A Classification Study of Respiratory Syncytial Virus (RSV) Inhibitors by Variable Selection with Random Forest.  
Jump to Full Text  
MedLine Citation:

PMID: 21541057 Owner: NLM Status: PubMednotMEDLINE 
Abstract/OtherAbstract:

Experimental pEC(50)s for 216 selective respiratory syncytial virus (RSV) inhibitors are used to develop classification models as a potential screening tool for a large library of target compounds. Variable selection algorithm coupled with random forests (VSRF) is used to extract the physicochemical features most relevant to the RSV inhibition. Based on the selected small set of descriptors, four other widely used approaches, i.e., support vector machine (SVM), Gaussian process (GP), linear discriminant analysis (LDA) and k nearest neighbors (kNN) routines are also employed and compared with the VSRF method in terms of several of rigorous evaluation criteria. The obtained results indicate that the VSRF model is a powerful tool for classification of RSV inhibitors, producing the highest overall accuracy of 94.34% for the external prediction set, which significantly outperforms the other four methods with the average accuracy of 80.66%. The proposed model with excellent prediction capacity from internal to external quality should be important for screening and optimization of potential RSV inhibitors prior to chemical synthesis in drug development. 
Authors:

Ming Hao; Yan Li; Yonghua Wang; Shuwei Zhang 
Related Documents
:

16265627  Multicomponent variance estimation for binary traits in familybased studies. 9803017  Genetic "clocks" and "soft" events: a twin model for pubertal development and other rec... 16542227  Casecontrol and caseonly designs with genotype and family history data: estimating re... 7755517  Sexspecific causal factors and effects of common environment for symptoms of anxiety a... 24101977  Understanding the contribution of habitats and regional variation to longterm populati... 20634997  An analysis of quantitative pcr reliability through replicates using the c method. 
Publication Detail:

Type: Journal Article Date: 20110221 
Journal Detail:

Title: International journal of molecular sciences Volume: 12 ISSN: 14220067 ISO Abbreviation: Int J Mol Sci Publication Date: 2011 
Date Detail:

Created Date: 20110504 Completed Date: 20110714 Revised Date: 20130529 
Medline Journal Info:

Nlm Unique ID: 101092791 Medline TA: Int J Mol Sci Country: Switzerland 
Other Details:

Languages: eng Pagination: 125980 Citation Subset:  
Affiliation:

School of Chemical Engineering, Dalian University of Technology, Dalian, Liaoning 116012, China; EMails: dluthm@yeah.net (M.H.); zswei@dlut.edu.cn (S.Z.). 
Export Citation:

APA/MLA Format Download EndNote Download BibTex 
MeSH Terms  
Descriptor/Qualifier:


Comments/Corrections 
Full Text  
Journal Information Journal ID (nlmta): Int J Mol Sci Journal ID (publisherid): ijms ISSN: 14220067 Publisher: Molecular Diversity Preservation International (MDPI) 
Article Information Download PDF © 2011 by the authors; licensee MDPI, Basel, Switzerland. openaccess: Received Day: 13 Month: 12 Year: 2010 Revision Received Day: 10 Month: 2 Year: 2011 Accepted Day: 11 Month: 2 Year: 2011 Electronic publication date: Day: 21 Month: 2 Year: 2011 collection publication date: Year: 2011 Volume: 12 Issue: 2 First Page: 1259 Last Page: 1280 ID: 3083704 PubMed Id: 21541057 DOI: 10.3390/ijms12021259 Publisher Id: ijms1201259 
A Classification Study of Respiratory Syncytial Virus (RSV) Inhibitors by Variable Selection with Random Forest  
Ming Hao1  
Yan Li1*  
Yonghua Wang2  
Shuwei Zhang1  
1 School of Chemical Engineering, Dalian University of Technology, Dalian, Liaoning 116012, China; EMails: dluthm@yeah.net (M.H.); zswei@dlut.edu.cn (S.Z.) 

2 Center of Bioinformatics, Northwest A&F University, Yangling, Shaanxi 712100, China; EMail: yh_wang@nwsuaf.edu.cn 

*Author to whom correspondence should be addressed; EMail: yanli@dlut.edu.cn; Tel.: +8641184986062; Fax: +8641184986063. 
Respiratory syncytial virus (RSV), a singlestranded RNA virus of negative genome polarity, is a member of the Pneumovirus genus of the Paramyxovirus family. RSV was first shown to occur in humans in 1957, after being recovered from two infants hospitalized with severe lower respiratory tract infections [^{1},^{2}]. Today, RSV is recognized as the leading cause of virusinduced lower respiratory tract disease among infants and children [^{3}]. Most children are infected with RSV before two years of age, reinfection is a common occurrence and morbidity due to complications is high among premature infants and those with underlying cardiopulmonary problems [^{4}]. Moreover, RSV infections have been associated with increased prevalence of asthma in later childhood [^{5}]. However, RSV was not recognized as a potentially serious problem in adults until the 1970s, when outbreaks of the virus occurred in longterm care facilities [^{6},^{7}]. Until a safe and effective antiviral can be developed for treatment of RSV infections, prevention of the infection by use of antiRSV antibodies appears to be the most acceptable approach. The main therapeutic agents include ribavirin [^{8}] and RSVIGIV [^{9}]. However, both of them pose some disadvantages. For example, ribavirin is not a specific antiviral agent and is teratogenic, while RSVIGIV is derived from blood, and consequently has the potential to transmit bloodborne pathogens. Thus, a search for more potent and selective inhibitors of RSV is clearly necessary. Recently, Nikitenko and coworkers have discovered a potent and selective inhibitor (RFI641) [^{10}]. Chapman et al. [^{11}] also reported the discovery and initial development of RSV604, a novel benzodiazepine with submicromolar antiRSV activity. In addition, with continuous efforts, Meanwell and colleagues have examined several of benzimidazole derivatives with highly potent RSV inhibition activity [^{12}–^{18}].
Traditionally, the biological activity of a drug candidate is obtained via costly and time consuming experiments. Thus the introduction of in silico methods, including the quantitative structureactivity relationship (QSAR) approaches in particular, has been explored in the drug development process for predicting the biological activity of drug candidates [^{19}–^{23}] prior to synthesis, thus attempting to eliminate undesirable compounds in a fast and costeffective manner. However, to our best knowledge, there is still no report of any computational models to classify RSV inhibition activity. Therefore, it is necessary to develop a predictive model to fill this gap.
