Prediction of factor Xa inhibitors by machine learning methods

Factor Xa (FXa) inhibitors have been explored as anticoagulants for treatment and prevention of thrombotic diseases. Molecular docking, pharmacophore, quantitative structure-activity relationships, and support vector machines (SVM) have been used for computer prediction of FXa inhibitors. These methods achieve promising prediction accuracies of 69-80% for FXa inhibitors and 85-99% for non-inhibitors. Prediction performance, particularly for inhibitors, may be further improved by exploring methods applicable to more diverse range of compounds and by using more appropriate set of molecular descriptors. We tested the capability of several machine learning methods (C4.5 decision tree, k-nearest neighbor, probabilistic neural network, and support vector machine) by using a much more diverse set of 1098 compounds (360 inhibitors and 738 non-inhibitors) than those in other studies. A feature selection method was used for selecting molecular descriptors appropriate for distinguishing FXa inhibitors and non-inhibitors. The prediction accuracies of these methods are 89.1-97.5% for FXa inhibitors and 92.3-98.1% for non-inhibitors. In particular, compared to other studies, support vector machine gives a substantially improved accuracy of 94.6% for FXa non-inhibitors and maintains a comparable accuracy of 98.1% for inhibitors, based-on a more rigorous test with more diverse range of compounds. Our study suggests that machine learning methods such as SVM are useful for facilitating the prediction of FXa inhibitors.

Bacillus chagannorensis sp. nov., a moderate halophile from a soda lake in Inner Mongolia, China

A Gram-positive, moderately halophilic, spore-forming bacterium, designated strain CG-15T, was isolated from a soda lake, Lake Chagannor, in the Inner Mongolia Autonomous Region, China. The cells were found to be motile short rods with ellipsoidal, terminal and deforming endospores. Strain CG-15T, a facultatively anaerobic bacterium, grew at pH 5.8–11.0 (optimally at pH 8.5), at 6–40 °C (optimally at 37 °C) and at salinities of 3–20 % (w/v) total salts (optimally at 7 % w/v). On the basis of the results of 16S rRNA gene sequence analysis, strain CG-15T was shown to belong to the genus Bacillus (phylum Firmicutes), showing the greatest phylogenetic similarity with respect to Bacillus saliphilus (96.0 %). The DNA G+C content of the novel isolate was found to be 53.8 mol%. The major cellular fatty acids of strain CG-15T were anteiso-C15 : 0, iso-C15 : 0 and anteiso-C17 : 0, and its polar lipids consisted of diphosphatidylglycerol, phosphatidylglycerol, phosphatidylethanolamine and three different unidentified phospholipids. The analysis of the quinones showed that MK-7 was the major menaquinone. The peptidoglycan type was A1γ, with meso-diaminopimelic acid as the diagnostic diamino acid. On the basis of the data from this polyphasic study, strain CG-15T represents a novel species of the genus Bacillus, for which the name Bacillus chagannorensis sp. nov. is proposed. The type strain is CG-15T (=CCM 7371T=CECT 7153T=CGMCC 1.6292T=DSM 18086T).

Functional characterization of TIP60 sumoylation in UV-irradiated DNA damage response

The histone acetyltransferase TIP60 regulates the DNA damage response following genotoxic stress by acetylating histone and remodeling chromatin. However, the molecular mechanisms underlying the TIP60-dependent response to UV-induced DNA damage remain poorly understood. To systematically analyse proteins that regulate TIP60 activity in response to UV irradiation, we performed a proteomic analysis of proteins selectively bound to TIP60 in response to UV irradiation using mass spectrometry and identified a novel regulatory mechanism by which TIP60 orchestrates transcriptional activation of p53-dependent checkpoint response in UV-irradiated cells. The initial step of this pathway involves UV-induced association of TIP60 with SUMO-conjugation enzymes and site-specific sumoylation of TIP60 at lysines 430 and 451 via Ubc9. This sumoylation initiates the relocation of TIP60 from nucleoplasm to the promyelocytic leukemia body, which is essential for the UV-irradiated DNA damage repair response via a p53-dependent pathway. Significantly, inhibition of TIP60 sumoylation by overexpression of non-sumoylatable mutant abrogates the p53-dependent DNA damage response, demonstrating the importance of TIP60 sumoylation in response to UV irradiation. Our biochemical characterization demonstrated that the sumoylation of TIP60 augments its acetyltransferase activity in vitro and in vivo. Thus, this study shed new light on the function and regulation of TIP60 activity in UV-irradiated DNA damage response.

