBDE (PBDEs) and polychlorinated biphenyl (PCBs) are typical persistence organic pollutant (POPs), have persistence, biological accumulation and three causes effect.Mussel is as the bivalve marine animal that dwells of the universal littoral end of a class, by extensively for the biological monitoring of marine pollution.Mussel has possessed 4 pacing itemss as indicator organism: geographic distribution is wide, be coastal and the sociales of mouth of sea area, biological physical efficiency pollution of circulation thing, can be to multiple environmental pollution deposits yields response.In addition, more thorough to the fundamental biological knowledge the characteristic study of mussel, thereby the indicator organism using mussel as marine pollution monitoring is generally accepted.Mussel is widely distributed, can biological concentration POPs, in the mussel body in Dalian Bay, Jiaozhou Bay, Ya Erhu and area, Pearl River Delta, found PAHs, PBDEs and PCBs at present.Eliminate speed (k _d) be an important kinetic parameter, can be used for estimating pollutant by the uptake rate of mussel and pollutant the biological concentration factor in mussel body.

Measuring is to obtain at present compound k _dthe main path of value.But because existing chemicals quantity is over 140,000 kinds, the chemicals management new legislation < < chemicals registration that starts to implement in full in June, 2007 according to European Union, assessment, authorize and statute of limitation > > (Registration, Evaluation, Authorization and Restriction of Chemicals, be called for short REACH rules) " estimation, the basic testing cost of each chemical substance approximately needs 8.5 ten thousand Euros (containing the survey fee of long-term environmental effect, not using), each novel substance complete detection expense approximately needs 570,000 Euros, this means if every kind of chemicals is all carried out to measuring, will bring huge financial expense.Carry out comprehensive experiment test, do not meet " the 3R principle " of the zoopery ethics in chemicals risk management, that is: the model of the number of minimizing (Reduce) experimental animal, the zooperal method of improvement (Refine), alternative (Replace) animal used as test yet.At present, " 3R principle " adopted by the research institution of European and American countries gradually, is able to clear and definite in some legislations in the Bing U.S., Britain, France and even whole Europe and government regulation.Therefore, if all obtain all relevant informations by experimental test, not only have the problem of testing expense costliness and animal ethics aspect, and the standard specimen of the chemicals having is difficult to obtain, lag behind in time, be difficult to meet the demand data of organic chemicals ecological risk assessment and supervision.Therefore, by a kind of biological model of eliminating speed of organic contaminant that can rapidly and efficiently obtain of quantitative structure-activity relationship (QSAR) method development, there is important application value.

QSAR refers to and is associated with the molecular structure of organic pollutants and the Quantitative Prediction Model of its physicochemical property, environmental behaviour and toxicology parameter (being referred to as activity), can reflect and disclose the molecular structure of organic contaminant and the inner link between its environmental behaviour and toxicological effect, than experimental test sooner, more cheap, the method is applied at aspects such as the environmental photochemistry behavior of organic contaminant, photic toxicity parameter, partition parameter and ecotoxicological effects.Therefore, the biology of compound is eliminated the foundation of speed QSAR model, with respect to the experimental technique based on traditional, determines k _dnumerical value, be more conducive to quickness and high efficiency evaluating chemical product bioaccumulation ability and carry out ecological risk assessment.

Oragnization for Economic Co-operation and Development (OECD), around the safety issue of chemicals, has carried out the applied research of QSAR technology.2004 is the application of standard QSAR technology, and OECD has proposed QSAR model construction and application guide rule.This guide rule requires the QSAR setting up should meet following standard: 1. for clearly defined environmental index; 2. there is clear and clear and definite mathematical algorithm; 3. defined the application domain of model; 4. model has suitable degree of fitting, stability and predictive ability; 5. preferably can carry out mechanism explain.2007, OECD issued about confirming and verify the policy paper of QSAR model.