Construction of a computational model often requires two conditions. The first factor is molecular descriptors, which are used to extract the structural information that is suitable for model development. The software Mold^{2} [^{24}] enables the rapid calculation of a large and diverse set of descriptors encoding twodimensional chemical structure information. Comparative analysis of Mold^{2} descriptors with those calculated by Cerius^{2}, Dragon or MolconnZ on several data sets has demonstrated that Mold^{2} descriptors can convey a similar amount of information as those widelyused software packages [^{24}]. Although a freely available software, it has been proven that Mold^{2} is suitable not only for QSAR [^{25}], but also for virtual screening large databases in drug development [^{24}].
Secondly, the adoption of appropriate classification approaches to establish models is another central element to obtain accurate prediction. Often used classification methods include the simple but interpretable linear discriminant analysis (LDA) and partial least square (PLS) [^{26}], and nonlinear, relatively difficult to interpret but often highly predictive methods such as artificial neural networks (ANN) [^{27}], support vector machine (SVM), random forest (RF), Gaussian process (GP) and so forth [^{28}–^{31}]. All of these methods have a proven record of many successful applications in computational modeling. However, several of these methods often suffer several limitations. For example, traditional statistical method like LDA can only handle data sets where the number of descriptors (p) is smaller than that of the molecules (n), unless again a preselection of the descriptors is executed (e.g., by using successive projections or genetic algorithms [^{32},^{33}], etc.). Also they are not flexible enough and do not explain nonlinear behavior [^{28}]. SVM, a relatively new nonlinear technique employed in classification problems [^{34},^{35}], is not robust to the presence of a large number of irrelevant descriptors [^{28}]. PLS is a popular computational method that expresses a dependent variable in terms of linear combinations of the independent variables commonly known as principal components. However, PLS may not be suitable for handling multiple mechanisms of action [^{28}], such as the nonlinear biological behaviors. Random forest, a new classification and regression tool, has been reported as combining relatively high prediction accuracy and a collection of desired features that make RF uniquely suited for modeling in cheminformatics [^{28}] including predicting a compound’s quantitative or categorical biological activity based on a quantitative description of the compound’s molecular structure. RF has shown excellent performance even when most predictive variables are noise, and be used when the number of variables is much larger than the number of observations, and returns measures of variable importance.
It is well known that an ideal classification model should have high performance with a lower number of descriptors. Thus, in the present work, to optimize the 2D (twodimensional) molecular descriptor subset, while simultaneously enhancing the statistical performance and efficiency of the model, the variable selection (VS) method by RF combined with backward elimination using outofbag (OOB) error is selected to perform a classification task for the current RSV inhibitors to investigate whether the proposed VSRF method can construct an ideal prediction model (i.e., high performance with less descriptors) for this dataset. This method was proposed originally for gene selection. The authors have proven that the novel approach can return very small sets of genes compared to the other alternative variable selection methods, while retaining predictive performance comparable to that of seven alternative stateofart methods [^{36}]. Although this method has been successfully applied to gene selection and microarray data [^{36}], there is still no record of attempts to develop computational models for small molecular inhibitors. To extend the range of application, we examined the VSRF method to classify the current dataset of RSV inhibitors. In addition, based on the performance evaluation, this method has also been compared with four other popular ones, i.e., SVM, GP, LDA, and kNN (k nearest neighbors) on the basis of the selected descriptors within the same data sets.
As a special kind of neural network that can be used for clustering, visualization, and abstraction tasks, selforganizing map (SOM) is especially suitable for data survey due to its prominent visualization properties. In our previous work, this technology has been successfully applied to dataset split [^{22},^{31}]. SOM creates a set of prototype vectors representing the dataset and carries out a topology preserving projection of the prototypes from the ddimensional input space onto a lowdimensional grid [^{37}], which is a convenient visualization space for showing the cluster structure of the data. In the present work, based on the SOM visualization of the whole data, the construction of the training and test sets was made [^{38}]. A small Kohonen network with 6 × 6 = 36 neurons was employed, producing a map with 36 positions. All the compounds with 272 molecular descriptors were placed onto the 36 positions (neurons) of the Kohonen map. Figure 1 demonstrates the distribution of the molecules, where the number corresponds to the series number of the compounds in Table S1 (Supporting Information). The training set is labeled in black and the prediction set in red. The purpose of performing the SOM simulation on the dataset was to guarantee that the representative points of the training set are distributed evenly within the whole area of the descriptor space occupied by the dataset and the representative points of the training set are close to those of the test set, which ensures the reliability of the simulation results.
A VSRF strategy has been developed successfully, with the final number of descriptors being reduced to six from the original 272 for the further study. Since it is recommended that the number of compounds in the training set should be at least fivetimes larger than that of the selected independent variables [^{39}], the model developed by VSRF obviously maintains the recommended ratio. Table 1 lists the selected descriptors together with their definitions and their values are listed in Table S2 (Supporting Information).
Based on the selected descriptors, five different statistical methods (VSRF, SVM, GP, LDA, kNN) were performed to compare their performance, and the detailed statistics are summarized in Table 2. The results predicted by these methods are presented in Table S3 (Supporting Information).
VSRF: Random forest effectively has only one tuning parameter, m_{try.} In the present work, the m_{try} value was tried from 1 to 6 and the optimal value determined by 10fold crossvalidation accuracy (Q_{cv} = 0.816). Ultimately, optimal RF results are obtained based on the m_{try} = 4 and 500 trees in the forest. The efficiency and robustness of the derived models are further evaluated by using the external prediction set. As shown in Table 2, for the external prediction set, the prediction accuracies of VSRF are 100% for high active RSV inhibitors and 88.46% for low active ones, with a total accuracy (Q) of 94.34%. The values of MCC and F are 0.89 and 0.96, respectively.
SVM: Similar to other multivariate statistical models, the performance of SVM depends on the combination of several parameters including the capacity parameter C, the kernel type K and its corresponding indices. C is a regularization parameter which controls the tradeoff between maximizing the margin and minimizing the training error. In this work, the grid search technology was employed to obtain the optimum parameters (C and sigma) using the R package caret [^{40}] on the basis of 10fold cross validation. Here, the function sigest in the kernlab package [^{41}] was used to provide a good estimate of the sigma parameter, so that only the C parameter was tuned. The final values used in the model are C = 10 and sigma = 0.284 with the highest 10fold crossvalidation accuracy (0.791). Using the determined optimal parameters, the SVM obtains statistical results of 85.19%, 80.77% and 83.02% for the sensitivity, specificity and Q of the test set, respectively. The MCC and F values are 0.66 and 0.84, respectively.