Halovivax ruber sp. nov., an extremely halophilic archaeon isolated from Lake Xilinhot, Inner Mongolia, China

A Gram-negative, pleomorphic, extremely halophilic archaeon, designated strain XH-70(T), was isolated from the saline Lake Xilinhot, in Inner Mongolia, China. It formed small (0.9-1.5 mm), red-pigmented, elevated colonies on agar medium. The strain required at least 2.5 M NaCl and 5 mM Mg(2+) for growth. The 16S rRNA gene sequence analysis indicated that strain XH-70(T) belongs to the family Halobacteriaceae, showing 99.5 % similarity to the type strain of Halovivax asiaticus and 94.7 and 94.6 % similarity, respectively, to the type strains of Natronococcus amylolyticus and Natronococcus occultus. Polar lipid analysis supported the placement of strain XH-70(T) in the genus Halovivax. DNA-DNA hybridization studies (32 % with Halovivax asiaticus CGMCC 1.4248(T)), as well as biochemical and physiological characterization, allowed strain XH-70(T) to be differentiated from Halovivax asiaticus. A novel species, Halovivax ruber sp. nov., is therefore proposed to accommodate this strain. The type strain is XH-70(T) (=CGMCC 1.6204(T)=DSM 18193(T)=JCM 13892(T)).

Aquisalimonas asiatica gen. nov., sp. nov., a moderately halophilic bacterium isolated from an alkaline, saline lake in Inner Mongolia, China

Two novel moderately halophilic, Gram-negative rods (strains CG12T and CG13) were isolated from Lake Chagannor in Inner Mongolia Autonomous Region, China. They were strictly aerobic and motile. They grew at pH 6.0–10.8 (optimally at pH 7.5–8.5), at 20–50 °C (optimally at 37 °C) and at salinities of 1–20 % (w/v) total salts (optimally at 7–10 %, w/v). Phylogenetic analysis of the two strains, based on a comparison of their 16S rRNA genes, led to their classification within the class Gammaproteobacteria, the closest recognized type strain being Alkalispirillum mobile DSM 12769T, with which they were found to share 94.4–94.6 % sequence similarity. On the basis of DNA–DNA hybridization data (showing 100 and 99 % relatedness for each other), the two isolates were found to be members of the same species. The DNA G+C contents of strains CG12T and CG13 were found to be 63.6 and 64.0 mol%, respectively. The major cellular fatty acids of strain CG12T, selected as the representative strain, were C18 : 1 ω7c, C16 : 0 and C12 : 0, and its polar lipids consisted of phosphatidylglycerol, diphosphatidylglycerol, phosphatidylethanolamine, phosphatidylcholine, a phosphoglycolipid and six unidentified phospholipids. On the basis of the polyphasic evidence from this study, strains CG12T and CG13 represent a novel genus and species, for which the name Aquisalimonas asiatica gen. nov., sp. nov. is proposed. The type strain of Aquisalimonas asiatica is CG12T (=CCM 7368T=CECT 7151T=CGMCC 1.6291T=DSM 18102T).

Application of support vector machines to in silico prediction of cytochrome p450 enzyme substrates and inhibitors

Cytochrome P450 enzymes are responsible for phase I metabolism of the majority of drugs and xenobiotics. Identification of the substrates and inhibitors of these enzymes is important for the analysis of drug metabolism, prediction of drug-drug interactions and drug toxicity, and the design of drugs that modulate cytochrome P450 mediated metabolism. The substrates and inhibitors of these enzymes are structurally diverse. It is thus desirable to explore methods capable of predicting compounds of diverse structures without over-fitting. Support vector machine is an attractive method with these qualities, which has been employed for predicting the substrates and inhibitors of several cytochrome P450 isoenzymes as well as compounds of various other pharmacodynamic, pharmacokinetic, and toxicological properties. This article introduces the methodology, evaluates the performance, and discusses the underlying difficulties and future prospects of the application of support vector machines to in silico prediction of cytochrome P450 substrates and inhibitors.