At present, existing many researchers apply QSAR method and have set up the biological speed k that eliminates of organic contaminant _dforecast model.If document " QSAR & Combinatorial Science.2011,28:537 – 541 " is based on quantum chemical descriptor and offset minimum binary (PLS) method, set up PAHs k in mussel (Elliptio complanata) body _dqSAR model.Model has good robustness, but is only applicable to single structure type compound (polycyclic aromatic hydrocarbon compounds), and model application domain is little.Document " Archives of Environment Contamination and Toxicology.2004,47:74 – 83 " is based on logK _oWvalue has built the k of PCBs _dvalue model, but model is too simple, does not consider the mechanism of action and performance evaluation.Document " Chemosphere.2007,69:362 – 370 " has been set up 8 PBDEs and 5 PCBs k in mussel body _dvalue and K _oWthe QSAR model of value, but the number of training set compound very little, also the mechanism explain to model not.Document " Science in China, Series B Chemistry.2009,52:1281 – 1286 " has been set up k based on PLS and theoretical molecule descriptor _dqSAR model, but the application domain of model is not characterized.Therefore be necessary to set up one and contain multiple types compound, and model structure is simple, prediction rule is transparent, the QSAR model of easy to understand and practical application, according to OECD guide rule, model is carried out to application domain sign and mechanism explain simultaneously.

Summary of the invention

For above-mentioned technical deficiency, the object of the invention is to develop and a kind ofly succinctly, fast, efficiently predict that persistence organic pollutant eliminates the method for speed in mussel body.

The technical solution adopted for the present invention to solve the technical problems is: prediction coastal zone persistence organic pollutant is eliminated the method for speed, comprises the following steps: the elimination speed that obtains organic contaminant by the QSAR model that adopts offset minimum binary method to build.

Described QSAR model is

${\log k}_d=1.34\&times;{10}^{-1}-1.07\&times;{10}^3\mathrm{\&omega;}-4.86\&times;{10}^{-3}\mathrm{polar}-7.65\&times;{\stackrel{\&OverBar;}{V_s}}^{-}-4.31\&times;{10}^{-2}\mathrm{\&tau;}$

Wherein, k _dfor the elimination speed of organic contaminant, ω represents molecule electrophilicity index, and polar represents mean molecule polarizability;

the mean value that represents negative electrostatic potential on molecular surface; τ represents the balance parameters of molecular surface electrostatic potential.

Described molecule electrophilicity index obtains by following formula

$\mathrm{\&omega;}=\frac{{\mathrm{\&mu;}}^2}{2\mathrm{\&eta;}},$ Chemical potential $\mathrm{\&mu;}=\frac{E_{\mathrm{LUMO}}+E_{\mathrm{HOMO}}}{2},$ Chemical hardness $\mathrm{\&eta;}=\frac{E_{\mathrm{LUMO}}-E_{\mathrm{HOMO}}}{2},$ E _hOMOit is the highest occupied molecular orbital energy of organic contaminant; E _lUMOit is the minimum track energy that do not occupy of organic contaminant.

On described molecular surface, the mean value of negative electrostatic potential obtains by following formula

${\stackrel{\&OverBar;}{V_S}}^{-}=\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{V_S}^{-}\left(r_j\right)$

Wherein, β is counting of the negative electrostatic potential calculating of molecular surface, V _s ^-(r _j) be the negative electrostatic potential of molecular surface of j point.

The balance parameters of described molecular surface electrostatic potential obtains by following formula

$\mathrm{\&tau;}=\frac{{\mathrm{\&sigma;}}_{+}^2{\mathrm{\&sigma;}}_{-}^2}{{\left({\mathrm{\&sigma;}}_{\mathrm{tot}}^2\right)}^2},$ ${\mathrm{\&sigma;}}_{\mathrm{tot}}^2={\mathrm{\&sigma;}}_{+}^2+{\mathrm{\&sigma;}}_{-}^2=\frac{1}{\mathrm{\&alpha;}}\underset{i=1}{\overset{\mathrm{\&alpha;}}{\mathrm{\&Sigma;}}}{[V^{+}\left(r_i\right)-{\stackrel{\&OverBar;}{V_s}}^{+}]}^2+\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{[V^{-}\left(r_j\right)-{\stackrel{\&OverBar;}{V_s}}^{-}]}^2$

Wherein,

with

be respectively the variance that molecule Bader face electrostatic potential distributes for the region electrostatic potential on the occasion of with negative value; α and β are respectively counting that the positive electrostatic potential of molecular surface and negative electrostatic potential calculate; V ⁺(r _i) and V ^-(r _j) represent respectively electrostatic potential positive on molecular surface and negative electrostatic potential; with

be respectively the mean value of the positive electrostatic potential of molecular surface and negative electrostatic potential.