GP: The Gaussian process method, based on clearly defined statistical principles and easily programmed [^{42}], was also adopted to classify the RSVrelated compounds. The optimal inverse kernel width for the Radial Basis kernel function (sigma) was finally fixed to 0.284 based on sigest function including the R package kernlab. Based on the 10crossvalidation, the final Q_{cv} of GP we derived is 0.78. As for the RF model, the GP model also presents 100% sensitivity, however, a low specificity of 76.92% for the test set. In addition, the values of Q, MCC and F are 88.68%, 0.79 and 0.9, respectively.
LDA: a widely used classification technology, LDA, was also performed to classify the current dataset based on the selected six descriptors. As shown in Table 2, no statistically satisfactory LDAbased model could be obtained, with the optimal one only depicting sensitivity of 74.07%, specificity of 80.77%, and overall accuracy of 77.36% for the test set. The value of Q_{cv} was just 0.675.
kNN: After 10fold crossvalidation, an optimal k = 17 was determined on the basis of the highest accuracy (Q_{cv} = 0.729). As seen from Table 2, the sensitivity and specificity for the prediction set are 81.48% and 65.38%, respectively. And the overall prediction accuracy for the test set is 73.58%. The values of MCC and F are 0.48 and 0.76. It is obvious that kNN, of the five statistical methods, is uniformly less able to predict than the others.
From the above discussion, it can be concluded that the developed VSRF model performed comparably with SVM and GP, demonstrated by the Q_{cv}(%) of VSRF, SVM and GP of 81.6%, 79.1% and 78%, respectively, in terms of crossvalidation. These models outperform those of the LDA and kNN, whose Q_{cv}(%) are 67.5% and 72.9%, respectively. High crossvalidation accuracy is necessary, but not sufficient for a model with high predictive ability [^{43}], thus an external validation is a better way to estimate the performance of the models. Therefore, a further investigation of Q(%) in the external prediction set was performed, where the VSRF model increases about 11.32% and 5.66% compared to the SVM and GP models, respectively. It should be noted that although GP shares the same prediction ability for high active compounds, for low active inhibitors the prediction accuracy decreases by 11.54% compared with VSRF. From this point of view, one can consider that the VSRF model is more favorable than others for the RSV inhibitors.
In addition, when comparing the other four models, it is observed that the LDA model is comparable to that of kNN, both of them presenting less overall accurate (Q) (77.36% for LDA and 73.58% for kNN) in the test set than the other models. The reason for LDA’s failure may be due to the existence of some nonlinear relationship between the molecular structures and the corresponding activity. For kNN, a possible reason for the low accuracy is that the method—based on the Euclidean distance—may not be the most effective approach for every problem just like the present one. Furthermore, for SVM and GP, their internal prediction ability is comparable, while the performance of GP is slightly better than SVM in terms of the external prediction. The area under the ROC (receiver operating characteristic) curve (AUC) [^{44},^{45}] is also considered as an important criterion for measuring the performance of the model. An AUC value of 1 indicates a theoretically perfect performance, while a value of 0.5 denotes no prediction ability. Clearly, the closer the AUC value is to 1, the better the model performance is. Figure 2 gives the ROC curves of VSRF, SVM, GP, LDA and kNN for the prediction set. The computed AUC values for the five statistical methods are 0.96, 0.89, 0.94, 0.86 and 0.78, respectively, also proving the good prediction ability and reliability of the VSRF model. Thus, our further analysis is only restricted to the VSRF model for prediction of RSV inhibition.
It should be noted that RF, as a new classification and regression tool, can well solve the small n and large p (n < p, that is the number of samples is smaller than that of descriptors) problems [^{28}] even without variable selection. Keeping this in mind, in order to estimate the effect of VSRF, we have compared both the statistical performance with and without variable selection. As shown in Table 3, for the training set, the statistical performance obtained with or without variable selection makes no difference, while the time cost of RF is approximately 20times more than that of VSRF. It must be pointed out that for the RF model without variable selection, the optimal m_{try} is obtained using grid search technology including the R package caret, and the search length is set to 10. For the test set, one can see that the statistics of VSRF outperform RF. The VSRF presents a sensitivity of 100%, while RF gives that of 92.59%, that is to say there are two high active compounds misclassified to low active ones by RF. According to above analysis, one can conclude that the VSRF model depicts not only high computation efficiency but also enhances prediction ability. Therefore, for the RSV inhibitor classification, the VSRF model gives very high statistical results with total accuracies of 100% and 94.34%, for the training and test set, respectively. In the final VSRF model, three compounds (No. 68, 120 and 124) are misclassified (Tables S1 and S3; Supporting Information). The reason for misclassification of compound 68 is unclear, since by comparison with compound 39, the former introduces a polar substituent CH_{2}COOH instead of Et, however, the activity decreases sharply suggesting the atomic polarizabilities may play a role in the RSV inhibition. Compounds 120 and 124 are misclassified as high active molecules by the VSRF model. By investigation of the correctly classified compounds, i.e., 115, 116, 118, 119, 123 and 125∼132 in Tables S1 and S3 (Supporting Information), it is revealed that all of them possess a linear R1 group at position 5. However, compounds 120 and 124 have a ringbased substituent at the same location, which we suppose may be the reason for the misclassification.
