On the fusion of threshold classifiers for categorization and dimensionality reduction

Optimizing Crowdsourced Land Use and Land Cover Data Collection: A Two-Stage Approach

Citizen science has become an increasingly popular approach to scientific data collection, where classification tasks involving visual interpretation of images is one prominent area of application, e.g., to support the production of land cover and land-use maps. Achieving a minimum accuracy in these classification tasks at a minimum cost is the subject of this study. A Bayesian approach provides an intuitive and reasonably straightforward solution to achieve this objective. However, its application requires additional information, such as the relative frequency of the classes and the accuracy of each user. While the former is often available, the latter requires additional data collection. In this paper, we present a two-stage approach to gathering this additional information. We demonstrate its application using a hypothetical two-class example and then apply it to an actual crowdsourced dataset with five classes, which was taken from a previous Geo-Wiki crowdsourcing campaign on identifying the size of agricultural fields from very high-resolution satellite imagery. We also attach the R code for the implementation of the newly presented approach.

Machine learning suggests sleep as a core factor in chronic pain

Patients with chronic pain have complex pain profiles and associated problems. Subgroup analysis can help identify key problems. We used a data-based approach to define pain phenotypes and their most relevant associated problems in 320 patients undergoing tertiary pain management. Unsupervised machine learning analysis of parameters "pain intensity," "number of pain areas," "pain duration," "activity pain interference," and "affective pain interference," implemented as emergent self-organizing maps, identified 3 patient phenotype clusters. Supervised analyses, implemented as different types of decision rules, identified "affective pain interference" and the "number of pain areas" as most relevant for cluster assignment. These appeared 698 and 637 times, respectively, in 1000 cross-validation runs among the most relevant characteristics in an item categorization approach in a computed ABC analysis. Cluster assignment was achieved with a median balanced accuracy of 79.9%, a sensitivity of 74.1%, and a specificity of 87.7%. In addition, among 59 demographic, pain etiology, comorbidity, lifestyle, psychological, and treatment-related variables, sleep problems appeared 638 and 439 times among the most important characteristics in 1000 cross-validation runs where patients were assigned to the 2 extreme pain phenotype clusters. Also important were the parameters "fear of pain," "self-rated poor health," and "systolic blood pressure." Decision trees trained with this information assigned patients to the extreme pain phenotype with an accuracy of 67%. Machine learning suggested sleep problems as key factors in the most difficult pain presentations, therefore deserving priority in the treatment of chronic pain.

Detecting Ordinal Subcascades

Ordinal classifier cascades are constrained by a hypothesised order of the semantic class labels of a dataset. This order determines the overall structure of the decision regions in feature space. Assuming the correct order on these class labels will allow a high generalisation performance, while an incorrect one will lead to diminished results. In this way ordinal classifier systems can facilitate explorative data analysis allowing to screen for potential candidate orders of the class labels. Previously, we have shown that screening is possible for total orders of all class labels. However, as datasets might comprise samples of ordinal as well as non-ordinal classes, the assumption of a total ordering might be not appropriate. An analysis of subsets of classes is required to detect such hidden ordinal substructures. In this work, we devise a novel screening procedure for exhaustive evaluations of all order permutations of all subsets of classes by bounding the number of enumerations we have to examine. Experiments with multi-class data from diverse applications revealed ordinal substructures that generate new and support known relations.

Analysis, identification and visualization of subgroups in genomics

Motivation

Cancer is a complex and heterogeneous disease involving multiple somatic mutations that accumulate during its progression. In the past years, the wide availability of genomic data from patients’ samples opened new perspectives in the analysis of gene mutations and alterations. Hence, visualizing and further identifying genes mutated in massive sets of patients are nowadays a critical task that sheds light on more personalized intervention approaches.

Results

Here, we extensively review existing tools for visualization and analysis of alteration data. We compare different approaches to study mutual exclusivity and sample coverage in large-scale omics data. We complement our review with the standalone software AVAtar (‘analysis and visualization of alteration data’) that integrates diverse aspects known from different tools into a comprehensive platform. AVAtar supplements customizable alteration plots by a multi-objective evolutionary algorithm for subset identification and provides an innovative and user-friendly interface for the evaluation of concurrent solutions. A use case from personalized medicine demonstrates its unique features showing an application on vaccination target selection.

