Recently, much attention has been given to models for identifying rumors in social media. Features that are helpful for automatic inference of credibility, veracity, reliability of information have been described. The ultimate goal is to train classification models that are able to recognize future high-impact rumors as early as possible, before the event unfolds. The generalization power of the models is greatly hindered by the domain-dependent distributions of the features, an issue insufficiently discussed. Here we study a large dataset consisting of rumor and non-rumor tweets commenting on nine breakingnews stories taking place in different locations of the world. We found that the distribution of most features are specific to the event and that this bias naturally affects the performance of the model. The analysis of the domain-specific feature distributions is insightful and hints to the distinct characteristics of the underlying social network for different countries, social groups, cultures and others.

Our goal is to build a Food and Drink (FD) gazetteer that can serve for classification of general, FD-related concepts, efficient faceted search or automated semantic enrichment. Fully supervised design of a domain-specific models ex novo is not scalable. Integration of several ready knowledge bases is tedious and does not ensure coverage. Completely data-driven approaches require a large amount of training data, which is not always available. For general domains (such as the FD domain), re-using encyclopedic knowledge bases like Wikipedia may be a good idea. We propose here a semi-supervised approach that uses a restricted Wikipedia as a base for the modeling, achieved by selecting a domain-relevant Wikipedia category as root for the model and all its subcategories, combined with expert and data-driven pruning of irrelevant categories.

The classification models obtained via maximum likelihood-based training do not necessarily reach the optimal \(F_\beta \)-measure for some user’s choice of \(\beta \) that is achievable with the chosen parametrization. In this work we link the weighted maximum entropy and the optimization of the expected \(F_\beta \)-measure, by viewing them in the framework of a general common multi-criteria optimization problem. As a result, each solution of the expected \(F_\beta \)-measure maximization can be realized as a weighted maximum likelihood solution within the maximum entropy model - a well understood and behaved problem for which standard (off the shelf) gradient methods can be used. Based on this insight, we present an efficient algorithm for optimization of the expected \(F_\beta \) using weighted maximum likelihood with dynamically adaptive weights.

The performance of NLP classifiers largely depends on the quality of the features considered for prediction (feature engineering). However, as the number of features increases, the more likely overfitting becomes and performance decreases. Also, due to the very large number of features, only slimple linear classifiers are considered, thus disregarding potentially predictive non-linear combinations of features. Here we propose an automated method for feature induction, which selects and includes in the model features and feature combinations which are likely to be useful for the prediction.The resulting model relies on a smaller feature set, is non-linear and is more accurate than the baseline, which is the model trained on the entire feature set. The method uses a greedy filtering approach based on various univariate measures of feature relevance and it is very fast in practice. Also, our feature induction method is independent of the classifier used: we (a) ==

We link the weighted maximum entropy and the optimization of the expected F measure, by viewing them in the framework of a general common multi-criteria optimization problem. As a result, each solution of the expectedF -measure maximization can be realized as a weighted maximum likelihood solution - a well understood and behaved problem. The specific structure of maximum entropy models allows us to approximate this characterization via the much simpler class-wise weighted maximum likelihood. Our approach reveals any probabilistic learning scheme as a specific trade-off between different objectives and provides the framework to link it to the expectedF -measure.

D4.3A is an annex to the D4.3 deliverable of WP4 of the MOLTO project. It aims to address the reviewers’ remarks and recommendations for D4.3, as well as to present a final overview of the prototypes built in the scope of MOLTO with respect to grammar-ontology interoperabilty. D4.3A also describes the work after M24 and gives a general overview of the achievements in MOLTO with focus on WP4 - Knowledge Engineering, WP7 - Patents use case, and WP8 - Cultural Heritage use case.

