Classification tree methods for development of decision rules for botulism and cyanide poisoning

Introduction

Identification of predictors of potential mass poisonings may increase the speed and accuracy with which patients are recognized, potentially reducing the number ultimately exposed and the degree to which they are affected. This analysis used a decision-tree method to sort such potential predictors.

Methods

Data from the Toxic Exposure Surveillance System were used to select cyanide and botulism cases from 1993 to 2005 for analysis. Cases of other poisonings from a single poison center were used as controls. After duplication was omitted and removal of cases from the control sample was completed, there remained 1,122 cyanide cases, 262 botulism cases, and 70,804 controls available for both analyses. Classification trees for each poisoning type were constructed, using 131 standardized clinical effects. These decision rules were compared with the current case surveillance definitions of one active poison center and the American Association of Poison Control Centers (AAPCC).

Results

The botulism analysis produced a 4-item decision rule with Sensitivity (Se) of 68% and Specificity (Sp) of 90%. Use of the single poison center and AAPCC definitions produced Se of 19.5% and 16.8%, and Sp of 99.5% and 83.2%, respectively. The cyanide analysis produced a 9-item decision rule with Se of 74% and Sp of 77%. The single poison center and AAPCC case definitions produced Se of 10.2% and 8.6%, and Sp of 99.8% and 99.8%, respectively.

Conclusions

These results suggest the possibility of improved poisoning case surveillance sensitivity using classification trees. This method produced substantially higher sensitivities, but not specificities, for both cyanide and botulism. Despite limitations, these results show the potential of a classification-tree approach in the detection of poisoning events.     

Bioanalytical data in decision making: discovery and development.

1. Bioanalysis is traditionally associated with the development phase of drugs; its use in discovery programmes is often ignored but can have a major impact. 2. Pharmacokinetic studies conducted in conjunction with pharmacology screening can provide additional information to that considered in conventional structure activity relationships. Such factors as half-life and bioavailability can be critical in designing improved drugs. 3. Analytical methods in discovery programmes may differ from those used in later development work: for instance bioassay allows a common assay system for a large number of project compounds. Moreover its use, when combined with conventional methods, such as h.p.l.c., allows active metabolites to be readily detected. 4. Bioanalytical data generated in discovery and pre-clinical programmes are a valuable guide to early clinical programmes. Plasma concentration-response data from these programmes can be compared with those obtained in man. Such comparisons are particularly valuable during the phase one-initial dose escalation study. To maximize this it is our practice to generate pharmacokinetic data between each dose increase.     

In silico tools used for compound selection during target-based drug discovery and development

Introduction: The target-based drug discovery process, including target selection, screening, hit-to-lead (H2L) and lead optimization stage gates, is the most common approach used in drug development. The full integration of in vitro and/or in vivo data with in silico tools across the entire process would be beneficial to R&D productivity by developing effective selection criteria and drug-design optimization strategies.

Areas covered: This review focuses on understanding the impact and extent in the past 5 years of in silico tools on the various stage gates of the target-based drug discovery approach.

Expert opinion: There are a large number of in silico tools available for establishing selection criteria and drug-design optimization strategies in the target-based approach. However, the inconsistent use of in vitro and/or in vivo data integrated with predictive in silico multiparameter models throughout the process is contributing to R&D productivity issues. In particular, the lack of reliable in silico tools at the H2L stage gate is contributing to the suboptimal selection of viable lead compounds. It is suggested that further development of in silico multiparameter models and organizing biologists, medicinal and computational chemists into one team with a single accountable objective to expand the utilization of in silico tools in all phases of drug discovery would improve R&D productivity.     