Availability

AVAtar is available at: https://github.com/sysbio-bioinf/avatar

Contact

hans.kestler@uni-ulm.de, phone: +49 (0) 731 500 24 500, fax: +49 (0) 731 500 24 502

Constraining classifiers in molecular analysis: invariance and robustness

Analysing molecular profiles requires the selection of classification models that can cope with the high dimensionality and variability of these data. Also, improper reference point choice and scaling pose additional challenges. Often model selection is somewhat guided by ad hoc simulations rather than by sophisticated considerations on the properties of a categorization model. Here, we derive and report four linked linear concept classes/models with distinct invariance properties for high-dimensional molecular classification. We can further show that these concept classes also form a half-order of complexity classes in terms of Vapnik-Chervonenkis dimensions, which also implies increased generalization abilities. We implemented support vector machines with these properties. Surprisingly, we were able to attain comparable or even superior generalization abilities to the standard linear one on the 27 investigated RNA-Seq and microarray datasets. Our results indicate that a priori chosen invariant models can replace ad hoc robustness analysis by interpretable and theoretically guaranteed properties in molecular categorization.

Using demographics toward efficient data classification in citizen science: a Bayesian approach

Public participation in scientific activities, often called citizen science, offers a possibility to collect and analyze an unprecedentedly large amount of data. However, diversity of volunteers poses a challenge to obtain accurate information when these data are aggregated. To overcome this problem, we propose a classification algorithm using Bayesian inference that harnesses diversity of volunteers to improve data accuracy. In the algorithm, each volunteer is grouped into a distinct class based on a survey regarding either their level of education or motivation to citizen science. We obtained the behavior of each class through a training set, which was then used as a prior information to estimate performance of new volunteers. By applying this approach to an existing citizen science dataset to classify images into categories, we demonstrate improvement in data accuracy, compared to the traditional majority voting. Our algorithm offers a simple, yet powerful, way to improve data accuracy under limited effort of volunteers by predicting the behavior of a class of individuals, rather than attempting at a granular description of each of them.

Assessing phenotype order in molecular data

Biological entities are key elements of biomedical research. Their definition and their relationships are important in areas such as phylogenetic reconstruction, developmental processes or tumor evolution. Hypotheses about relationships like phenotype order are often postulated based on prior knowledge or belief. Evidence on a molecular level is typically unknown and whether total orders are reflected in the molecular measurements is unclear or not assessed. In this work we propose a method that allows a fast and exhaustive screening for total orders in large datasets. We utilise ordinal classifier cascades to identify discriminable molecular representations of the phenotypes. These classifiers are constrained by an order hypothesis and are highly sensitive to incorrect assumptions. Two new error bounds, which are introduced and theoretically proven, lead to a substantial speed-up and allow the application to large collections of many phenotypes. In our experiments we show that by exhaustively evaluating all possible candidate orders, we are able to identify phenotype orders that best coincide with the high-dimensional molecular profiles.

Producing knowledge by admitting ignorance: Enhancing data quality through an "I don't know" option in citizen science

The "noisy labeler problem" in crowdsourced data has attracted great attention in recent years, with important ramifications in citizen science, where non-experts must produce high-quality data. Particularly relevant to citizen science is dynamic task allocation, in which the level of agreement among labelers can be progressively updated through the information-theoretic notion of entropy. Under dynamic task allocation, we hypothesized that providing volunteers with an "I don't know" option would contribute to enhancing data quality, by introducing further, useful information about the level of agreement among volunteers. We investigated the influence of an "I don't know" option on the data quality in a citizen science project that entailed classifying the image of a highly polluted canal into "threat" or "no threat" to the environment. Our results show that an "I don't know" option can enhance accuracy, compared to the case without the option; such an improvement mostly affects the true negative rather than the true positive rate. In an information-theoretic sense, these seemingly meaningless blank votes constitute a meaningful piece of information to help enhance accuracy of data in citizen science.