In excess of 12% of human cancer incidents have a viral cofactor. Epidemiological studies of idiopathic human cancers indicate that additional tumor viruses remain to be discovered. Recent advances in sequencing technology have enabled systematic screenings of human tumor transcriptomes for viral transcripts. However, technical problems such as low abundances of viral transcripts in large volumes of sequencing data, viral sequence divergence, and homology between viral and human factors significantly confound identification of tumor viruses. We have developed a novel computational approach for detecting viral transcripts in human cancers that takes the aforementioned confounding factors into account and is applicable to a wide variety of viruses and tumors. We apply the approach to conducting the first systematic search for viruses in neuroblastoma, the most common cancer in infancy. The diverse clinical progression of this disease as well as related epidemiological and virological findings are highly suggestive of a pathogenic cofactor. However, a viral etiology of neuroblastoma is currently contested. We mapped 14 transcriptomes of neuroblastoma as well as positive and negative controls to the human and all known viral genomes in order to detect both known and unknown viruses. Analysis of controls, comparisons with related methods, and statistical estimates demonstrate the high sensitivity of our approach. Detailed investigation of putative viral transcripts within neuroblastoma samples did not provide evidence for the existence of any known human viruses. Likewise, de-novo assembly and analysis of chimeric transcripts did not result in expression signatures associated with novel human pathogens. While confounding factors such as sample dilution or viral clearance in progressed tumors may mask viral cofactors in the data, in principle, this is rendered less likely by the high sensitivity of our approach and the number of biological replicates analyzed. Therefore, our results suggest that frequent viral cofactors of metastatic neuroblastoma are unlikely.

MOTIVATIONRecurrent DNA breakpoints in cancer genomes indicate the presence of critical functional elements for tumor development. Identifying them can help determine new therapeutic targets. High-dimensional DNA microarray experiments like arrayCGH afford the identification of DNA copy number breakpoints with high precision, offering a solid basis for computational estimation of recurrent breakpoint locations.RESULTSWe introduce a method for identification of recurrent breakpoints (consensus breakpoints) from copy number aberration datasets. The method is based on weighted kernel counting of breakpoints around genomic locations. Counts larger than expected by chance are considered significant. We show that the consensus breakpoints facilitate consensus segmentation of the samples. We apply our method to three arrayCGH datasets and show that by using consensus segmentation we achieve significant dimension reduction, which is useful for the task of prediction of tumor phenotype based on copy number data. We use our approach for classification of neuroblastoma tumors from different age groups and confirm the recent recommendation for the choice of age cut-off for differential treatment of 18 months. We also investigate the (epi)genetic properties at consensus breakpoint locations for seven datasets and show enrichment in overlap with important functional genomic regions.AVAILABILITYImplementation in R of our approach can be found at http://www.mpi-inf.mpg.de/ ∼laura/FeatureGrouping.html.CONTACTlaura@mpi-inf.mpg.de.SUPPLEMENTARY INFORMATIONSupplementary data are available at Bioinformatics online.

In the CLEF 2012 the BulTreeBank Group of LMD, IICT, BAS is participating for QA4MRE task for Bulgarian. The system represented in the paper exploits an NLP Pipeline for Bulgarian in order to process the questions, answers and the supporting texts. Then we represent the results of the analysis as a bag of linguistic units lemmas, dependency relations. These bags of words are the match between the question plus answer and the sentences in the text. The answer that maximizes the overlap is selected as the correct one. Since the system is deterministic we have only one run. The score achieved by the run is 0.29. The other two runs are performed as baseline runs with randomly selected answers. Their scores are 0.20 and 0.12, respectively. Thus, the using of linguistic units in the overlapping estimation provides significant improvements over the baseline.

We describe three language-independent methods for the task of answer validation. All methods are based on a scoring mechanism that reflects the degree of similarity between the question-answer pairs and the supporting text. We evaluate the proposed methods when using various string similarity metrics, such as exact matching, Levenshtein, Jaro and Jaro-Winkler. In addition to this baseline approach, we take advantage of the multilingual QA4MRE dataset, and devise an ensemble method, which chooses the answer indicated as correct by the largest number of analyses of the individual translations. Finally, we present a language-augmented method that enriches the questions and answers with paraphrases obtained by means of machine translation. Our methods depend on parameters which we estimate using the dataset from CLEF2011. We show that all of the described approaches achieve a significant improvement over the random baseline, and that both majority voting and language augmentation lead to superior accuracy as compared with the original method.