By using feature selection, the most appropriate sets of molecular descriptors for predicting the RSV low and high active inhibitors are extracted from the VSRF models, some of which probably provide new insights into the physicochemical characteristics of RSV inhibition by specific classes of compounds. D299, one of the topological descriptors, is a molecular branching index that is calculated from the algebraic formulas derived by Lovasz and Pelikan for special types of trees such as path or star and for particular eigenvalues [^{46}]. The highest molecular branching corresponds to the most branched graphs. This is in agreement with the previous result that the topology of the side chain is important to modulate physical properties [^{12}]. D347 stands for molecular topological path index of order 07. The path counts are molecular descriptors obtained from an Hdepleted molecular graph and are vertex invariants encoding that molecular environment, defined as the number of path lengths m starting from the ith vertex to any other vertex in the graph. A path (or selfavoiding walk) is a walk without any repeated vertices [^{47}]. The path length is the number of edges associated with the path, and this value is increased with the ring size, ring numbers, and the ramification number [^{48}]. Of the selected six descriptors, D503 and D513 belong to 2D autocorrelation classes, which represent the topological structure of the compounds but are more complex in nature than the classical topological descriptors. Computation of these descriptors involves the summations of different autocorrelation functions corresponding to different structural lags and leads to different autocorrelation vectors corresponding to the lengths of substructural fragments. Hence, it can distinguish the details of important substructural differences. In the previous work, the 2D autocorrelation descriptors have been proven advantageous for establishing a QSAR model [^{49}–^{53}]. For the present work, the Moran’s index I [^{53},^{54}] is employed for the classification of RSV inhibitors:
(1)
I=n2L∑ijδij(pki−p¯k)(pkj−p¯k)∑i(pki−p¯k) 
(2)
δij={1if dij=10if dij≠1 
The 2D autocorrelation descriptors can be obtained by summing up the products of certain properties of the two atoms located at a given topological distance or spatial lag. The most important factor in interpreting them in the model is the topological distance, once weighted equally. In point of this fact, the best model selected an optimum descriptor combination, which includes van der Waals volumes and atomic polarizabilities as the most relevant key features (Table 1). This result illustrates that a certain distribution of these properties is necessary to distinguish the RSV inhibitors.
The last selected two descriptors (D513 and D528) belong to topological charge indices. D513, molecular topological order3 charge index (GGI3) represents the three eigenvalues of the corrected adjacency matrix of a molecule. D528, the mean molecular topological order8 charge index (JGI8), is a kind of Galvez topological charge index which evaluates the charge transfers between pairs of atoms and the global charge transfers in the molecule [^{55}]. Galvez charge indices GGIK and JGIK are computed as follows:
(3)
GGIK=∑i=1, j=i+1i=N−1, j=NCTijδ(k,Dij) 
(4)
JGIK=GGIKN−1 
From the aforementioned discussion, it can be seen that the activity of these RSV inhibitors is mainly influenced by several factors including the molecular branching index and atomic polarizabilities. These results are to some extent in agreement with the corresponding related experimental conclusions [^{12},^{13},^{18}]. For example, Yu et al. reported that the topology of the side chain of RSV inhibitors is important, while we also find that the corresponding descriptors (D299 and D347) play a part in RSV inhibition. The studies on a series of benzotriazole derivatives as RSV inhibitors [^{13}] revealed a broad tolerance for substituent size and functionality, our selected 2D autocorrelation descriptors also disclose such information. In reference [^{12}], the authors reported that the polar functionality provides considerable latitude to modulate both the pharmaceutical and pharmacokinetic properties, which is found also to be of considerable importance in the quest for orally effective RSV inhibitors. In addition, reference [^{18}] illustrated polarity in the oxime substituent in a series of compounds with potent antiviral activity in cell culture that combined good metabolic stability in vitro with high cell membrane permeability, and the descriptor D503 also depicts the role that atomic polarizabilities plays in RSV inhibition.
As expected, besides the robust, sparse and predictive features, an ideal classification model would still be interpretable. In many cases, gaining an intuitive interpretation of important features from the twodimensional QSAR is not always simple. For the present work, it should be pointed out that our explanations for the current descriptors are just broad due to nonlinear model types and abstract descriptors. However, in terms of developing a highly predictive classification model, the proposed VSRF model in this work could allow this task.
A large, diverse dataset of 216 RSV inhibitors collected from the literature [^{12}–^{18}] published by the same research group with converted molar pEC_{50} (−logEC_{50}) values ranging from less than 3.563 to 8.699 mole were used as the dataset in the present study. These EC_{50} values were the results of two experiments performed on consecutive weeks with the data from individual experiments shown in parentheses. Based on the inhibitory activity, the dataset is split into two classes, i.e., 107 low active compounds with pEC_{50} < 6.5 and 109 high active ones with pEC50 > 7.5. Table 4 depicts several representative compounds together with their classification labels. All information of the dataset with their diverse scaffolds of structures is provided in Table S1 (Supporting Information).
In the present work, the two dimensional structures of all RSV inhibitors were built with the ISIS/Draw 2.3 program [^{56}], and converted to SDF format by Open Babel software package (http://openbabel.sourceforge.net/). The final structures were transferred into Mold^{2} [^{24}], a free program available to public to calculate molecular descriptors. The Mold^{2} software package can calculate 777 molecular descriptors solely from 2D chemical structures, and the models generated using Mold^{2} descriptors were reported comparable to those generated using descriptors from the compared commercial software packages [^{24}]. In our work, all original 777 Mold^{2} molecular descriptors were calculated, and then underwent a preprocessing process (also called unsupervised selection of descriptors) as follows: (1) descriptors containing larger than 85% zero values were removed; (2) zero and near zero variance predictors were removed because such descriptors may cause the model to crash or the fit to be unstable; and (3) one of the two descriptors that have the absolute correlations above 0.95 was omitted. After these steps, the number of original descriptors was reduced to 272 for further research.
Rational division of an experimental SAR (structureactivity relationship) dataset into the respective training and test sets for model development and validation is very important. The methods often used include random sampling (RS), KennardStone (KS), Kmean clustering, and selforganizing map, etc. The basic rule should be that the points of the training set are distributed evenly within the whole area covered by the dataset, and that the condition of closeness of the test set points to those of the training set is satisfied [^{57}].
For the independent prediction set, we performed our selection on the basis of their distribution in the chemical space, which is defined by Kohonen neural network [^{58}]. The Kohonen neural network of dimension 6 × 6 was applied, which enables one to map objects into 36 positions. Similar objects were mapped into the same position (x, y coordinates in a Kohonen map). Only one part of a representative object from each position in the Kohonen map was chosen for the training set, respecting the original proportion among the different classes and the predefined 3:1 ratio between the training and the test objects. The rest were put into the test set. The selforganizing map simulations were carried out using internally developed Clanguage program. The training set was used for the development of the classification models, and the independent prediction set was used for the assessment of the system. The training and independent test sets contain 163 (81 low active and 82 high active) and 53 (26 low active and 27 high active) compounds, respectively, with approximately onefourth of the respective groups assigned in the independent prediction set.