Combined microRNA and mRNA microfluidic TaqMan array cards for the diagnosis of malignancy of multiple types of pancreatico-biliary tumors in fine-needle aspiration material

Pancreatic ductal adenocarcinoma (PDAC) continues to carry the lowest survival rates among all solid tumors. A marked resistance against available therapies, late clinical presentation and insufficient means for early diagnosis contribute to the dismal prognosis. Novel biomarkers are thus required to aid treatment decisions and improve patient outcomes.

We describe here a multi-omics molecular platform that allows for the first time to simultaneously analyze miRNA and mRNA expression patterns from minimal amounts of biopsy material on a single microfluidic TaqMan Array card. Expression profiles were generated from 113 prospectively collected fine needle aspiration biopsies (FNAB) from patients undergoing surgery for suspect masses in the pancreas. Molecular classifiers were constructed using support vector machines, and rigorously evaluated for diagnostic performance using 10×10fold cross validation. The final combined miRNA/mRNA classifier demonstrated a sensitivity of 91.7%, a specificity of 94.5%, and an overall diagnostic accuracy of 93.0% for the differentiation between PDAC and benign pancreatic masses, clearly outperfoming miRNA-only classifiers. The classification algorithm also performed very well in the diagnosis of other types of solid tumors (acinar cell carcinomas, ampullary cancer and distal bile duct carcinomas), but was less suited for the diagnostic analysis of cystic lesions.

We thus demonstrate that simultaneous analysis of miRNA and mRNA biomarkers from FNAB samples using multi-omics TaqMan Array cards is suitable to differentiate suspect solid pancreatic masses with high precision.

Genetic Factors of the Disease Course After Sepsis: Rare Deleterious Variants Are Predictive

Sepsis is a life-threatening organ dysfunction caused by dysregulated host response to infection. For its clinical course, host genetic factors are important and rare genomic variants are suspected to contribute. We sequenced the exomes of 59 Greek and 15 German patients with bacterial sepsis divided into two groups with extremely different disease courses. Variant analysis was focusing on rare deleterious single nucleotide variants (SNVs).

We identified significant differences in the number of rare deleterious SNVs per patient between the ethnic groups. Classification experiments based on the data of the Greek patients allowed discrimination between the disease courses with estimated sensitivity and specificity > 75%. By application of the trained model to the German patients we observed comparable discriminatory properties despite lower population-specific rare SNV load. Furthermore, rare SNVs in genes of cell signaling and innate immunity related pathways were identified as classifiers discriminating between the sepsis courses.

Sepsis patients with favorable disease course after sepsis, even in the case of unfavorable preconditions, seem to be affected more often by rare deleterious SNVs in cell signaling and innate immunity related pathways, suggesting a protective role of impairments in these processes against a poor disease course.

•

Rare SNV load is higher in the Greek vs. German population.

•

Subsets of rare deleterious SNVs are predictive for the disease course after sepsis.

•

Patients with favorable disease course seem to carry protective deleterious variants in sepsis related pathways.

Sepsis is a life-threatening disease caused by improper response to infection. Only little is known about the role of genetic factors. From > 4000 patients we selected the most extreme cases showing either a favorable or adverse disease course. We determined rare (< 1/200) protein-damaging genetic variants, as they may have a large effect. Using a computational model that includes knowledge on genes we can predict the disease course with > 75% accuracy. Surprisingly, favorable courses can be expected if defense mechanisms are damaged and prevented from overshooting. This underlines the relevance of rare variants for better understanding of sepsis and may offer new treatment options.

BiTrinA--multiscale binarization and trinarization with quality analysis

MOTIVATION

When processing gene expression profiles or other biological data, it is often required to assign measurements to distinct categories (e.g. 'high' and 'low' and possibly 'intermediate'). Subsequent analyses strongly depend on the results of this quantization. Poor quantization will have potentially misleading effects on further investigations. We propose the BiTrinA package that integrates different multiscale algorithms for binarization and for trinarization of one-dimensional data with methods for quality assessment and visualization of the results. By identifying measurements that show large variations over different time points or conditions, this quality assessment can determine candidates that are related to the specific experimental setting.

AVAILABILITY AND IMPLEMENTATION

BiTrinA is freely available on CRAN.

CONTACT

hans.kestler@leibniz-fli.de or hans.kestler@uni-ulm.de

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.