Halorubrum orientale sp. nov., a halophilic archaeon isolated from Lake Ejinor, Inner Mongolia, China

A motile, pleomorphic, red-pigmented archaeon, strain EJ-52T, was isolated from water from Lake Ejinor, a saline lake in Inner Mongolia, China. Analysis of the almost-complete 16S rRNA gene sequence showed that the isolate was phylogenetically related to species of the genus Halorubrum, being most closely related to Halorubrum saccharovorum ATCC 29252T (96.1% sequence similarity), Halorubrum lacusprofundi JCM 8891T (95.9%), Halorubrum tibetense AS 1.3239T (95.2%), Halorubrum alcaliphilum AS 1.3528T (95.2%) and Halorubrum vacuolatum JCM 9060T (95.1%). The polar lipids of strain EJ-52T were C20C20 derivatives of phosphatidylglycerol phosphate and phosphatidylglycerol phosphate methyl ester and a sulfated diglycosyl diether. Strain EJ-52T requires at least 2.5 M NaCl for growth and grows optimally at 3.4 M NaCl. The strain grows at 25-50 degrees C, with optimal growth occurring at 35-45 degrees C. Mg2+ is not required. The DNA G+C content is 64.2 mol%. On the basis of the data obtained in this study, strain EJ52T represents a novel species, for which the name Halorubrum orientale sp. nov. is proposed. The type strain is EJ-52T (=CECT 7145T=JCM 13889T=CGMCC 1.6295T).

MODEL—molecular descriptor lab: A web-based server for computing structural and physicochemical features of compounds

Molecular descriptors represent structural and physicochemical features of compounds. They have been extensively used for developing statistical models, such as quantitative structure activity relationship (QSAR) and artificial neural networks (NN), for computer prediction of the pharmacodynamic, pharmacokinetic, or toxicological properties of compounds from their structure. While computer programs have been developed for computing molecular descriptors, there is a lack of a freely accessible one. We have developed a web-based server, MODEL (Molecular Descriptor Lab), for computing a comprehensive set of 3,778 molecular descriptors, which is significantly more than the approximately 1,600 molecular descriptors computed by other software. Our computational algorithms have been extensively tested and the computed molecular descriptors have been used in a number of published works of statistical models for predicting variety of pharmacodynamic, pharmacokinetic, and toxicological properties of compounds. Several testing studies on the computed molecular descriptors are discussed. MODEL is accessible at http://jing.cz3.nus.edu.sg/cgi-bin/model/model.cgi free of charge for academic use.

Classification of a diverse set of Tetrahymena pyriformis toxicity chemical compounds from molecular descriptors by statistical learning methods

Toxicity of various compounds has been measured in many studies by their toxic effects against Tetrahymena pyriformis. Efforts have also been made to use computational quantitative structure-activity relationship (QSAR) and statistical learning methods (SLMs) for predicting Tetrahymena pyriformis toxicity (TPT) at impressive accuracies. Because of the diversity of compounds and toxicity mechanisms, it is desirable to explore additional methods and to examine if these methods are applicable to more diverse sets of compounds. We tested several SLMs (logistic regression, C4.5 decision tree, k-nearest neighbor, probabilistic neural network, support vector machines) for their capability in predicting TPT by using 1129 compounds (841 TPT and 288 non-TPT agents) which are more diverse than those in other studies. A feature selection method was used for improving prediction performance and selecting molecular descriptors responsible for distinguishing TPT and non-TPT agents. The prediction accuracies are 86.9% approximately 94.2% for TPT and 71.2% approximately 87.5% for non-TPT agents based on 5-fold cross-validation studies, which are comparable to some of earlier studies despite the use of more diverse sets of compounds. The selected molecular descriptors are consistent with those used in other studies and experimental findings. These suggest that SLMs are useful for predicting TPT potential of diverse sets of compounds and for characterizing the molecular descriptors associated with TPT.

Prediction of estrogen receptor agonists and characterization of associated molecular descriptors by statistical learning methods

