A ranking method for the concurrent learning of compounds with various activity profiles

Background

In this study, we present a SVM-based ranking algorithm for the concurrent learning of compounds with different activity profiles and their varying prioritization. To this end, a specific labeling of each compound was elaborated in order to infer virtual screening models against multiple targets. We compared the method with several state-of-the-art SVM classification techniques that are capable of inferring multi-target screening models on three chemical data sets (cytochrome P450s, dehydrogenases, and a trypsin-like protease data set) containing three different biological targets each.

Results

The experiments show that ranking-based algorithms show an increased performance for single- and multi-target virtual screening. Moreover, compounds that do not completely fulfill the desired activity profile are still ranked higher than decoys or compounds with an entirely undesired profile, compared to other multi-target SVM methods.

Conclusions

SVM-based ranking methods constitute a valuable approach for virtual screening in multi-target drug design. The utilization of such methods is most helpful when dealing with compounds with various activity profiles and the finding of many ligands with an already perfectly matching activity profile is not to be expected.

Electronic supplementary material

The online version of this article (doi:10.1186/s13321-014-0050-6) contains supplementary material, which is available to authorized users.

Brain connectivity and novel network measures for Alzheimer's disease classification

We compare a variety of different anatomic connectivity measures, including several novel ones, that may help in distinguishing Alzheimer's disease (AD) patients from controls. We studied diffusion-weighted magnetic resonance imaging from 200 subjects scanned as part of the Alzheimer's Disease Neuroimaging Initiative. We first evaluated measures derived from connectivity matrices based on whole-brain tractography; next, we studied additional network measures based on a novel flow-based measure of brain connectivity, computed on a dense 3-dimensional lattice. Based on these 2 kinds of connectivity matrices, we computed a variety of network measures. We evaluated the measures' ability to discriminate disease with a repeated, stratified 10-fold cross-validated classifier, using support vector machines, a supervised learning algorithm. We tested the relative importance of different combinations of features based on the accuracy, sensitivity, specificity, and feature ranking of the classification of 200 people into normal healthy controls and people with early or late mild cognitive impairment or AD.

Ranking clinical indicators for stroke pattern identification according to clinical importance in traditional Korean medicine

Differences in the value that individual clinicians place on each clinical indicator, when performing pattern identification, can be a significant obstacle in achieving accurate and consistent pattern identification. In this study, we aimed to identify and rank the importance of clinical indicators that correspond to each pattern. During the fundamental study to standardize and objectify the identification of patterns in traditional Korean medicine (TKM) treatment for stroke (the SOPI-Stroke project), we developed the Third Standardization for Pattern Identification (K-SPI-Stroke-III), which includes 44 clinical indicators that belong to one of the four patterns (Dampness-Phlegm, Fire-Heat, Qi deficiency, and Yin deficiency). Between September 2006 and December 2010, a total of 105 TKM doctors from 11 hospitals performed four examinations on 4,012 subjects based on the clinical indicators in the K-SPI-Stroke-III, and they were asked to leave a special mark on the indicators they thought were important, regardless of the number of the indicators. A total of 1,018 patients were included in the study. Among the 44 clinical indicators of the K-SPI-Stroke-III, four indicators were highly significant in differentiating between patterns and the frequency within a pattern. In contrast, another seven indicators were almost useless. Meanwhile, six confounding indicators had outlying values for differentiating between patterns and the frequency within a pattern. In conclusion, the four significant indicators should be actively considered when identifying patterns for strokes in clinical practice. Regarding the six confounding indicators, a further study should be conducted to modify some parts of the current K-SPI-Stroke-III.