Described organic contaminant comprises palycyclic aromatic, PBDE, polychlorinated biphenyl.

The present invention has following beneficial effect and advantage:

1. the present invention adopts partial least squares regression algorithm, chooses descriptor built forecast model based on Analysis on Mechanism, and the QSAR model of foundation has the good goodness of fit, robustness and predictive ability.Model simple, transparency is strong, is convenient to understand and practical application.

2. QSAR model of the present invention, clear and definite application domain scope, model is contained the organic compound of various structures, can provide basic data for risk assessment and the management work of persistence organic pollutant.

3. adopt the inventive method can predict the elimination speed of dissimilar persistence organic pollutant in mussel body, the method is with low cost, easy and quick, can save in a large number the required manpower of experiment test, expense and time.

4. foundation and checking that the biology the present invention relates to is eliminated rate prediction method strictly develop and use guide rule according to the QSAR model of OECD regulation, therefore use the Prediction of biodegradability result of patent of the present invention, can and manage for organic chemicals ecological risk assessment important Data support is provided, there is important theory and realistic meaning.

Accompanying drawing explanation

Fig. 1 is method flow diagram of the present invention;

Fig. 2 is logk _dmeasured value and the fitted figure of predicted value;

Fig. 3 is model application domain phenogram.

Embodiment

Below in conjunction with embodiment, the present invention is described in further detail.

The method can directly be predicted its elimination speed according to molecular structure of compounds, and then the environmental persistence of target compound is predicted and evaluated, for Risk Assessment of Chemicals and management provide necessary basic data.As shown in Figure 1, comprise the steps:

(1) data collection and division: collect 28 palycyclic aromatics (PAHs), 8 PBDEs (PBDEs) and 28 polychlorinated biphenyl (PCBs) of measuring in document for predicting the elimination speed k in mussel (Elliptio complanata) body _dvalue, is divided into training set and checking collection according to 4:1 at random by above-mentioned data;

(2) QSAR model construction: according to Analysis on action mechanism, calculate 13 kinds of quantum chemical descriptors, adopt offset minimum binary method to build logk _dwith the quantitative structure activity relationship QSAR model of training set compound molecule descriptor, expression formula is:

${\log k}_d=1.34\&times;{10}^{-1}-1.07\&times;{10}^3\mathrm{\&omega;}-4.86\&times;{10}^{-3}\mathrm{polar}-7.65\&times;{\stackrel{\&OverBar;}{V_s}}^{-}-4.31\&times;{10}^{-2}\mathrm{\&tau;}$

Wherein ω represents molecule electrophilicity index; Polar represents mean molecule polarizability;

the mean value that represents negative electrostatic potential on molecular surface; τ represents the balance parameters of molecular surface electrostatic potential.

(3) checking of QSAR model and application domain characterize: square Q of outside prediction related coefficient for the result of QSAR model ² _eXTrMSE represents with root-mean-square error, and the compound application domain of QSAR model adopts the method for Williams figure to characterize.

The present invention has collected 28 palycyclic aromatics (PAHs), 8 PBDEs (PBDEs) and 28 polychlorinated biphenyl (PCBs) the elimination speed kd value in mussel (Elliptio complanata) body altogether.

The technical solution used in the present invention comprises the steps:

(1) for guaranteeing the data accuracy for modeling, the measured value of experiment that document is collected is assessed and is analyzed.Choose the experimental data that adopts identical experiment method as far as possible.Compound comprises palycyclic aromatic, PBDE and polychlorinated biphenyl.Ratio random division training set and checking collection with 4:1.

(2) mechanism anatomy and Molecular structure descriptor chooses.By inference, the elimination speed of compound in mussel body may be distributed and interact relevant with its balance.This process comprises: distribution capability, dipole-dipole interaction, hydrogen bond and the electrostatic interaction of compound in biologic facies.