MOTIVATION:Classification and feature selection of genomics or transcriptomics data is often hampered by the large number of features as compared with the small number of samples available. Moreover, features represented by probes that either have similar molecular functions (gene expression analysis) or genomic locations (DNA copy number analysis) are highly correlated. Classical model selection methods such as penalized logistic regression or random forest become unstable in the presence of high feature correlations. Sophisticated penalties such as group Lasso or fused Lasso can force the models to assign similar weights to correlated features and thus improve model stability and interpretability. In this article, we show that the measures of feature relevance corresponding to the above-mentioned methods are biased such that the weights of the features belonging to groups of correlated features decrease as the sizes of the groups increase, which leads to incorrect model interpretation and misleading feature ranking.RESULTS:With simulation experiments, we demonstrate that Lasso logistic regression, fused support vector machine, group Lasso and random forest models suffer from correlation bias. Using simulations, we show that two related methods for group selection based on feature clustering can be used for correcting the correlation bias. These techniques also improve the stability and the accuracy of the baseline models. We apply all methods investigated to a breast cancer and a bladder cancer arrayCGH dataset and in order to identify copy number aberrations predictive of tumor phenotype.AVAILABILITY:R code can be found at: http://www.mpi-inf.mpg.de/~laura/Clustering.r.

MOTIVATION In life sciences, interpretability of machine learning models is as important as their prediction accuracy. Linear models are probably the most frequently used methods for assessing feature relevance, despite their relative inflexibility. However, in the past years effective estimators of feature relevance have been derived for highly complex or non-parametric models such as support vector machines and RandomForest (RF) models. Recently, it has been observed that RF models are biased in such a way that categorical variables with a large number of categories are preferred. RESULTS In this work, we introduce a heuristic for normalizing feature importance measures that can correct the feature importance bias. The method is based on repeated permutations of the outcome vector for estimating the distribution of measured importance for each variable in a non-informative setting. The P-value of the observed importance provides a corrected measure of feature importance. We apply our method to simulated data and demonstrate that (i) non-informative predictors do not receive significant P-values, (ii) informative variables can successfully be recovered among non-informative variables and (iii) P-values computed with permutation importance (PIMP) are very helpful for deciding the significance of variables, and therefore improve model interpretability. Furthermore, PIMP was used to correct RF-based importance measures for two real-world case studies. We propose an improved RF model that uses the significant variables with respect to the PIMP measure and show that its prediction accuracy is superior to that of other existing models. AVAILABILITY R code for the method presented in this article is available at http://www.mpi-inf.mpg.de/ approximately altmann/download/PIMP.R CONTACT: altmann@mpi-inf.mpg.de, laura.tolosi@mpi-inf.mpg.de SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Somatic genetic alterations in cancers have been linked with response to targeted therapeutics by creation of specific dependency on activated oncogenic signaling pathways. However, no tools currently exist to systematically connect such genetic lesions to therapeutic vulnerability. We have therefore developed a genomics approach to identify lesions associated with therapeutically relevant oncogene dependency. Using integrated genomic profiling, we have demonstrated that the genomes of a large panel of human non-small cell lung cancer (NSCLC) cell lines are highly representative of those of primary NSCLC tumors. Using cell-based compound screening coupled with diverse computational approaches to integrate orthogonal genomic and biochemical data sets, we identified molecular and genomic predictors of therapeutic response to clinically relevant compounds. Using this approach, we showed that v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations confer enhanced Hsp90 dependency and validated this finding in mice with KRAS-driven lung adenocarcinoma, as these mice exhibited dramatic tumor regression when treated with an Hsp90 inhibitor. In addition, we found that cells with copy number enhancement of v-abl Abelson murine leukemia viral oncogene homolog 2 (ABL2) and ephrin receptor kinase and v-src sarcoma (Schmidt-Ruppin A-2) viral oncogene homolog (avian) (SRC) kinase family genes were exquisitely sensitive to treatment with the SRC/ABL inhibitor dasatinib, both in vitro and when it xenografted into mice. Thus, genomically annotated cell-line collections may help translate cancer genomics information into clinical practice by defining critical pathway dependencies amenable to therapeutic inhibition.