An oil-free microemulsion for intravenous delivery of diallyl trisulfide: formulation and evaluation

The aim of the present study was to develop an oil-free o/w microemulsion, Cremophor EL:ethanol-propylene glycol:saline, for diallyl trisulfide (DATS) for intravenous (i.v.) administration to modify the safety and pharmacokinetics of DATS. The ternary diagram was constructed to identify the regions of dilutable microemulsions, and the optimal composition of microemulsion was determined by evaluation of injection safety such as hemolysis, intravenous stimulation and injection anaphylaxis compared to commercial formulation Chentian(?). Promising microemulsion with modified injection safety was developed that could incorporate 100 mg/g of DATS. The droplet size of the microemulsion was about 26 nm in diameter with narrow distribution (polydispersity index: 0.14). Acute toxicity test showed that median lethal dose (LD(50)) of DATS microemulsion was 1.69-fold higher than that of Chentian(?). Pharmacokinetics was assessed by comparing with the commercial injection after intravenous administration to rats at a dose of 30 mg/kg. The developed microemulsion showed significant higher area under the drug concentration-time curve and lower clearance and distribution volume than those of Chentian(?) (p<0.05). This helped DATS to reach higher level in vessel, and circulate in the blood stream for a longer time resulting in better therapeutic effect. In conclusion, microemulsion would be a promising intravenous delivery system for DATS.     

Different decision tree induction strategies for a medical decision problem

The paper presents a comparative study of selected recognition methods for the medical decision problem -acute abdominal pain diagnosis. We consider if it is worth using expert knowledge and learning set at the same time. The article shows two groups of decision tree approaches to the problem under consideration. The first does not use expert knowledge and generates classifier only on the basis of learning set. The second approach utilizes expert knowledge for specifying the decision tree structure and learning set for determining mode of decision making in each node based on Bayes decision theory. All classifiers are evaluated on the basis of computer experiments.     

Method development and validation for the GC-FID assay of p-cymene in tea tree oil formulation

This paper describes the development and validation of an isothermal gas chromatography-flame ionisation detection (GC-FID) method for the assay of pure tea tree oil. The chromatographic conditions of the method employ a 5% carbowax packed column (20 m x 0.25 mm), isothermal elution with hydrogen at a column flow of 36 ml/min, injector and detector temperature at 220 degrees C and oven temperature at 100 degrees C, and a 1.5 microl injection volume. Samples and standard were diluted in hexane. The calibration curve for p-cymene was linear (r2=0.9995) from 20 to 120% range of the analytical concentration of 100 microg/ml. The precision of this method was calculated as the relative standard deviation (R.S.D.) was 0.66% (n=6). The R.S.D. for intermediate precision study was 0.13 and recovery of the p-cymene ranged between 93.39 and 97.86%. The limits of detection and quantitation were determined to be 2.08 and 10.39 ng/ml, respectively.     

The European Lead Factory: An updated HTS compound library for innovative drug discovery

Through the European Lead Factory model, industry-standard high-throughput screening and hit validation are made available to academia, small and medium-sized enterprises, charity organizations, patient foundations, and participating pharmaceutical companies. The compound collection used for screening is built from a unique diversity of sources. It brings together compounds from companies with different therapeutic area heritages and completely new compounds from library synthesis. This generates structural diversity and combines molecules with complementary physicochemical properties. In 2019, the screening library was updated to enable another 5 years of running innovative drug discovery projects. Here, we investigate the physicochemical and diversity properties of the updated compound collection. We show that it is highly diverse, drug-like, and complementary to commercial screening libraries.     

Decision tree models for data mining in hit discovery

INTRODUCTION: Decision tree induction (DTI) is a powerful means of modeling data without much prior preparation. Models are readable by humans, robust and easily applied in real-world applications, features that are mutually exclusive in other commonly used machine learning paradigms. While DTI is widely used in disciplines ranging from economics to medicine, they are an intriguing option in pharmaceutical research, especially when dealing with large data stores. AREAS COVERED: This review covers the automated technologies available for creating decision trees and other rules efficiently, even from large datasets such as chemical libraries. The authors discuss the need for properly documented and validated models. Lastly, the authors cover several case studies in hit discovery, drug metabolism and toxicology, and drug surveillance, and compare them with other established techniques. EXPERT OPINION: DTI is a competitive and easy-to-use tool in basic research as well as in hit and drug discovery. Its strengths lie in its ability to handle all sorts of different data formats, the visual nature of the models, and the small computational effort needed for implementation in real-world systems. Limitations include lack of robustness and over-fitted models for certain types of data. As with any modeling technique, proper validation and quality measures are of utmost importance.     