Based on this, choose 13 theoretical quantum chemical descriptors for QSAR modeling: electron density volume (V), mean molecule polarizability (polar) are for characterizing hole interaction; Molecule highest occupied molecular orbital can (E _hOMO), the minimum track that do not occupy of molecule can (E _lUMO), the most positive net charge (qH of hydrogen atom in molecule ⁺), on electronegativity index (ω), molecular surface the most just and the most negative electrostatic potential (V _{s, max}, V _{s, min}), the mean value of positive electrostatic potential and negative electrostatic potential on molecular surface the balance parameters (τ) of the dispersion degree of molecular surface electrostatic potential (П), electrostatic potential is selected is used for characterizing hydrogen bond and electrostatic interaction.

(3) calculating of molecule descriptor.V is defined as the electron density volume in the space that electron density is 0.001 electronics/cube bohr.Quantum chemical descriptor all adopts Gaussian09 program software to calculate.Detailed process is: first in Chemoffice, generate initial molecular structure, and utilize PM3 method wherein to carry out initial optimization, the graphic file obtaining is changed into Gaussian input file.Then use the B3LYP method in the density functional theory (DFT) in Gaussian09 routine package, compound molecule is carried out to structure optimization in 6-31G (d, p) base group level, obtain its stable molecular structure.Then the geometric configuration of having optimized is carried out to frequency analysis, to guarantee that system without void frequently.In the calculating of above-mentioned geometry optimization and DFT, all use self-consistent field (SCRF) and integral equation form polarization continuum Model (IEFPCM) to consider solvent (water) effect.In SCRF model, solute molecule is positioned in vacuum cavity, and solvent is regarded as continuous non-structure but have the medium of certain DIELECTRIC CONSTANT ε (ε=78.4 of water).ω is calculated by formula below:

$\mathrm{\&omega;}=\frac{{\mathrm{\&mu;}}^2}{2\mathrm{\&eta;}}---\left(1\right)$

$\mathrm{\&mu;}=\frac{E_{\mathrm{LUMO}}+E_{\mathrm{HOMO}}}{2}---\left(2\right)$

$\mathrm{\&eta;}=\frac{E_{\mathrm{LUMO}}-E_{\mathrm{HOMO}}}{2}---\left(3\right)$

Here, μ is chemical potential; η is Chemical hardness; E _hOMOthe highest occupied molecular orbital energy of compound, E _lUMOit is the minimum track energy that do not occupy of compound.

In the structure that the calculating of molecular surface electrostatic potential has been optimized at Gaussian09, carry out.Electrostatic potential is calculated and is adopted lattice Method, and a cube lattice precision set is Cube=fine.For each molecule, can obtain nearly 100 like this ³the electrostatic potential of individual point.Then, the electrostatic potential of these points is carried out to statistical computation and just obtain desired parameters.Calculated respectively V _{s, max}, V _{s, min},

П and τ.Design parameter is defined as follows:

The mean value of the positive and negative electrostatic potential of molecular surface

${\stackrel{\&OverBar;}{V_S}}^{+}=\frac{1}{\mathrm{\&alpha;}}\underset{i=1}{\overset{\mathrm{\&alpha;}}{\mathrm{\&Sigma;}}}{V_S}^{+}\left(r_i\right)---\left(4\right)$

${\stackrel{\&OverBar;}{V_S}}^{-}=\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{V_S}^{-}\left(r_j\right)---\left(5\right)$

Average mark divergence (П)

$\mathrm{\&Pi;}=\frac{1}{\mathrm{\&alpha;}+\mathrm{\&beta;}}\underset{i=1}{\overset{\mathrm{\&alpha;}+\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}|V\left(r_i\right)-\stackrel{\&OverBar;}{V_s}|---\left(6\right)$

$\stackrel{\&OverBar;}{V_s}=\frac{\mathrm{\&alpha;}{\stackrel{\&OverBar;}{V_s}}^{+}+\mathrm{\&beta;}{\stackrel{\&OverBar;}{V_s}}^{-}}{\mathrm{\&alpha;}+\mathrm{\&beta;}}---\left(7\right)$

The equilibrium constant (τ)