BACKGROUND: The purpose of this study was to prove the feasibility of a longmer oligonucleotide microarray platform to profile gene copy number alterations in prostate cancer cell lines and to quickly indicate novel candidate genes, which may play a role in carcinogenesis. METHODS/RESULTS AND FINDINGS: Genome-wide screening for regions of genetic gains and losses on nine prostate cancer cell lines (PC3, DU145, LNCaP, CWR22, and derived sublines) was carried out using comparative genomic hybridization on a 35,000 feature oligonucleotide microarray (arrayCGH). Compared to conventional chromosomal CGH, more deletions and small regions of gains, particularly in pericentromeric regions and regions next to the telomeres, were detected. As validation of the high-resolution of arrayCGH we further analyzed a small amplicon of 1.7 MB at 9p13.3, which was found in CWR22 and CWR22-Rv1. Increased copy number was confirmed by fluorescence in situ hybridization using the BAC clone RP11-165H19 from the amplified region comprising the two genes interleukin 11 receptor alpha (IL11-RA) and dynactin 3 (DCTN3). Using quantitative real time PCR (qPCR) we could demonstrate that IL11-RA is the gene with the highest copy number gain in the cell lines compared to DCTN3 suggesting IL11-RA to be the amplification target. Screening of 20 primary prostate carcinomas by qPCR revealed an IL11-RA copy number gain in 75% of the tumors analyzed. Gain of DCTN3 was only found in two cases together with a gain of IL11-RA. CONCLUSIONS/SIGNIFICANCE: ArrayCGH using longmer oligonucleotide microarrays is feasible for high-resolution analysis of chomosomal imbalances. Characterization of a small gained region at 9p13.3 in prostate cancer cell lines and primary prostate cancer samples by fluorescence in situ hybridization and quantitative PCR has revealed interleukin 11 receptor alpha gene as a candidate target of amplification with an amplification frequency of 75% in prostate carcinomas. Frequent amplification of IL11-RA in prostate cancer is a potential mechanism of IL11-RA overexpression in this tumor type.

Analysis of ArrayCGH Data for the Estimation of Genetic Tumor Progression Laura Toloşi Master of Science Department of Computer Science Saarland University 2006 In cancer research, prediction of time to death or relapse is important for a meaningful tumor classification and selecting appropriate therapies. The accumulation of genetic alterations during tumor progression can be used for the assessment of the genetic status of the tumor. ArrayCGH technology is used to measure genomic amplifications and deletions, with a high resolution that allows the detection of down to single genes copy number changes. We propose an automated method for analysis of cancer mutations accumulation based on statistical analysis of arrayCGH data. The method consists of the four steps: arrayCGH smoothing, aberrations detection, consensus analysis and oncogenetic tree models estimation. For the second and third steps, we propose new algorithmic solutions. First, we use the adaptive weights smoothing-based algorithm GLAD for identifying regions of constant copy number. Then, in order to select regions of gain and loss, we fit robust normals to the smoothed Log2Ratios of each CGH array and choose appropriate significance cutoffs. The consensus analysis step consists of an automated selection of recurrent aberrant regions when multiple CGH experiments on the same tumor type are available. We propose to associate p-values to each measured genomic position and to select the regions where the p-value is sufficiently small. The aberrant regions computed by our method can be further used to estimate evolutionary trees, which model the dependencies between genetic mutations and can help to predict tumor progression stages and survival times. We applied our method to two arrayCGH data sets obtained from prostate cancer and glioblastoma patients, respectively. The results confirm previous knowledge on the genetic mutations specific to these types of cancer, but also bring out new regions, often reducing to single genes, due to the high resolution of arrayCGH measurements. An oncogenetic tree mixture model fitted to the Prostate Cancer data set shows two distinct evolutionary patterns discriminating between two different cell lines. Moreover, when used as clustering features, the genetic mutations our algorithm outputs separate well arrays representing 4 different cell lines, proving that we extract meaningful information.