Specific estrogen receptor (ER) agonists have been used for hormone replacement therapy, contraception, osteoporosis prevention, and prostate cancer treatment. Some ER agonists and partial-agonists induce cancer and endocrine function disruption. Methods for predicting ER agonists are useful for facilitating drug discovery and chemical safety evaluation. Structure-activity relationships and rule-based decision forest models have been derived for predicting ER binders at impressive accuracies of 87.1-97.6% for ER binders and 80.2-96.0% for ER non-binders. However, these are not designed for identifying ER agonists and they were developed from a subset of known ER binders. This work explored several statistical learning methods (support vector machines, k-nearest neighbor, probabilistic neural network and C4.5 decision tree) for predicting ER agonists from comprehensive set of known ER agonists and other compounds. The corresponding prediction systems were developed and tested by using 243 ER agonists and 463 ER non-agonists, respectively, which are significantly larger in number and structural diversity than those in previous studies. A feature selection method was used for selecting molecular descriptors responsible for distinguishing ER agonists from non-agonists, some of which are consistent with those used in other studies and the findings from X-ray crystallography data. The prediction accuracies of these methods are comparable to those of earlier studies despite the use of significantly more diverse range of compounds. SVM gives the best accuracy of 88.9% for ER agonists and 98.1% for non-agonists. Our study suggests that statistical learning methods such as SVM are potentially useful for facilitating the prediction of ER agonists and for characterizing the molecular descriptors associated with ER agonists.

Natrinema ejinorense sp. nov., isolated from a saline lake in Inner Mongolia, China

A Gram-negative, non-motile, neutrophilic, pleomorphic and extremely halophilic archaeon, strain EJ-57T, was isolated from saline Lake Ejinor in Inner Mongolia, China. Strain EJ-57T was able to grow at 25–50 °C, required at least 1.8 M NaCl for growth (optimum at 3.4 M NaCl) and grew over a pH range from 6.0 to 8.5 (optimum at pH 7.0). Hypotonic treatment with less than 1.5 M NaCl caused cell lysis. Analysis of the almost complete 16S rRNA gene sequence indicated that the isolate represented a member of the genus Natrinema in the family Halobacteriaceae. Strain EJ-57T was most closely related to Natrinema versiforme JCM 10478T (96.2 % sequence similarity), Natrinema pallidum NCIMB 777T (95.9 % sequence similarity), Natrinema altunense JCM 12890T (95.8 % sequence similarity) and Natrinema pellirubrum NCIMB 786T (95.5 % sequence similarity). However, DNA–DNA hybridization experiments showed that strain EJ-57T was not related to these species, with levels of DNA–DNA relatedness equal to or below 39 %. The major polar lipids of the isolate were C20C20 and C20C25 derivatives of phosphatidylglycerol, phosphatidylglycerol phosphate methyl ester and the disulfated glycolipid S2-DGA-1. The G+C content of the genomic DNA was 64.7 mol%. Comparative analysis of phenotypic characteristics between strain EJ-57T and recognized Natrinema species supported the conclusion that EJ-57T represents a novel species within this genus, for which the name Natrinema ejinorense sp. nov. is proposed. The type strain is EJ-57T (=CECT 7144T=JCM 13890T=CGMCC 1.6202T).

CYP2A6 polymorphism reveals differences in Japan and the existence of a specific variant in Ovambo and Turk populations

CYP2A6 is a polymorphic enzyme, and CYP2A6 genotype has been shown to be associated with smoking habits and lung cancer. We investigated CYP2A6 polymorphism in Japanese from four different geographic areas of Japan and in the Ovambo and Turk populations. Using two polymerase chain reaction restriction fragment length polymorphisms (PCR-RFLPs), we identified the functionally important variants of CYP2A6: *1A, *1B, *1F, *1G, *4A, and *4D. In the Japanese population the highest frequencies of the CYP2A6*1A allele were observed in subjects from the Fukuoka (Kyushu Island) and Ehime (Shikoku Island) prefectures, whereas subjects in Shimane and Tottori (both located on the Japan Sea side of Honshu Island) showed the highest frequencies of the CYP2A6*1B allele. In the Tottori and Shimane groups no subject was homozygous for the CYP2A6*4A allele, a whole gene deletion type that is prevalent among Asians. In the Ovambo and Turk populations the CYP2A6*1A allele was predominant. Furthermore, two alleles undetected in the Japanese were observed in these latter two ethnic groups: CYP2A6*1G was found solely in the Ovambos, and CYP2A6*1F was found solely in the Turks. The present study is the first to show interprefecture differences in CYP2A6 polymorphism in Japanese who live in relatively close but distinct geographic areas; this is also the first study to evaluate CYP2A6 variations among these Japanese and the Ovambo and Turk populations. The distribution results of these alleles could help to define the true significance of CYP2A6 polymorphism as a genetic susceptibility marker in worldwide populations.