VSRF: Random forest model was constructed according to the described original RF algorithm [^{59}]. RF is an ensemble of single decision trees, whose assembly produces a corresponding number of outputs and the outputs of all trees are aggregated to obtain one final prediction. The training algorithm of the RF for classification can be briefly summarized as follows: (1) Draw N bootstrap samples from the original training set. (2) Construct an unpruned tree T_{p} (p = 1, …, N) with each training set B_{p}. At each node, rather than choosing the best split among all predictors, randomly sample m_{try} of the predictors and then choose the best split from among those variables. The tree is grown to maximum size and not pruned back. (3) Predict the N trees by majority vote for classification. RF algorithm is the same as Bagging when m_{try} = p and the tree growing algorithm used in RF is CART (classification and regression tree). The RF algorithm can be efficient especially when the number of descriptors (p) is very large. This is because RF only tests the m_{try} of the descriptors rather than the p, where the default m_{try} is the square root of the number of descriptors for classification. Thus, m_{try} is very small, so that the search is very fast.
RF possesses its own reliable statistical characteristics based on OOB set prediction, which could be used for validation and model selection with no crossvalidation performed. It was shown that the prediction accuracy of an OOB set and a 5fold cross validation procedure was nearly the same [^{28}]. Although RF performs relatively well “off the shelf” without expending much effort on parameter tuning or variable selection [^{28}], it is also important for carrying out some tentative investigations on the changes of m_{try} or descriptor selection to optimize the performance of RF. In the current study, the optimal m_{try} was determined when the prediction accuracy reached the highest based on the 10fold crossvalidation.
Random forest, as a new classification and regression tool, has not been frequently applied in QSAR, QSPR (quantitative structureproperty relationship) [^{25},^{28},^{60},^{61}]. Thus it should be of value to investigate whether the RF can be applied to obtain better statistical performance for the current dataset of RSV inhibitors. Here, only a brief introduction about RF is presented, since more details can be found in corresponding literatures [^{28},^{59}]. In the present work, the RF algorithm was employed using the R package randomForest [^{62}].
As expected, an ideal classification model should possess high prediction ability with a small set of descriptors. Thus, variable selection with random forest was used to implement this task. Here, we simply introduce the VSRF. To select optimal descriptors, random forests were iteratively fitted, at each iteration building a new forest after discarding those descriptors with the smallest variable importance; the selected set of descriptor is the one that yields the smallest OOB error rate. In this algorithm, all forests result from eliminating, iteratively, a fraction, fraction.dropped, of the descriptors (the least important ones) used in the previous iteration. By default, fraction.dropped = 0.2, which allows for relatively fast operation, coherent with the idea of an “aggressive variable selection” approach, and increases the resolution as the number of descriptors considered becomes smaller. After fitting all forests, the OOB error rates from all the fitted random forests were examined. And the solution with the smallest number of descriptors whose error rate is within μ standard errors of the minimum error rate of all forests is chosen. Setting μ = 0 is the same as selecting the set of descriptors that leads to the smallest error rate. Setting μ = 1 is similar to the common “1 s.e. rule”, used in the classification trees [^{36}]. In our work, the μ = 1 was adopted, since this strategy can lead to solutions with fewer descriptors than selecting the solution with the smallest error rate, while achieving an error rate that is not different, within sampling error, from the “best solution”. More details on the VSRF can be found in literature [^{36}]. The variable selection from random forest was performed using the R package varSelRF [^{63}]. All parameters were adopted by default.
SVM: Support vector machines are a relatively new type of learning algorithm originally introduced by Vapnik and coworkers [^{64}]. Due to its many attractive features and promising empirical performances, SVM is gaining increasing popularity in many fields [^{65},^{66}], and thus was also performed in the present work. Since there have been a number of excellent introductions into SVM [^{35},^{64},^{67}], only a briefly description of the main idea of SVM classification is presented here.
For the classification task, briefly, this involves the optimization of Lagrangian multipliers α_{i} with constraints 0 ≤ α_{i} ≤ C and ∑α_{i} y_{i} = 0 to yield a decision function as follows:
(5)
f(x)=sign(∑i=1lyiαiK(x,xi)+b) 
GP: Preliminarily used in QSAR field, the Gaussian process (GP) was also introduced in the present study to classify the RSV inhibitors. Pioneering work was made by Burden [^{42}] who demonstrated GP applications in QSAR modeling of data sets of compounds active at the benzodiazepine and muscarinic receptors, etc. In addition, the authors of these references [^{68}–^{71}] have also reported the successful use of GP in statistical predictions of a series of pharmacokinetic properties. Recently, GP was also reported to be applied both in an automatic QSAR modeling of ADME (absorption, distribution, metabolism, excretion) properties [^{72}], and the multivariate spectroscopic calibration [^{73}]. All these works confirmed the possibility of GP as a promising machine learning tool, to be used in QSAR studies. In view of this, the present study is dedicated to introducing GP in classification modeling of RSV inhibitors.
A Gaussian process is defined simply as a collection of random variables which have a joint Gaussian distribution. It is completely characterized by its mean and covariance function. In the GP, the kernel function used in training and prediction contains (1) Radial Basis kernel function “Gaussian”; (2) Polynomial kernel function; (3) Linear kernel function; (4) Hyperbolic tangent kernel function; (5) Laplacian kernel function; (6) Bessel kernel function; (7) ANOVA RBF kernel function; and (8) Spline kernel. In the present work, the popular Radial Basis kernel function was chosen, with the kernel parameters determined by sigest function implemented in the R package kernlab.