$\mathrm{\&tau;}=\frac{{\mathrm{\&sigma;}}_{+}^2{\mathrm{\&sigma;}}_{-}^2}{{\left({\mathrm{\&sigma;}}_{\mathrm{tot}}^2\right)}^2}---\left(8\right)$

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="tot"\; filenames="HDA0000454030980000012.png"\; state="\{\{state\}\}">tot</figure-callout>}}^2={\mathrm{\&sigma;}}_{+}^2+{\mathrm{\&sigma;}}_{-}^2=\frac{1}{\mathrm{\&alpha;}}\underset{i=1}{\overset{\mathrm{\&alpha;}}{\mathrm{\&Sigma;}}}{[V^{+}\left(r_i\right)-{\stackrel{\&OverBar;}{V_s}}^{+}]}^2+\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{[V^{-}\left(r_j\right)-{\stackrel{\&OverBar;}{V_s}}^{-}]}^2---\left(9\right)$

Wherein, s represents molecular surface; α and β are respectively counting that the positive electrostatic potential of molecular surface and negative electrostatic potential calculate; V ⁺(r _i) and V ^-(r _j) represent respectively electrostatic potential positive on molecular surface and negative electrostatic potential;

with be respectively the mean value of the positive electrostatic potential of molecular surface and negative electrostatic potential; the average electrostatic potential that represents molecular surface;

with

be respectively molecule Bader face electrostatic potential for " on the occasion of " and the variance of the region electrostatic potential distribution of " negative value ".V _s ⁺(r _i), V _s ^-(r _j) be respectively the positive electrostatic potential of molecular surface, a j negative electrostatic potential of molecular surface of putting of i point, V (r _i) be the molecular surface electrostatic potential of i point.

(4) foundation of QSAR model.Use the Molecular structure descriptor calculating in offset minimum binary (PLS) and step (3), obtain as drag:

${\log k}_d=1.34\&times;{10}^{-1}-1.07\&times;{10}^3\mathrm{\&omega;}-4.86\&times;{10}^{-3}\mathrm{polar}-7.65\&times;{\stackrel{\&OverBar;}{V_s}}^{-}-4.31\&times;{10}^{-2}\mathrm{\&tau;}$

Wherein ω represents molecule electrophilicity index; Polar represents mean molecule polarizability;

the mean value that represents negative electrostatic potential on molecular surface; τ represents the balance parameters of molecular surface electrostatic potential.

In model, logk _dbe expressed as the function of 4 descriptor variables, training set data n=51.Related coefficient square (R ²) be 0.952, standard deviation (SE)=0.119, level of significance (p) <0.0001, illustrates that model has the good goodness of fit.Accumulation cross validation coefficient (Q ² _cUM)=0.947, shows that model has good robustness (Fig. 2).

(5) checking and the application domain of the QSAR model of setting up.

The predictive ability of QSAR model need to be checked by external certificate.N=13 data of checking collection.The result of external certificate can be by square (Q of outside prediction related coefficient ² _eXT) and the root-mean-square error RMSE of external certificate result represent.These two parameter-definitions are:

${{\mathrm{<figure-callout\; id="2"\; label="Q\; EXT"\; filenames="HDA0000454030980000012.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{\mathrm{EXT}}}^2=1-\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(y_i-{\hat{y}}_i)}^2}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(y_i-{\stackrel{\&OverBar;}{y}}_{\mathrm{EXT}})}^2}---\left(10\right)$

$\mathrm{RMSE}=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(y_i-{\hat{y}}_i)}^2}{n}}---\left(11\right)$

Wherein, the number of n representation compound, y _iwith

the measured value and the predicted value that represent respectively i compound;

mean value for compound activity measured value;

the mean value that represents external certificate collection predicted value.

The Williams figure of the present invention's employing based on leverage evaluated the application domain of the QSAR model of setting up, as shown in Figure 3.Williams figure is that impact (leverage) value of integrated application Molecular structure descriptor is (with h _irepresent, i represents different compounds) and pass through the descriptor field that standardized cross validation residual error characterizes QSAR model, this method has been used to the evaluation of QSAR model application domain.