Prestige rankings of chronic diseases and disabilities. A survey among professionals in the disability field

Disabled people constitute the world's largest minority; too little is known about the internal structure and valuations of that minority. We investigate whether prestige rankings of different chronic diseases and disabilities can be elicited from a community of disability non-governmental organizations (NGOs). A survey was performed in a sample of NGO professionals in Norway in late 2013 and early 2014. Two copies of a questionnaire was sent to 92 national and regional offices of disability NGOs requesting a response from either senior employees or elected officials, preferably one of each. Outcome measures were ratings on a 1-9 scale of the prestige these respondents believed most professionals in their field would accord to a sample of 38 different conditions. We find that there is a prestige hierarchy of chronic diseases and disabilities in the disability field. In this hierarchy, somatic conditions that are strongly associated with medical treatment were placed higher than either conditions that are characterized by permanence, or conditions are associated with psychosomatic etiologies. The elicited prestige hierarchy is at odds with prevalent normative positions in the disability field; there is a lack of fit between some of the field's central political goals and its internal evaluations. We propose that its structure can be explained through a) influence from the medical field, b) organization history, size, and prominence, and c) issues of credibility, shame and blame. Further research should be conducted into the structure and valuations of the disability field in general and people with disabilities and chronic diseases in particular.

Using risk-ranking of metals to identify which poses the greatest threat to freshwater organisms in the UK

Freshwater aquatic organisms face the challenge of being exposed to a multitude of chemicals discharged by the human population. The objective of this study was to rank metals according to the threat they pose to aquatic organisms. This will contribute to a wider Chemical Strategy for freshwater which will risk-rank all chemicals based on their potential risk to wildlife in a UK setting. The method involved comparing information on ecotoxicological thresholds with measured concentrations in rivers. The bioconcentration factor was also considered as a ranking method. The metals; Ag, Al, As, Cd, Cr, Cu, Fe, Hg, Mn, Ni, Pb and Zn, were analysed using this approach. Triclosan and lindane were used as comparative organic pollutants. Using a range of ranking techniques, Cu, Al and Zn came top of the list of concern, with Cu coming first.

Analysis of tennis champions' career: how did top-ranked players perform the previous years?

Professional tennis is a highly competitive sport ranked through an objective, merit-based, mathematical system. The objective of this study was to determine and analyze how top 1°-ranked professional tennis players perform in the previous three years. Data of ranking position of all top 1°-ranked female and male players in the professional era and their performance in Grand Slams tournaments were collected from tennis stakeholders’ websites and analyzed.

Top 1° male players’ ranking position averaged 2.17 ± 1.92 (CI 95%:1.56/2.78), 7.02 ± 18.073, and 10.73 ± 29.31, at 1, 2, and 3 previous years, respectively. Top 1° female players’ ranking position averaged 2.21 ± 1.59 (CI 95%:1.61/2.71), 4.78 ± 4.09, and 14.97 ± 26.75, at 1, 2, and 3 previous years, respectively. All top 1° male and female players were ranked within the 1°-10° and 1°-6° positions the previous year, respectively; the majority of tennis champions won at least one Grand Slam tournament during the year before reaching the top 1° ranking position (females = 69%; males = 65%), and during the same season (females = 82%; males = 92%). Female and male top 1°-ranked player maintained that position in the following year in 48.7% and 52.5% of cases, respectively.

In conclusion, tennis players need to perform at high level during at least three years prior to reach the top 1° position in the professional tennis ranking. Both, male and female champions, showed similar patterns of performance in their professional career.

CombiSimilarity, an innovative method to compare environmental and health data sets with different attribute sizes example: eighteen Organochlorine Pesticides in soil and human breast milk samples

Human health and the health of the environment have entwined. In this paper we underpin this position by presenting a modeling approach named CombiSimilarity, which has been developed by the first author in the software tool PyHasse comprising a wide variety of partial ordering tools. A case study of 18 Organochlorine Pesticides (OCPs) detected in soil as well as in human breast milk samples in the Taurus Mountains in Turkey is carried out. Seven soil samples and 44 breast milk samples were measured. We seek to answer the question whether the contamination pattern in breast milk is associated with the contamination pattern in soil by studying the mutual quantitative relationships of the chemicals involved. We could demonstrate that there is a similarity with respect to the concentration profiles between the soil and breast milk pollution. Therefore the hypothesis may be formulated that the concentrations of chemicals in the milk samples are strongly related to the soil contamination. This supports the concept that soil could be a surrogate for human exposure at background locations.