LDA: LDA is a pattern recognition method providing a classification model based on the combination of variables that best predicts the category or group to which a given compounds belongs. The basic theory of LDA is to classify the dependents by dividing an ndimensional descriptor space into two regions that are separated by a hyperplane defined by a linear discriminant function. In this study, the independent variables were the calculated molecular descriptors, and the discrimination property was EC_{50} (represented by either high active or low active). Statistical analyses were performed using the R package MASS [^{74}].
kNN: kNN measures the Euclidean distance between a tobeclassified vector x and each individual vector x_{i} in the training set [^{75}]. A total of k number of vectors nearest to the vector x are used to determine its class, f(x):
(6)
f^(x)←arg maxv∈V ∑i=1kδ[v, f(xi)] 
As in the case of all discriminative methods [^{22},^{31}], the performance of statistical learning methods can be measured by a series of parameters including the quantity of true positives (TP), true negatives (TN), false positives (FP), false negatives (FN), sensitivity (SE) (also called recall), SE=TP/(TP + FN), which is the prediction accuracy for the high active compounds in this work, and specificity (SP), SP = TN/(TN + FP), which is the prediction accuracy for the low active inhibitors, Precision = TP/(TP + FP), which is the positive predictive value. The overall prediction accuracy (Q), Matthews correlation coefficient (MCC) and Fmeasure, a function of recall and precision which indicated the accuracy of real and estimated class, respectively, are also used to measure the prediction accuracies and can be given as follows:
(7)
Q=TP+TNTP+TN+FP+FN 
(8)
MCC=TP×TN−FN×FP(TP+FN)(TP+FP)(TN+FN)(TN+FP) 
(9)
F−measue=2×recall×precisionrecall+precision 
In the present work, based on the uptodate largest dataset (to our best knowledge) of 216 structurally diverse RSV inhibitors, a VSRF classification model with good predictive performance (the overall Q = 94.34% for the prediction set) has been built.
By explanation of the selected descriptors, we conclude that the topological structure and electronic factors play a central role in the RSV inhibition. Moreover, a comparison with four other statistical methods, i.e., SVM, GP, LDA and kNN, demonstates that the VSRF model presents better statistics both for the training and test sets. Through a comparison of RF statistical performance with and without variable selection based on these RSV inhibitors, the proposed VSRF method not only improves the prediction ability but also enhances computational efficiency. Therefore, we hope that this method and the derived model will be of help for predictive tasks to screen new and potent RSV inhibitors in early drug development.
This work is financially supported by the National Natural Science Foundation of China (Grant No. 10801025). The authors thank the R Development Core Team for affording the free R2.10 software.
References
1..  Chanock R,Roizman B,Myers R. Recovery from infants with respiratory illness of a virus related to chimpanzee coryza agent (CCA)Am. J. EpidemiolYear: 195766281290 
2..  Cianci C,Genovesi E,Lamb L,Medina I,Yang Z,Zadjura L,Yang H,D’Arienzo C,Sin N,Yu K. Oral efficacy of a respiratory syncytial virus inhibitor in rodent models of infectionAntimicrob. Agents ChemotherYear: 2004482448245415215093 
3..  Cianci C,Yu K,Combrink K,Sin N,Pearce B,Wang A,Civiello R,Voss S,Luo G,Kadow K. Orally active fusion inhibitor of respiratory syncytial virusAntimicrob. Agents ChemotherYear: 20044841342214742189 
4..  Greensill J,McNamara P,Dove W,Flanagan B,Smyth R,Hart C. Human metapneumovirus in severe respiratory syncytial virus bronchiolitisEmerg. Infect. DisYear: 2003937237512643835 
5..  Sigurs N,Gustafsson P,Bjarnason R,Lundberg F,Schmidt S,Sigurbergsson F,Kjellman B. Severe respiratory syncytial virus bronchiolitis in infancy and asthma and allergy at age 13Am. J. Respir. Crit. Care MedYear: 200517113714115516534 
6..  Hart R. An outbreak of respiratory syncytial virus infection in an old people’s homeJ. InfectYear: 198482592616736667 
7..  Falsey AR,Hennessey PA,Formica MA,Cox C,Walsh EE. Respiratory syncytial virus infection in elderly and highrisk adultsN. Engl. J. MedYear: 20053521749175915858184 
8..  Ding WD,Mitsner B,Krishnamurthy G,Aulabaugh A,Hess CD,Zaccardi J,Cutler M,Feld B,Gazumyan A,Raifeld Y,et al. Novel and specific respiratory syncytial virus inhibitors that target virus fusionJ. Med. ChemYear: 199841267126759667956 
9..  Sidwell R,Barnard D. Respiratory syncytial virus infections: Recent prospects for controlAntiviral ResYear: 20067137939016806515 
10..  Nikitenko A,Raifeld Y,Wang T. The discovery of RFI641 as a potent and selective inhibitor of the respiratory syncytial virusBioorg. Med. Chem. LettYear: 2001111041104411327584 
11..  Chapman J,Abbott E,Alber D,Baxter R,Bithell S,Henderson E,Carter M,Chambers P,Chubb A,Cockerill G. RSV604, a novel inhibitor of respiratory syncytial virus replicationAntimicrob. Agents ChemotherYear: 2007513346335317576833 
12..  Yu KL,Zhang Y,Civiello RL,Kadow KF,Cianci C,Krystal M,Meanwell NA. Fundamental structureactivity relationships associated with a new structural class of respiratory syncytial virus inhibitorBioorg. Med. Chem. LettYear: 2003132141214412798322 
13..  Yu KL,Zhang Y,Civiello RL,Trehan AK,Pearce BC,Yin Z,Combrink KD,Gulgeze HB,Wang XA,Kadow KF,et al. Respiratory syncytial virus inhibitors. Part 2: Benzimidazol2one derivativesBioorg. Med. Chem. LettYear: 2004141133113714980651 