The H value of training set compound (cap matrix) can be tried to achieve by formula below:

H=X(X ^TX) ^–1X ^T (12)

In formula, the matrix that X is n * k, n is the number of training set compound, and k is the number of predictive variable (Molecular structure descriptor), and X matrix has formed the descriptor space of training set compound.

Leverage (the h of each compound _i) value (being lever value) is the corresponding cornerwise value of H, can be calculated by formula (13):

h _i=x _i ^T(X ^TX) ^–1x _i (13)

In formula, x _iit is the row vector of i molecular structure of compounds descriptor.

Warning value (h ^*) be defined as:

h ^*=3(k+1)/n (14)

Wherein, the number that k is descriptor.

For QSAR model of the present invention, checking collection compound k in application domain _dthe result of prediction is: n=13, Q ² _eXT=0.892, RMSE=0.160.This QSAR model of surface has good predictive ability (Fig. 2).

Embodiment 1

2,6-dimethylnaphthalene: adopting Williams figure method to calculate its lever value is 0.2349<h ^*(early warning value)=0.294, residual (SE)=-0.0288>-3, illustrates that this compound is in QSAR model application domain.Based on the mechanism of action, adopt Gaussian09 software, according to narration method in invention, calculate four descriptor values in model.

The abatement of pollution speed logk of 2,6-dimethylnaphthalene in mussel body _dmeasured value be :-0.577.As follows based on QSAR model prediction step:

logk _d=0.134–1070×(0.000674)–0.00486×(174.4643)–76.5×(-0.01197)–0.0431×(0.018006)=-0.527。Predicted value extremely conforms to measured value.

Embodiment 2

Dibenzo furans: adopting Williams figure method to calculate its lever value is 0.1335<h ^*(early warning value)=0.294, residual (SE)=0.2374<3, illustrates that this compound is in QSAR model application domain.Based on the mechanism of action, adopt Gaussian09 software, according to narration method in invention, calculate four descriptor values in model.

The abatement of pollution speed logk of Dibenzo furans in mussel body _dmeasured value be :-0.635.As follows based on QSAR model prediction step:

logk _d=0.134–1070×(0.000815)–0.00486×(169.273)–76.5×(-0.012547)–0.0431×(0.208105)=-0.527。Predicted value extremely conforms to measured value.

Embodiment 3

1-methylphenanthrene: adopting Williams figure method to calculate its lever value is 0.0414<h ^*(early warning value)=0.294, residual (SE)=0.3421<3, illustrates that this compound is in QSAR model application domain.Based on the mechanism of action, adopt Gaussian09 software, according to narration method in invention, calculate four descriptor values in model.

The abatement of pollution speed logk of 1-methylphenanthrene in mussel body _dmeasured value be :-0.858.As follows based on QSAR model prediction step:

logk _d=0.134–1070×(0.000699)–0.00486×(224.2747)–76.5×(-0.0117)–0.0431×(0.120829)=-0.860。Predicted value extremely conforms to measured value.

Embodiment 4

BDE-47: adopting Williams figure method to calculate its lever value is 0.0320<h ^*(early warning value)=0.294, residual (SE)=-0.3495>-3, illustrates that this compound is in QSAR model application domain.Based on the mechanism of action, adopt Gaussian09 software, according to narration method in the present invention, calculate four descriptor values in model.

The abatement of pollution speed logk of BDE-47 in mussel body _dmeasured value be :-1.721.As follows based on QSAR model prediction step:

logk _d=0.134–1070×(0.000954)–0.00486×(264.171)–76.5×(-0.00788)–0.0431×(0.187566)=-1.648。Predicted value extremely conforms to measured value.

Embodiment 5

PCB-97: adopting Williams figure method to calculate its lever value is 0.0577<h ^*(early warning value)=0.294, residual (SE)=-0.1619>-3, illustrates that this compound is in QSAR model application domain.Based on the mechanism of action, adopt Gaussian09 software, according to narration method in the present invention, calculate four descriptor values in model.

The abatement of pollution speed logk of PCB-97 in mussel body _dmeasured value be :-1.721.As follows based on QSAR model prediction step:

logk _d=0.134–1070×(0.001234)–0.00486×(216.0167)–76.5×(-0.00904)–0.0431×(0.190156)=-1.626。Predicted value extremely conforms to measured value.