Combination chemotherapy for high-risk gestational trophoblastic tumour

BACKGROUND

Gestational trophoblastic disease (GTD) includes gestational trophoblastic tumour and hydatidiform mole. Many women of reproductive age are affected by this disease although its incidence differs by geographical location. A number of chemotherapy regimens are used for treating the disease, such as methotrexate, actinomycin D and cyclophosphamide (MAC), methotrexate, actinomycin D, cyclophosphamide, doxorubicin, melphalan, hydroxyurea and vincristine (CHAMOC), etoposide, methotrexate and actinomycin (EMA) plus cyclophosphamide and vincristine (CO) (EMA-CO), etoposide, methotrexate and actinomycin (EMA) plus etoposide and cisplatin(EP) (EMA-EP). The efficacy of these drugs has not been systematically reviewed.

OBJECTIVES

To determine the efficacy and safety of combination chemotherapy in treating high-risk GTT.

SEARCH STRATEGY

Electronic searches of MEDLINE, EMB, Cochrane Central Register of Controlled Trials (CENTRAL) and CBM were carried out. Four journals were handsearched and other searching methods were used for identifying more studies.

SELECTION CRITERIA

The review included randomized controlled trials (RCTs) or quasi-RCTs of combination chemotherapy for treating high-risk GTT. Patients with placental-site trophoblastic tumour (PSTT), who had received chemotherapy in the previous two weeks, or patients with chemotherapy intolerance were excluded.

DATA COLLECTION AND ANALYSIS

Two investigators independently collected data using a data extraction form. Meta-analysis was not performed and the review was conducted as a narrative review.

MAIN RESULTS

One study with 42 participants was included in this review. It indicated that a MAC regimen was better than a CHAMOCA regimen for high-risk GTT because of lower toxicity. The quality of the study was unclear.

AUTHORS' CONCLUSIONS

The methodological limitations of the included study prevent any firm conclusions about the best combination chemotherapy regimen for high-risk GTT. High quality studies are required.

Halostagnicola larsenii gen. nov., sp. nov., an extremely halophilic archaeon from a saline lake in Inner Mongolia, China

Strain XH-48T was isolated from the sediment of Lake Xilinhot, a saline lake in Inner Mongolia (China). The organism is pleomorphic, neutrophilic and requires at least 2.5 M (15 %) NaCl, but not MgCl2, for growth; it exhibits optimal growth at 3.4 M (20 %) NaCl. The G+C content of its DNA is 61 mol%. 16S rRNA gene sequence analysis revealed that strain XH-48T is a member of the family Halobacteriaceae, but there were low levels of similarity with other members of this family. The highest sequence similarity values (94.5 and 93.3 %) were obtained with the 16S rRNA genes of Natrialba aegyptiaca and Natrialba asiatica, respectively. Polar lipid analyses revealed that strain XH-48T contains phosphatidylglycerol and phosphatidylglyceromethylphosphate, derived from both C20C20 and C20C25 glycerol diethers, and two unidentified glycolipids. On the basis of the data obtained, the novel isolate cannot be classified within any recognized genus. Strain XH-48T should be placed within a novel genus and species within the family Halobacteriaceae, order Halobacteriales, for which the name Halostagnicola larsenii gen. nov., sp. nov. is proposed. The type strain of Halostagnicola larsenii is strain XH-48T (=DSM 17691T=CGMCC 1.5338T=JCM 13463T=CECT 7116T).

Prediction of compounds with specific pharmacodynamic, pharmacokinetic or toxicological property by statistical learning methods

Computational methods for predicting compounds of specific pharmacodynamic, pharmacokinetic, or toxicological property are useful for facilitating drug discovery and drug safety evaluation. The quantitative structure-activity relationship (QSAR) and quantitative structure-property relationship (QSPR) methods are the most successfully used statistical learning methods for predicting compounds of specific property. More recently, other statistical learning methods such as neural networks and support vector machines have been explored for predicting compounds of higher structural diversity than those covered by QSAR and QSPR. These methods have shown promising potential in a number of studies. This article is intended to review the strategies, current progresses and underlying difficulties in using statistical learning methods for predicting compounds of specific property. It also evaluates algorithms commonly used for representing structural and physicochemical properties of compounds.

Spin-dependent transport through a magnetic carbon nanotube-molecule junction

The electronic structure and spin-dependent conductance of a magnetic junction consisting of two Fe-doped carbon nanotubes and a C60 molecule are investigated using a first-principles approach that combines the density functional theory with the nonequilibrium Greens function technique. The tunneling magnetoresistance ratio is found to be 11%. The density of states and transmission coefficient through the molecular junction are analyzed and compared to layered magnetic tunneling junctions. Our findings suggest new possibilities for experiments and for future technology.