14..  Yu KL,Wang XA,Civiello RL,Trehan AK,Pearce BC,Yin Z,Combrink KD,Gulgeze HB,Zhang Y,Kadow KF,et al. Respiratory syncytial virus fusion inhibitors. Part 3: Watersoluble benzimidazol2one derivatives with antiviral activity in vivoBioorg. Med. Chem. LettYear: 2006161115112216368233 
15..  Yu KL,Sin N,Civiello RL,Wang XA,Combrink KD,Gulgeze HB,Venables BL,Wright JJK,Dalterio RA,Zadjura L,et al. Respiratory syncytial virus fusion inhibitors. Part 4: Optimization for oral bioavailabilityBioorg. Med. Chem. LettYear: 20071789590117169560 
16..  Wang XA,Cianci CW,Yu KL,Combrink KD,Thuring JW,Zhang Y,Civiello RL,Kadow KF,Roach J,Li Z,et al. Respiratory syncytial virus fusion inhibitors. Part 5: Optimization of benzimidazole substitution patterns towards derivatives with improved activityBioorg. Med. Chem. LettYear: 2007174592459817576060 
17..  Combrink KD,Gulgeze HB,Thuring JW,Yu KL,Civiello RL,Zhang Y,Pearce BC,Yin Z,Langley DR,Kadow KF,et al. Respiratory syncytial virus fusion inhibitors. Part 6: An examination of the effect of structural variation of the benzimidazol2one heterocycle moietyBioorg. Med. Chem. LettYear: 2007174784479017616396 
18..  Sin N,Venables BL,Combrink KD,Gulgeze HB,Yu KL,Civiello RL,Thuring J,Wang XA,Yang Z,Zadjura L,et al. Respiratory syncytial virus fusion inhibitors. Part 7: Structureactivity relationships associated with a series of isatin oximes that demonstrate antiviral activity in vivoBioorg. Med. Chem. LettYear: 2009194857486219596574 
19..  Roy PP,Roy K. QSAR studies of CYP2D6 inhibitor aryloxypropanolamines using 2D and 3D descriptorsChem. Biol. Drug DesYear: 20097344245519291105 
20..  Hemmateenejad B,Miri R,Akhond M,Shamsipur M. QSAR study of the calcium channel antagonist activity of some recently synthesized dihydropyridine derivatives. An application of genetic algorithm for variable selection in MLR and PLS methodsChemom. Intell. Lab. SystYear: 2002649199 
21..  Agrafiotis D,Bandyopadhyay D,Wegner J,Van Vlijmen H. Recent advances in chemoinformaticsJ. Chem. Inf. ModelYear: 2007471279129317511441 
22..  Sun X,Li Y,Liu X,Ding J,Wang Y,Shen H,Chang Y. Classification of bioaccumulative and nonbioaccumulative chemicals using statistical learning approachesMol. DiversYear: 20081215716918937041 
23..  Roy K,Leonard LT. Classical QSAR modeling of antiHIV 2,3diaryl1,3thiazolidin4onesQSAR Comb. SciYear: 200524579592 
24..  Hong H,Xie Q,Ge W,Qian F,Fang H,Shi L,Su Z,Perkins R,Tong W. Mold^{2}, molecular descriptors from 2D structures for chemoinformatics and toxicoinformaticsJ. Chem. Inf. ModelYear: 2008481337134418564836 
25..  Hao M,Li Y,Wang Y,Zhang S. Prediction of PKCθ inhibitory activity using the random forest algorithmInt. J. Mol. SciYear: 2010113413343320957104 
26..  Wang Y,Li Y,Wang B. An in silico method for screening nicotine derivatives as cytochrome P450 2A6 selective inhibitors based on kernel partial least squaresInt. J. Mol. SciYear: 20078166179 
27..  Wang Z,Li Y,Ai C,Wang Y. In silico prediction of estrogen receptor subtype binding affinity and selectivity using statistical methods and molecular docking with 2arylnaphthalenes and 2arylquinolinesInt. J. Mol. SciYear: 2010113434345820957105 
28..  Svetnik V,Liaw A,Tong C,Culberson JC,Sheridan RP,Feuston BP. Random forest: A classification and regression tool for compound classification and QSAR modelingJ. Chem. Inf. Comput. SciYear: 2003431947195814632445 
29..  Obrezanova O,Segall M. Gaussian processes for classification: QSAR modeling of ADMET and target activityJ. Chem. Inf. ModelYear: 2010501053106120433177 
30..  Zhou P,Chen X,Wu Y,Shang Z. Gaussian process: An alternative approach for QSAM modeling of peptidesAmino AcidsYear: 20103819921219123053 
31..  Li Y,Wang Y,Ding J,Wang Y,Chang Y,Zhang S. In silico prediction of androgenic and nonandrogenic compounds using random forestQSAR Comb. SciYear: 200928396405 
32..  Pontes M,Galvãob R,Araújo M,Moreira P,Neto O,Joséa G,Saldanha T. The successive projections algorithm for spectral variable selection in classification problemsChemom. Intell. Lab. SystYear: 2005781118 
33..  Bakken G,Jurs P. Classification of multidrugresistance reversal agents using structurebased descriptors and linear discriminant analysisJ. Med. ChemYear: 2000434534454111087578 
34..  Pourbasheer E,Riahi S,Ganjali M,Norouzi P. QSAR study on melanocortin4 receptors by support vector machineEur. J. Med. ChemYear: 2010451087109320031282 
35..  Doucet JP,Barbault F,Xia HR,Panaye A,Fan B. Nonlinear SVM approaches to QSPR/QSAR studies and drug designCurr. Comput. Aided Drug DesYear: 20073263289 
36..  DíazUriarte R,Alvarez de Andrés S. Gene selection and classification of microarray data using random forestBMC bioinformaticsYear: 200673131316393334 
37..  Vesanto J,Alhoniemi E. Clustering of the selforganizing mapIEEE Trans. Neural NetworksYear: 200011586600 
38..  Zupan J,Novič M,Ruisánchez I. Kohonen and counterpropagation artificial neural networks in analytical chemistryChemom. Intell. Lab. SystYear: 199738123 
39..  Eriksson L,Jaworska J,Worth A,Cronin M,McDowell R,Gramatica P. Methods for reliability and uncertainty assessment and for applicability evaluations of classificationand regressionbased QSARsEnviron. Health PerspectYear: 20031111361137512896860 
40..  Kuhn M. caret: Classification and Regression Training. CRAN: Wien, Austria, 2010; Available online: http://cran.rproject.org/web/packages/caret/index.html (accessed on 11 February 2011).. 
41..  Karatzoglou A,Smola A,Hornik K. kernlab: Kernelbased Machine Learning Lab. CRAN: Wien, Austria, 2010; Available online: http://cran.rproject.org/web/packages/kernlab/index.html (accessed on 11 February 2011).. 