Evaluation of single-cell gel electrophoresis data: combination of variance analysis with sum of ranking differences

Specimens of the mussel Mytilus galloprovincialis were collected from five sites in the Boka Kotorska Bay (Adriatic Sea, Montenegro) during the period summer 2011-autumn 2012. Three types of tissue, haemolymph, digestive gland were used for assessment of DNA damage. Images of randomly selected cells were analyzed with a fluorescence microscope and image analysis by the Comet Assay IV Image-analysis system. Three parameters, viz. tail length, tail intensity and Olive tail moment were analyzed on 4200 nuclei per cell type. We observed variations in the level of DNA damage in mussels collected at different sites, as well as seasonal variations in response. Sum of ranking differences (SRD) was implemented to compare use of different types of cell and different measure of comet tail per nucleus. Numerical scales were transferred into ranks, range scaling between 0 and 1; standardization and normalization were carried out. SRD selected the best (and worst) combinations: tail moment is the best for all data treatment and for all organs; second best is tail length, and intensity ranks third (except for digestive gland). The differences were significant at the 5% level. Whereas gills and haemolymph cells do not differ significantly, cells of the digestive gland are much more suitable to estimate genotoxicity. Variance analysis decomposed the effect of different factors on the SRD values. This unique combination has provided not only the relative importance of factors, but also an overall evaluation: the best evaluation method, the best data pre-treatment, etc., were chosen even for partially contradictory data. The rank transformation is superior to any other way of scaling, which is proven by ordering the SRD values by SRD again, and by cross validation.

Quantitative assessments of municipal waste management systems: using different indicators to compare and rank programs in New York State

The primary objective of waste management technologies and policies in the United States is to reduce the harmful environmental impacts of waste, particularly those relating to energy consumption and climate change. Performance indicators are frequently used to evaluate the environmental quality of municipal waste systems, as well as to compare and rank programs relative to each other in terms of environmental performance. However, there currently is no consensus on the best indicator for performing these environmental evaluations. The purpose of this study is to examine the common performance indicators used to assess the environmental benefits of municipal waste systems to determine if there is agreement between them regarding which system performs best environmentally. Focus is placed on how indicator selection influences comparisons between municipal waste management programs and subsequent system rankings. The waste systems of ten municipalities in the state of New York, USA, were evaluated using each common performance indicator and Spearman correlations were calculated to see if there was a significant association between system rank orderings. Analyses showed that rank orders of waste systems differ substantially when different indicators are used. Therefore, comparative system assessments based on indicators should be considered carefully, especially those intended to gauge environmental quality. Insight was also gained into specific factors which may lead to one system achieving higher rankings than another. However, despite the insufficiencies of indicators for comparative quality assessments, they do provide important information for waste managers and they can assist in evaluating internal programmatic performance and progress. To enhance these types of assessments, a framework for scoring indicators based on criteria that evaluate their utility and value for system evaluations was developed. This framework was used to construct an improved model for waste system performance assessments.

A FTA-based method for risk decision-making in emergency response

Decision-making problems in emergency response are usually risky and uncertain due to the limited decision data and possible evolvement of emergency scenarios. This paper focuses on a risk decision-making problem in emergency response with several distinct characteristics including dynamic evolvement process of emergency, multiple scenarios, and impact of response actions on the emergency scenarios. A method based on Fault Tree Analysis (FTA) is proposed to solve the problem. By analyzing the evolvement process of emergency, the Fault Tree (FT) is constructed to describe the logical relations among conditions and factors resulting in the evolvement of emergency. Given different feasible response actions, the probabilities of emergency scenarios are estimated by FTA. Furthermore, the overall ranking value of each action is calculated, and a ranking of feasible response actions is determined. Finally, a case study on H1N1 infectious diseases is given to illustrate the feasibility and validity of the proposed method.

Selecting pesticides for inclusion in drinking water quality guidelines on the basis of detection probability and ranking

Pesticides released into the environment may pose both ecological and human health risks. Governments set the regulations and guidelines for the allowable levels of active components of pesticides in various exposure sources, including drinking water. Several pesticide risk indicators have been developed using various methodologies, but such indicators are seldom used for the selection of pesticides to be included in national regulations and guidelines. The aim of the current study was to use risk indicators for the selection of pesticides to be included in regulations and guidelines. Twenty-four risk indicators were created, and a detection rate was defined to judge which indicators were the best for selection. The combination of two indicators (local sales of a pesticide for the purposes of either rice farming or other farming, divided by the guideline value and annual precipitation, and amended with the scores from the physical and chemical properties of the pesticide) gave the highest detection rates. In this case study, this procedure was used to evaluate 134 pesticides that are currently unregulated in the Japanese Drinking Water Quality Guidelines, from which 44 were selected as pesticides to be added to the primary group in the guidelines. The detection probability of the 44 pesticides was more than 72%. Among the 102 pesticides currently in the primary group, 17 were selected for withdrawal from the group.

