In the last two decades, Computer Aided Detection (CAD) systems were developed to help radiologists analyse screening mammograms, however benefits of current CAD technologies appear to be contradictory, therefore they should be improved to be ultimately considered useful. Since 2012, deep convolutional neural networks (CNN) have been a tremendous success in image recognition, reaching human performance. These methods have greatly surpassed the traditional approaches, which are similar to currently used CAD solutions. Deep CNN-s have the potential to revolutionize medical image analysis. We propose a CAD system based on one of the most successful object detection frameworks, Faster R-CNN. The system detects and classifies malignant or benign lesions on a mammogram without any human intervention. The proposed method sets the state of the art classification performance on the public INbreast database, AUC = 0.95. The approach described here has achieved 2nd place in the Digital Mammography DREAM Challenge with AUC = 0.85. When used as a detector, the system reaches high sensitivity with very few false positive marks per image on the INbreast dataset. Source code, the trained model and an OsiriX plugin are published online at https://github.com/riblidezso/frcnn_cad.

A search technique locating network modules, i.e. internally densely connected groups of nodes in directed networks is introduced by extending the clique percolation method originally proposed for undirected networks. After giving a suitable definition for directed modules we investigate their percolation transition in the Erdős–Rényi graph both analytically and numerically. We also analyse four real-world directed networks, including Google's own web-pages, an email network, a word association graph and the transcriptional regulatory network of the yeast Saccharomyces cerevisiae. The obtained directed modules are validated by additional information available for the nodes. We find that directed modules of real-world graphs inherently overlap and the investigated networks can be classified into two major groups in terms of the overlaps between the modules. Accordingly, in the word-association network and Google's web-pages, overlaps are likely to contain in-hubs, whereas the modules in the email and transcriptional regulatory network tend to overlap via out-hubs.

Sewage metagenomics has risen to prominence in urban population surveillance of pathogens and antimicrobial resistance (AMR). Unknown species with similarity to known genomes cause database bias in reference-based metagenomics. To improve surveillance, we seek to recover sewage genomes and develop a quantification and correlation workflow for these genomes and AMR over time. We use longitudinal sewage sampling in seven treatment plants from five major European cities to explore the utility of catch-all sequencing of these population-level samples. Using metagenomic assembly methods, we recover 2332 metagenome-assembled genomes (MAGs) from prokaryotic species, 1334 of which were previously undescribed. These genomes account for ~69% of sequenced DNA and provide insight into sewage microbial dynamics. Rotterdam (Netherlands) and Copenhagen (Denmark) show strong seasonal microbial community shifts, while Bologna, Rome, (Italy) and Budapest (Hungary) have occasional blooms of Pseudomonas-dominated communities, accounting for up to ~95% of sample DNA. Seasonal shifts and blooms present challenges for effective sewage surveillance. We find that bacteria of known shared origin, like human gut microbiota, form communities, suggesting the potential for source-attributing novel species and their ARGs through network community analysis. This could significantly improve AMR tracking in urban environments.

Objective: Relapsing polychondritis (RP) is a rare autoimmune inflammatory disease that attacks mainly cartilaginous structures or causes serious damage in proteoglycan-rich structures (the eyes, heart, blood vessels, inner ear). This study shows results regarding the epidemiology, progression, and associations of this highly variable disease by collecting all cases from a 124-million-person-year Central European nationwide cohort. Methods: We used the Hungarian Health Care Database to identify all persons with possible RP infection. We followed patients who had International Classification of Diseases 10th edition code M94.1 at least once in their inpatient or outpatient records between January 1, 2002 and December 31, 2013 in Hungary. We classified these patients into disease severity groups by their drug consumption patterns between January 1, 2010 and December 31, 2013. We analyzed the regional distribution of RP incidences as well. Overall maps of comorbidity are presented with network layouts. Results: We identified 256 patients with RP among cumulatively 11.5 million registered inhabitants. We classified these patients into four severity classes as "extremely mild" (n=144), "mild" (n=22), "moderate" (n=41), and "severe" (n=4). Two additional groups were defined for patients without available drug data as "suspected only" (n=23) and "confirmed but unknown treatment" (n=22). The age and sex distributions of patients were similar to worldwide statistics. Indeed, the overall survival was good (95% confidence interval for 5 years was 83.6%–92.9% and for 10 years was 75.0%–88.3% which corresponds to the overall survival of the general population in Hungary), and the associations with other autoimmune disorders were high (56%) in Hungary. Almost any disease can occur with RP; however, the symptoms of chromosomal abnormalities are only incidental. Spondylosis can be a sign of the activation of RP, while Sjögren syndrome is the most frequent autoimmune association. Regional distribution of incidences suggests arsenic drinking water and sunlight exposure as possible triggering factors. Conclusion: The good survival rate of RP in Hungary is probably associated with the early diagnosis of the disease.

We investigate the fundamental statistical features of tagged (or annotated) networks having a rich variety of attributes associated with their nodes. Tags (attributes, annotations, properties, features, etc) provide essential information about the entity represented by a given node, thus, taking them into account represents a significant step towards a more complete description of the structure of large complex systems. Our main goal here is to uncover the relations between the statistical properties of the node tags and those of the graph topology. In order to better characterize the networks with tagged nodes, we introduce a number of new notions, including tag-assortativity (relating link probability to node similarity), and new quantities, such as node uniqueness (measuring how rarely the tags of a node occur in the network) and tag-assortativity exponent. We apply our approach to three large networks representing very different domains of complex systems. A number of the tag related quantities display analogous behaviour (e.g. the networks we studied are tag-assortative, indicating possible universal aspects of tags versus topology), while some other features, such as the distribution of the node uniqueness, show variability from network to network allowing for pin-pointing large scale specific features of real-world complex networks. We also find that for each network the topology and the tag distribution are scale invariant, and this self-similar property of the networks can be well characterized by the tag-assortativity exponent, which is specific to each system.