42..  Burden F. Quantitative structureactivity relationship studies using gaussian processesJ. Chem. Inf. Comput. SciYear: 20014183083511410065 
43..  Golbraikh A,Tropsha A. Beware of q^{2}!J. Mol. Graph. ModelYear: 20022026927611858635 
44..  Triballeau N,Acher F,Brabet I,Pin JP,Bertrand HO. Virtual screening workflow development guided by the “receiver operating characteristic” curve approach. Application to highthroughput docking on metabotropic glutamate receptor subtype 4J. Med. ChemYear: 2005482534254715801843 
45..  Bradley A. The use of the area under the ROC curve in the evaluation of machine learning algorithmsPattern RecognitYear: 19973011451159 
46..  Lovasz L,Pelikan J. On the eigenvalues of treesPeriodica Mathematica HungaricaYear: 19733175182 
47..  Helguera AM,RodriguezBorges JE,GarciaMera X,Fernandez F,Natalia M,Cordeiro DS. Probing the anticancer activity of nucleoside analogues: A QSAR model approach using an internally consistent training setJ. Med. ChemYear: 2007501537154517341060 
48..  Randić M,Wilkins C. Graph theoretical approach to recognition of structural similarity in moleculesJ. Chem. Inf. Comput. SciYear: 1979193137 
49..  SaízUrra L,González M,Teijeira M. 2Dautocorrelation descriptors for predicting cytotoxicity of naphthoquinone ester derivatives against oral human epidermoid carcinomaBioorg. Med. ChemYear: 2007153565357117368033 
50..  Caballero J,Garriga M,Fernández M. 2D Autocorrelation modeling of the negative inotropic activity of calcium entry blockers using Bayesianregularized genetic neural networksBioorg. Med. ChemYear: 2006143330334016442799 
51..  Bauknecht H,Zell A,Bayer H,Levi P,Wagener M,Sadowski J,Gasteiger J. Locating biologically active compounds in mediumsized heterogeneous datasets by topological autocorrelation vectors: Dopamine and benzodiazepine agonistsJ. Chem. Inf. Comput. SciYear: 199636120512138941996 
52..  Moreau G,Broto P. The autocorrelation of a topological structure: A new molecular descriptorNouv. J. ChimYear: 19804359360 
53..  Wagener M,Sadowski J,Gasteiger J. Autocorrelation of molecular surface properties for modeling corticosteroid binding globulin and cytosolic Ah receptor activity by neural networksJ. Am. Chem. SocYear: 199511777697775 
54..  Moran P. Notes on continuous stochastic phenomenaBiometrikaYear: 195037172315420245 
55..  Galvez J,Garcia R,Salabert MT,Soler R. Charge indexes. New topological descriptorsJ. Chem. Inf. Comput. SciYear: 199434520525 
56..  ISIS Draw 2.3. MDL Information Systems, Inc.: San Leandro, CA, USA, 2010 
57..  Golbraikh A,Tropsha A. Predictive QSAR modeling based on diversity sampling of experimental datasets for the training and test set selectionJ. Comput. Aided Mol. DesYear: 20021635736912489684 
58..  Kohonen T. The selforganizing mapProc. Inst. Electrical Electronics EngYear: 19907814641480 
59..  Breiman L. Random forestsMach. LearnYear: 200145532 
60..  Polishchuk PG,Muratov EN,Artemenko AG,Kolumbin OG,Muratov NN,Kuz’min VE. Application of random forest approach to QSAR prediction of aquatic toxicityJ. Chem. Inf. ModelYear: 2009492481248819860412 
61..  Palmer D,O’Boyle N,Glen R,Mitchell J. Random forest models to predict aqueous solubilityJ. Chem. Inf. ModelYear: 20074715015817238260 
62..  Breiman L,Cutler A,Liaw A,Wiener M. randomForest: Breiman and Cutler’s Random Forests for Classification and Regression. CRAN: Wien, Austria, 2010; Available online: http://cran.rproject.org/web/packages/randomForest/index.html (accessed on 11 February 2011).. 
63..  DiazUriarte R. varSelRF: Variable Selection Using Random Forests. CRAN: Wien, Austria, 2010; Available online: http://cran.rproject.org/web/packages/varSelRF/index.html (accessed on 11 February 2011).. 
64..  Guyon I,Weston J,Barnhill S,Vapnik V. Gene selection for cancer classification using support vector machinesMach. LearnYear: 200246389422 
65..  Riahi S,Pourbasheer E,Dinarvand R,Ganjali MR,Norouzi P. Exploring QSARs for antiviral activity of 4alkylamino6(2hydroxyethyl)2methylthiopyrimidines by support vector machineChem. Biol. Drug DesYear: 20087220521618715229 
66..  Kriegl JM,Arnhold T,Beck B,Fox T. A support vector machine approach to classify human cytochrome P450 3A4 inhibitorsJ. Comput. Aided Mol. DesYear: 20051918920116059671 
67..  Furey T,Cristianini N,Duffy N,Bedarski D,Schummer M,Haussler D. Support vector machine classification and validation of cancer tissue samples using microarray expression dataBioinformaticsYear: 20001690691411120680 
68..  Enot D,Gautier R,Marouille J. Gaussian process: An efficient technique to solve quantitative structureproperty relationship problemsSAR QSAR Environ. ResYear: 20011246146911813811 
69..  Tiño P,Nabney IT,Williams BS,Lösel J,Sun Y. Nonlinear prediction of quantitative structureactivity relationshipsJ. Chem. Inf. Comput. SciYear: 2004441647165315446822 
70..  Schwaighofer A,Schroeter T,Mika S,Laub J,Ter Laak A,Sülzle D,Ganzer U,Heinrich N,Müller K. Accurate solubility prediction with error bars for electrolytes: A machine learning approachJ. Chem. Inf. ModelYear: 20074740742417243756 
71..  Schroeter T,Schwaighofer A,Mika S,Ter Laak A,Suelzle D,Ganzer U,Heinrich N,Müller K. Predicting lipophilicity of drugdiscovery molecules using gaussian process modelsChem. Med. ChemYear: 200721265126717576646 
72..  Obrezanova O,Csányi G,Gola JMR,Segall MD. Gaussian processes: A method for automatic QSAR modeling of ADME propertiesJ. Chem. Inf. ModelYear: 2007471847185717602549 
73..  Chen T,Morris J,Martin E. Gaussian process regression for multivariate spectroscopic calibrationChemom. Intell. Lab. SystYear: 2007875971 
74..  MASS: Main Package of Venables and Ripley’s MASS. CRAN: Wien, Austria, 2010; Available online: http://cran.rproject.org/web/packages/MASS/index.html (accessed on 11 February 2011). 
75..  Gunturi SB,Narayanan R. In silico ADME modeling 3: Computational models to predict human intestinal absorption using sphere exclusion and kNN QSAR methodsQSAR Comb. SciYear: 200726653668 
Article Categories:
Keywords: RSV, variable selection, Mold^{2} descriptors, random forest. 
Previous Document: Molecular Arrangement in SelfAssembled AzobenzeneContaining Thiol Monolayers at the Individual Dom...
Next Document: Application of molecular topology for the prediction of reaction yields and antiinflammatory activi...