Decision tree models for data mining in hit discovery

Introduction: Decision tree induction (DTI) is a powerful means of modeling data without much prior preparation. Models are readable by humans, robust and easily applied in real-world applications, features that are mutually exclusive in other commonly used machine learning paradigms. While DTI is widely used in disciplines ranging from economics to medicine, they are an intriguing option in pharmaceutical research, especially when dealing with large data stores.

Areas covered: This review covers the automated technologies available for creating decision trees and other rules efficiently, even from large datasets such as chemical libraries. The authors discuss the need for properly documented and validated models. Lastly, the authors cover several case studies in hit discovery, drug metabolism and toxicology, and drug surveillance, and compare them with other established techniques.

Expert opinion: DTI is a competitive and easy-to-use tool in basic research as well as in hit and drug discovery. Its strengths lie in its ability to handle all sorts of different data formats, the visual nature of the models, and the small computational effort needed for implementation in real-world systems. Limitations include lack of robustness and over-fitted models for certain types of data. As with any modeling technique, proper validation and quality measures are of utmost importance.     

布洛芬静脉注射制剂的开发与临床应用

非甾体类解热镇痛药大多为口服制剂,在口服困难的患者中存在很大的缺陷,2009年美国FDA首次批准上市了首个治疗疼痛和发热的布洛芬静脉注射制剂。本文介绍了布洛芬注射液的制剂开发、药理学、药动学、临床应用、不良反应以及与其他药物的相互作用等,供临床参考,以促进临床安全合理用药。     

Development of a decision tree to classify the most accurate tissue-specific tissue to plasma partition coefficient algorithm for a given compound

Physiologically based pharmacokinetic (PBPK) modeling is a tool used in drug discovery and human health risk assessment. PBPK models are mathematical representations of the anatomy, physiology and biochemistry of an organism and are used to predict a drug’s pharmacokinetics in various situations. Tissue to plasma partition coefficients (Kp), key PBPK model parameters, define the steady-state concentration differential between tissue and plasma and are used to predict the volume of distribution. The experimental determination of these parameters once limited the development of PBPK models; however, in silico prediction methods were introduced to overcome this issue. The developed algorithms vary in input parameters and prediction accuracy, and none are considered standard, warranting further research. In this study, a novel decision-tree-based Kp prediction method was developed using six previously published algorithms. The aim of the developed classifier was to identify the most accurate tissue-specific Kp prediction algorithm for a new drug. A dataset consisting of 122 drugs was used to train the classifier and identify the most accurate Kp prediction algorithm for a certain physicochemical space. Three versions of tissue-specific classifiers were developed and were dependent on the necessary inputs. The use of the classifier resulted in a better prediction accuracy than that of any single Kp prediction algorithm for all tissues, the current mode of use in PBPK model building. Because built-in estimation equations for those input parameters are not necessarily available, this Kp prediction tool will provide Kp prediction when only limited input parameters are available. The presented innovative method will improve tissue distribution prediction accuracy, thus enhancing the confidence in PBPK modeling outputs.     

BindingDB and ChEMBL: online compound databases for drug discovery

Public domain repositories of compound structures and activity data are indispensable tools for many aspects of pharmaceutical research, especially in academic environments. Such databases provide essential resources for structure-activity data mining and the evaluation of chemoinformatics and drug design methods. They are also important to support scientific interactions between commercial and academic environments. This editorial highlights two major public domain compound data repositories, BindingDB and ChEMBL, which have different origins. BindingDB has evolved in an academic setting (and continues to be developed there) and ChEMBL in a biotechnology environment. The ChEMBL database is now maintained and further developed at the European Bioinformatics Institute Outstation of the European Molecular Biology Laboratory. These databases mostly contain structures and activity data taken from the scientific literature, covering different stages of compound exploration and optimization efforts, and provide a substantial body of complementary compound activity information. Together with PubChem bioassays, ChEMBL and BindingDB provide the foundation of compound data analysis in the public domain.     