One-step purification of mammalian deoxyribonucleases I and differences among pancreas, parotid, and pancreas-parotid (mixed) types based on species-and organ-specific N-linked glycosylation

Mammalian deoxyribonucleases I (DNase I) are classified into three types, namely, pancreas, parotid, and pancreas-parotid (mixed), based on differences in their tissue concentrations. In this study, DNase I purification by concanavalin A-wheat germ agglutinin mixture-agarose column from rat (parotid type), rabbit (mixed type), and pig (pancreas type) is described. This method permits a relatively easy one-step purification of DNase I from rat and rabbit parotid glands, the rat submaxillary gland, and porcine pancreas. To elucidate differences among the three types, these DNases I were subjected to enzymatic deglycosylation either by peptide N-glycosidase F (PNGase F) or endoglycosidase H (Endo H). Following deglycosylation, digests were separated on DNA-casting polyacrylamide gel electrophoresis. PNGase F produced a single lower mobility product in all samples. Endo H produced a double band in rat and rabbit parotid glands and porcine pancreas, and a single band in the rabbit pancreas corresponding with the PNGase F product. DNase I activity of the porcine pancreas was completely extinguished by deglycosylation, while that of the parotid glands and rabbit pancreas was unaffected. Our results suggest that the distinct properties of DNase I exhibited by the three types may be attributed to differences in the extent of post-translational N-linked glycosylation of the enzyme.

Effect of molecular descriptor feature selection in support vector machine classification of pharmacokinetic and toxicological properties of chemical agents

Statistical-learning methods have been developed for facilitating the prediction of pharmacokinetic and toxicological properties of chemical agents. These methods employ a variety of molecular descriptors to characterize structural and physicochemical properties of molecules. Some of these descriptors are specifically designed for the study of a particular type of properties or agents, and their use for other properties or agents might generate noise and affect the prediction accuracy of a statistical learning system. This work examines to what extent the reduction of this noise can improve the prediction accuracy of a statistical learning system. A feature selection method, recursive feature elimination (RFE), is used to automatically select molecular descriptors for support vector machines (SVM) prediction of P-glycoprotein substrates (P-gp), human intestinal absorption of molecules (HIA), and agents that cause torsades de pointes (TdP), a rare but serious side effect. RFE significantly reduces the number of descriptors for each of these properties thereby increasing the computational speed for their classification. The SVM prediction accuracies of P-gp and HIA are substantially increased and that of TdP remains unchanged by RFE. These prediction accuracies are comparable to those of earlier studies derived from a selective set of descriptors. Our study suggests that molecular feature selection is useful for improving the speed and, in some cases, the accuracy of statistical learning methods for the prediction of pharmacokinetic and toxicological properties of chemical agents.

Prediction of Genotoxicity of Chemical Compounds by Statistical Learning Methods

Various toxicological profiles, such as genotoxic potential, need to be studied in drug discovery processes and submitted to the drug regulatory authorities for drug safety evaluation. As part of the effort for developing low cost and efficient adverse drug reaction testing tools, several statistical learning methods have been used for developing genotoxicity prediction systems with an accuracy of up to 73.8% for genotoxic (GT+) and 92.8% for nongenotoxic (GT-) agents. These systems have been developed and tested by using less than 400 known GT+ and GT- agents, which is significantly less in number and diversity than the 860 GT+ and GT- agents known at present. There is a need to examine if a similar level of accuracy can be achieved for the more diverse set of molecules and to evaluate other statistical learning methods not yet applied to genotoxicity prediction. This work is intended for testing several statistical learning methods by using 860 GT+ and GT- agents, which include support vector machines (SVM), probabilistic neural network (PNN), k-nearest neighbor (k-NN), and C4.5 decision tree (DT). A feature selection method, recursive feature elimination, is used for selecting molecular descriptors relevant to genotoxicity study. The overall accuracies of SVM, k-NN, and PNN are comparable to and those of DT lower than the results from earlier studies, with SVM giving the highest accuracies of 77.8% for GT+ and 92.7% for GT- agents. Our study suggests that statistical learning methods, particularly SVM, k-NN, and PNN, are useful for facilitating the prediction of genotoxic potential of a diverse set of molecules.