Rapid analysis of multiclass antibiotic residues and some of their metabolites in hospital, urban wastewater and river water by ultra-highperformance liquid chromatography coupled to quadrupole-linear ion trap tandem mass spectrometry

The present work describes the development of a fast and robust analytical method for the determination of 53 antibiotic residues, covering various chemical groups and some of their metabolites, in environmental matrices that are considered important sources of antibiotic pollution, namely hospital and urban wastewaters, as well as in river waters. The method is based on automated off-line solid phase extraction (SPE) followed by ultra-high-performance liquid chromatography coupled to quadrupole linear ion trap tandem mass spectrometry (UHPLC–QqLIT). For unequivocal identification and confirmation, and in order to fulfill EU guidelines, two selected reaction monitoring (SRM) transitions per compound are monitored (the most intense one is used for quantification and the second one for confirmation). Quantification of target antibiotics is performed by the internal standard approach, using one isotopically labeled compound for each chemical group, in order to correct matrix effects. The main advantages of the method are automation and speed-up of sample preparation, by the reduction of extraction volumes for all matrices, the fast separation of a wide spectrum of antibiotics by using ultra-high-performance liquid chromatography, its sensitivity (limits of detection in the low ng/L range) and selectivity (due to the use of tandem mass spectrometry) The inclusion of β-lactam antibiotics (penicillins and cephalosporins), which are compounds difficult to analyze in multi-residue methods due to their instability in water matrices, and some antibiotics metabolites are other important benefits of the method developed. As part of the validation procedure, the method developed was applied to the analysis of antibiotics residues in hospital, urban influent and effluent wastewaters as well as in river water samples

Classifier combination for in vivo magnetic resonance spectra of brain tumours

Peer-reviewed

Classifier combination for in vivo magnetic resonance spectra of brain tumours

In this paper we present a multi-stage classifier for magnetic resonance spectra of human brain tumours which is being developed as part of a decision support system for radiologists. The basic idea is to decompose a complex classification scheme into a sequence of classifiers, each specialising in different classes of tumours and trying to reproducepart of the WHO classification hierarchy. Each stage uses a particular set of classification features, which are selected using a combination of classical statistical analysis, splitting performance and previous knowledge.Classifiers with different behaviour are combined using a simple voting scheme in order to extract different error patterns: LDA, decision trees and the k-NN classifier. A special label named "unknown¿ is used when the outcomes of the different classifiers disagree. Cascading is alsoused to incorporate class distances computed using LDA into decision trees. Both cascading and voting are effective tools to improve classification accuracy. Experiments also show that it is possible to extract useful information from the classification process itself in order to helpusers (clinicians and radiologists) to make more accurate predictions and reduce the number of possible classification mistakes.

Shared feature extraction for nearest neighbor face recognition

In this paper, we propose a new supervised linearfeature extraction technique for multiclass classification problemsthat is specially suited to the nearest neighbor classifier (NN).The problem of finding the optimal linear projection matrix isdefined as a classification problem and the Adaboost algorithmis used to compute it in an iterative way. This strategy allowsthe introduction of a multitask learning (MTL) criterion in themethod and results in a solution that makes no assumptions aboutthe data distribution and that is specially appropriated to solvethe small sample size problem. The performance of the methodis illustrated by an application to the face recognition problem.The experiments show that the representation obtained followingthe multitask approach improves the classic feature extractionalgorithms when using the NN classifier, especially when we havea few examples from each class

A Sparse Bayesian Approach for Joint Feature Selection and Classifier Learning

Peer-reviewed

Readout electronics for low dark count pixel detectors based on geiger mode avalanche photodiodes fabricated in conventional CMOS technologies for future linear colliders

The high sensitivity and excellent timing accuracy of Geiger mode avalanche photodiodes makes them ideal sensors as pixel detectors for particle tracking in high energy physics experiments to be performed in future linear colliders. Nevertheless, it is well known that these sensors suffer from dark counts and afterpulsing noise, which induce false hits (indistinguishable from event detection) as well as an increase of the necessary area of the readout system. In this work, we present a comparison between APDs fabricated in a high voltage 0.35 µm and a high integration 0.13 µm commercially available CMOS technologies that has been performed to determine which of them best fits the particle collider requirements. In addition, a readout circuit that allows low noise operation is introduced. Experimental characterization of the proposed pixel is also presented in this work.