Claims (6)

1. prediction coastal zone persistence organic pollutant is eliminated the method for speed, it is characterized in that comprising the following steps: the elimination speed that obtains organic contaminant by the QSAR model that adopts offset minimum binary method to build.

2. prediction coastal zone persistence organic pollutant according to claim 1 is eliminated the method for speed, it is characterized in that: described QSAR model is

${\log k}_d=1.34\&times;{10}^{-1}-1.07\&times;{10}^3\mathrm{\&omega;}-4.86\&times;{10}^{-3}\mathrm{polar}-7.65\&times;{\stackrel{\&OverBar;}{V_s}}^{-}-4.31\&times;{10}^{-2}\mathrm{\&tau;}$

Wherein, k _dfor the elimination speed of organic contaminant, ω represents molecule electrophilicity index, and polar represents mean molecule polarizability;

the mean value that represents negative electrostatic potential on molecular surface; τ represents the balance parameters of molecular surface electrostatic potential.

3. prediction coastal zone persistence organic pollutant according to claim 1 is eliminated the method for speed, it is characterized in that: described molecule electrophilicity index obtains by following formula

$\mathrm{\&omega;}=\frac{{\mathrm{\&mu;}}^2}{2\mathrm{\&eta;}},$ Chemical potential $\mathrm{\&mu;}=\frac{E_{\mathrm{LUMO}}+E_{\mathrm{HOMO}}}{2},$ Chemical hardness $\mathrm{\&eta;}=\frac{E_{\mathrm{LUMO}}-E_{\mathrm{HOMO}}}{2}$

E _hOMOit is the highest occupied molecular orbital energy of organic contaminant; E _lUMOit is the minimum track energy that do not occupy of organic contaminant.

4. prediction coastal zone persistence organic pollutant according to claim 1 is eliminated the method for speed, it is characterized in that: on described molecular surface, the mean value of negative electrostatic potential obtains by following formula

${\stackrel{\&OverBar;}{V_S}}^{-}=\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{V_S}^{-}\left(r_j\right)$

Wherein, β is counting of the negative electrostatic potential calculating of molecular surface, V _s ^-(r _j) be the negative electrostatic potential of molecular surface of j point.

5. prediction coastal zone persistence organic pollutant according to claim 1 is eliminated the method for speed, it is characterized in that: the balance parameters of described molecular surface electrostatic potential obtains by following formula

$\mathrm{\&tau;}=\frac{{\mathrm{\&sigma;}}_{+}^2{\mathrm{\&sigma;}}_{-}^2}{{\left({\mathrm{\&sigma;}}_{\mathrm{tot}}^2\right)}^2},$

${\mathrm{\&sigma;}}_{\mathrm{tot}}^2={\mathrm{\&sigma;}}_{+}^2+{\mathrm{\&sigma;}}_{-}^2=\frac{1}{\mathrm{\&alpha;}}\underset{i=1}{\overset{\mathrm{\&alpha;}}{\mathrm{\&Sigma;}}}{[V^{+}\left(r_i\right)-{\stackrel{\&OverBar;}{V_s}}^{+}]}^2+\frac{1}{\mathrm{\&beta;}}\underset{j=1}{\overset{\mathrm{\&beta;}}{\mathrm{\&Sigma;}}}{[V^{-}\left(r_j\right)-{\stackrel{\&OverBar;}{V_s}}^{-}]}^2$

Wherein, with

be respectively the variance that molecule Bader face electrostatic potential distributes for the region electrostatic potential on the occasion of with negative value; α and β are respectively counting that the positive electrostatic potential of molecular surface and negative electrostatic potential calculate; V ⁺(r _i) and V ^-(r _j) represent respectively electrostatic potential positive on molecular surface and negative electrostatic potential;

with

be respectively the mean value of the positive electrostatic potential of molecular surface and negative electrostatic potential.

6. prediction coastal zone persistence organic pollutant according to claim 1 is eliminated the method for speed, it is characterized in that: described organic contaminant comprises palycyclic aromatic, PBDE, polychlorinated biphenyl.

Images