BindingDB and ChEMBL: online compound databases for drug discovery

Public domain repositories of compound structures and activity data are indispensable tools for many aspects of pharmaceutical research, especially in academic environments. Such databases provide essential resources for structure-activity data mining and the evaluation of chemoinformatics and drug design methods. They are also important to support scientific interactions between commercial and academic environments. This editorial highlights two major public domain compound data repositories, BindingDB and ChEMBL, which have different origins. BindingDB has evolved in an academic setting (and continues to be developed there) and ChEMBL in a biotechnology environment. The ChEMBL database is now maintained and further developed at the European Bioinformatics Institute Outstation of the European Molecular Biology Laboratory. These databases mostly contain structures and activity data taken from the scientific literature, covering different stages of compound exploration and optimization efforts, and provide a substantial body of complementary compound activity information. Together with PubChem bioassays, ChEMBL and BindingDB provide the foundation of compound data analysis in the public domain.     

Combinatorial compound libraries for drug discovery: an ongoing challenge

Almost 20 years of combinatorial chemistry have emphasized the power of numbers, a key issue for drug discovery in the current genomic era, in which it has been estimated that there might be more than 10,000 potential targets for which it would be desirable to have small-molecule modulators. Combinatorial chemistry is best described as the industrialization of chemistry; the chemistry has not changed, just the way in which it is now carried out, which is principally by exploiting instrumentation and robotics coupled to the extensive use of computers to efficiently control the process and analyse the vast amounts of resulting data. Many researchers have contributed to the general concepts as well as to the technologies in present use. However, some interesting challenges still remain to be solved, and these are discussed here in the context of the application of combinatorial chemistry to drug discovery.     

Inclusion compound formulation of hirsutenone with beta-cyclodextrin

Inclusion complexation with β-cyclodextrin (β-CD) was performed to improve the stability of hirsutenone (HST), a naturally occurring immunomodulator that is labile in aqueous solutions. HST-β-CD inclusion complexes were prepared using a solvent evaporation method. Briefly, solutions of HST (dissolved in isopropyl alcohol) and β-CD (dissolved in distilled water) were mixed and evaporated under vacuum by using a rotary evaporator. Phase solubility studies of the product revealed 1:1 or 1:2 complex formation, with an apparent stability constant of 249.2 M^?1. Differential scanning calorimetry showed a shift of endothermic peaks and nuclear magnetic resonance spectra displayed shift changes in H-3,5,6 protons, located inside the β-CD cavity, in inclusion complexes. These data provided strong evidence for inclusion complex formation. Characterization using infrared spectroscopy was hindered because of interfering β-CD vibrations. Inclusion complex stability was evaluated in the solid and aqueous solution states. Rate constants (×10^?2, day^?1) of HST and HST-β-CD were 13.2 and 9.79, respectively, in an aqueous solution at 25 °C; the corresponding values in the solid state were 0.14 and 0.18. The present study therefore showed successful formation of HST-β-CD, but the stability of HST within this inclusion compound was not markedly improved.     

An improved topical formulation for urea

The solubility of urea in different polar solvents was studied by means of determination of their dielectric constants. The most appropriate solvents turned out to be water and a water/propylene glycol (1:1 in volume) mixture. The best solvents were then used in the study of different semisolid vehicles for topical use (cetylic excipient, Beeler's base and Carbopol gel), which show different physicochemical characteristics. The final formulation contained a 40% (w/w) concentration in urea, since this value is most often used in Dermatology and, at the same time, it is the most problematic from a technological point of view. The stability of the different preparations was investigated by conductimetric and rheological determinations. The results are discussed in terms of both the solubility and stability of the active principles and the organoleptic and rheological characteristics of the final preparations.