A FTA-based method for risk decision-making in emergency response

Decision-making problems in emergency response are usually risky and uncertain due to the limited decision data and possible evolvement of emergency scenarios. This paper focuses on a risk decision-making problem in emergency response with several distinct characteristics including dynamic evolvement process of emergency, multiple scenarios, and impact of response actions on the emergency scenarios. A method based on Fault Tree Analysis (FTA) is proposed to solve the problem. By analyzing the evolvement process of emergency, the Fault Tree (FT) is constructed to describe the logical relations among conditions and factors resulting in the evolvement of emergency. Given different feasible response actions, the probabilities of emergency scenarios are estimated by FTA. Furthermore, the overall ranking value of each action is calculated, and a ranking of feasible response actions is determined. Finally, a case study on H1N1 infectious diseases is given to illustrate the feasibility and validity of the proposed method.

Scalable and Axiomatic Ranking of Network Role Similarity

A key task in analyzing social networks and other complex networks is role analysis: describing and categorizing nodes according to how they interact with other nodes. Two nodes have the same role if they interact with equivalent sets of neighbors. The most fundamental role equivalence is automorphic equivalence. Unfortunately, the fastest algorithms known for graph automorphism are nonpolynomial. Moreover, since exact equivalence is rare, a more meaningful task is measuring the role similarity between any two nodes. This task is closely related to the structural or link-based similarity problem that SimRank addresses. However, SimRank and other existing similarity measures are not sufficient because they do not guarantee to recognize automorphically or structurally equivalent nodes. This paper makes two contributions. First, we present and justify several axiomatic properties necessary for a role similarity measure or metric. Second, we present RoleSim, a new similarity metric which satisfies these axioms and which can be computed with a simple iterative algorithm. We rigorously prove that RoleSim satisfies all these axiomatic properties. We also introduce Iceberg RoleSim, a scalable algorithm which discovers all pairs with RoleSim scores above a user-defined threshold θ. We demonstrate the interpretative power of RoleSim on both both synthetic and real datasets.

Ranking of concern, based on environmental indexes, for pharmaceutical and personal care products: an application to the Spanish case

A wide range of Pharmaceuticals and Personal Care Products (PPCPs) are present in the environment, and many of their adverse effects are unknown. The emergence of new compounds or changes in regulations have led to dynamical studies of occurrence, impact and treatment, which consider geographical areas and trends in consumption and innovation in the pharmaceutical industry. A Quantitative study of Structure-Activity Relationship ((Q)SAR) was performed to assess the possible adverse effects of ninety six PPCPs and metabolites with negligible experimental data and establish a ranking of concern, which was supported by the EPA EPI Suite™ interface. The environmental and toxicological indexes, the persistence (P), the bioaccumulation (B), the toxicity (T) (extensive) and the occurrence in Spanish aquatic environments (O) (intensive) were evaluated. The most hazardous characteristics in the largest number of compounds were generated by the P index, followed by the T and B indexes. A high number of metabolites has a concern score equal to or greater than their parent compounds. Three PBT and OPBT rankings of concern were proposed using the total and partial ranking method (supported by a Hasse diagram) by the Decision Analysis by Ranking Techniques (DART) tool, which was recently recommended by the European Commission. An analysis of the sensibility of the relative weights of these indexes has been conducted. Hormones, antidepressants (and their metabolites), blood lipid regulators and all of the personal care products considered in this study were at the highest levels of risk according to the PBT and OPBT total rankings. Furthermore, when the OPBT partial ranking was performed, X-ray contrast media, H2 blockers and some antibiotics were included at the highest level of concern. It is important to improve and incorporate useful indexes for the predicted environmental impact of PPCPs and metabolites and thus focus experimental analysis on the compounds that require urgent attention.